gm(1)

🎀 NAME 🎀

Okay bestie, meet gm(1) — the command-line utility to create, edit, compare, convert, or display images. It’s the main-character GraphicsMagick executable, and it’s here to make your terminal do the absolute most. 💅✨

✨ SYNOPSIS ✨

No cap, here’s the exact shape of every gm sub-command. Keep these literal, bestie — copy-paste is self-care. 🫶

gm animate [ options ... ] file [ [ options ... ] file ... ]

gm batch [ options ... ] [ script ]

gm benchmark [ options ... ] subcommand

gm compare [ options ... ] reference-image [ options ... ] compare-image
[ options ... ]

gm composite [ options ... ] change-image base-image [ mask-image ]
output-image

gm conjure [ options ] script.msl [ [ options ] script.msl ]

gm convert [ [ options ... ] [ input-file ...  ] [ options ... ] ]
output-file

gm display [ options ... ] file ...  [ [options ... ]file ... ]

gm identify file [ file ... ]

gm import [ options ... ] file

gm mogrify [ options ... ] file ...

gm montage [ options ... ] file [ [ options ... ] file ... ] output-file

gm time subcommand

gm version

🌸 DESCRIPTION 🌸

GraphicsMagick’s gm provides a suite of utilities for creating, comparing, converting, editing, and displaying images. All of the utilities are provided as sub-commands of a single gm executable. The gm executable returns the exit code 0 to indicate success, or 1 to indicate failure:

• animate — displays an animation (e.g. a GIF file) on any workstation display running an X server. Cute moving pics? Obsessed. 🎞️
• batch — executes an arbitrary number of the utility commands (e.g. convert) in the form of a simple linear batch script in order to improve execution efficiency, and/or to allow use as a subordinate co-process under the control of an arbitrary script or program. Efficiency queen behavior. 👑
• benchmark — executes one of the other utility commands (e.g. convert) for a specified number of iterations, or execution time, and reports execution time and other profiling information such as CPU utilization. Benchmark provides various operating modes including executing the command with a varying number of threads, and alternate reporting formats such as comma-separated value (CSV). Numbers, threads, CSV — the whole nerdy slay. 📊
• compare — compares two images and reports difference statistics according to specified metrics and/or outputs an image with a visual representation of the differences. It may also be used to test if images are similar within a particular range and specified metric, returning a truth value to the executing environment. Spot-the-difference but make it command line. 🔍
• composite — composites images (blends or merges images together) to create new images. Collab energy, but for pixels. 🫶
• conjure — interprets and executes scripts in the Magick Scripting Language (MSL). Script magic, literally. 🪄
• convert — converts an input file using one image format to an output file with the same or differing image format while applying an arbitrary number of image transformations. Glow-up pipeline, no cap. 🔥
• display — is a machine architecture independent image processing and display facility. It can display an image on any workstation display running an X server. Put the image on the screen and let it serve. 🖥️
• identify — describes the format and characteristics of one or more image files. It will also report if an image is incomplete or corrupt. Detective mode: activated. 🕵️‍♀️
• import — reads an image from any visible window on an X server and outputs it as an image file. You can capture a single window, the entire screen, or any rectangular portion of the screen. Screenshot goblin era. 📸
• mogrify — transforms an image or a sequence of images. These transforms include image scaling, image rotation, color reduction, and others. The transmogrified image overwrites the original image. ⚠️ That’s the whole warning, bestie — it replaces the original, so maybe keep a copy if you’re feeling precious. 🙈
• montage — creates a composite by combining several separate images. The images are tiled on the composite image with the name of the image optionally appearing just below the individual tile. Mood board, but make it files. 🌈
• time — executes a subcommand and reports the user, system, and total execution time consumed. Timing the tea, literally. ⏱️
• version — reports the GraphicsMagick release version, maximum sample-depth, copyright notice, supported features, and the options used while building the software. Receipts: provided. 🧾

The GraphicsMagick utilities recognize the following image formats:

Okay so this next bit is the whole format menu — every file format GraphicsMagick knows about, what it's called, and which directions it works in. It's basically the Cheesecake Factory menu of image formats: long, iconic, a little overwhelming, but everything's labeled so you're fine. 💅

The three little letters next to each name are the mode, and we'll decode those in a sec. Read left to right: r = can read, w = can write, + = multi-image. A - just means "nope, not that one, bestie."

---

• 3FR r-- — Hasselblad Photo RAW
• 8BIM rw- — Photoshop resource format
• 8BIMTEXT rw- — Photoshop resource text format
• 8BIMWTEXT rw- — Photoshop resource wide text format
• APP1 rw- — Raw application information
• APP1JPEG rw- — Raw JPEG binary data
• ART r-- — PF1: 1st Publisher
• ARW r-- — Sony Alpha DSLR RAW
• AVS rw+ — AVS X image
• BIE rw- — Joint Bi-level Image experts Group interchange format
• BMP rw+ — Microsoft Windows bitmap image
• BMP2 -w- — Microsoft Windows bitmap image v2
• BMP3 -w- — Microsoft Windows bitmap image v3
• CACHE --- — Magick Persistent Cache image format
• CALS rw- — Continuous Acquisition and Life-cycle Support Type 1 image
• CAPTION r-- — Caption (requires separate size info)
• CIN rw- — Kodak Cineon Format
• CMYK rw- — Raw cyan, magenta, yellow, and black samples (8 or 16 bits, depending on the image depth)
• CMYKA rw- — Raw cyan, magenta, yellow, black, and matte samples (8 or 16 bits, depending on the image depth)
• CR2 r-- — Canon Photo RAW
• CRW r-- — Canon Photo RAW
• CUR r-- — Microsoft Cursor Icon
• CUT r-- — DR Halo
• DCM r-- — Digital Imaging and Communications in Medicine image
• DCR r-- — Kodak Photo RAW
• DCX rw+ — ZSoft IBM PC multi-page Paintbrush
• DNG r-- — Adobe Digital Negative
• DPX rw- — Digital Moving Picture Exchange
• EPDF rw- — Encapsulated Portable Document Format
• EPI rw- — Adobe Encapsulated PostScript Interchange format
• EPS rw- — Adobe Encapsulated PostScript
• EPS2 -w- — Adobe Level II Encapsulated PostScript
• EPS3 -w- — Adobe Level III Encapsulated PostScript
• EPSF rw- — Adobe Encapsulated PostScript
• EPSI rw- — Adobe Encapsulated PostScript Interchange format
• EPT rw- — Adobe Encapsulated PostScript with MS-DOS TIFF preview
• EPT2 rw- — Adobe Level II Encapsulated PostScript with MS-DOS TIFF preview
• EPT3 rw- — Adobe Level III Encapsulated PostScript with MS-DOS TIFF preview
• EXIF rw- — Exif digital camera binary data
• FAX rw+ — Group 3 FAX (Not TIFF Group3 FAX!)
• FITS rw- — Flexible Image Transport System
• FRACTAL r-- — Plasma fractal image
• FPX rw- — FlashPix Format
• GIF rw+ — CompuServe graphics interchange format
• GIF87 rw- — CompuServe graphics interchange format (version 87a)
• GRADIENT r-- — Gradual passing from one shade to another
• GRAY rw+ — Raw gray samples (8/16/32 bits, depending on the image depth)
• HISTOGRAM -w- — Histogram of the image
• HRZ r-- — HRZ: Slow scan TV
• HTML -w- — Hypertext Markup Language and a client-side image map
• ICB rw+ — Truevision Targa image
• ICC rw- — ICC Color Profile
• ICM rw- — ICC Color Profile
• ICO r-- — Microsoft icon
• ICON r-- — Microsoft icon
• IDENTITY r-- — Hald CLUT identity image
• IMAGE r-- — GraphicsMagick Embedded Image
• INFO -w+ — Image descriptive information and statistics
• IPTC rw- — IPTC Newsphoto
• IPTCTEXT rw- — IPTC Newsphoto text format
• IPTCWTEXT rw- — IPTC Newsphoto wide text format
• JBG rw+ — Joint Bi-level Image experts Group interchange format
• JBIG rw+ — Joint Bi-level Image experts Group interchange format
• JNG rw- — JPEG Network Graphics
• JP2 rw- — JPEG-2000 JP2 File Format Syntax
• JPC rw- — JPEG-2000 Code Stream Syntax
• JPEG rw- — Joint Photographic Experts Group JFIF format
• JPG rw- — Joint Photographic Experts Group JFIF format
• K25 r-- — Kodak Photo RAW
• KDC r-- — Kodak Photo RAW
• LABEL r-- — Text image format
• M2V rw+ — MPEG-2 Video Stream
• MAP rw- — Colormap intensities and indices
• MAT r-- — MATLAB image format
• MATTE -w+ — MATTE format
• MIFF rw+ — Magick Image File Format
• MNG rw+ — Multiple-image Network Graphics
• MONO rw- — Bi-level bitmap in least-significant-byte-first order
• MPC rw+ — Magick Persistent Cache image format
• MPEG rw+ — MPEG-1 Video Stream
• MPG rw+ — MPEG-1 Video Stream
• MRW r-- — Minolta Photo Raw
• MSL r-- — Magick Scripting Language
• MTV rw+ — MTV Raytracing image format
• MVG rw- — Magick Vector Graphics
• NEF r-- — Nikon Electronic Format
• NULL r-- — Constant image of uniform color
• OTB rw- — On-the-air bitmap
• P7 rw+ — Xv thumbnail format
• PAL rw- — 16bit/pixel interleaved YUV
• PALM rw- — Palm Pixmap
• PBM rw+ — Portable bitmap format (black and white)
• PCD rw- — Photo CD
• PCDS rw- — Photo CD
• PCL -w- — Page Control Language
• PCT rw- — Apple Macintosh QuickDraw/PICT
• PCX rw- — ZSoft IBM PC Paintbrush
• PDB rw+ — Palm Database ImageViewer Format
• PDF rw+ — Portable Document Format
• PEF r-- — Pentax Electronic File
• PFA r-- — TrueType font
• PFB r-- — TrueType font
• PGM rw+ — Portable graymap format (gray scale)
• PGX r-- — JPEG-2000 VM Format
• PICON rw- — Personal Icon
• PICT rw- — Apple Macintosh QuickDraw/PICT
• PIX r-- — Alias/Wavefront RLE image format
• PLASMA r-- — Plasma fractal image
• PNG rw- — Portable Network Graphics
• PNG24 rw- — Portable Network Graphics, 24 bit RGB opaque only
• PNG32 rw- — Portable Network Graphics, 32 bit RGBA semitransparency OK
• PNG8 rw- — Portable Network Graphics, 8-bit indexed, binary transparency only
• PNM rw+ — Portable anymap
• PPM rw+ — Portable pixmap format (color)
• PREVIEW -w- — Show a preview an image enhancement, effect, or f/x
• PS rw+ — Adobe PostScript
• PS2 -w+ — Adobe Level II PostScript
• PS3 -w+ — Adobe Level III PostScript
• PSD rw- — Adobe Photoshop bitmap
• PTIF rw- — Pyramid encoded TIFF
• PWP r-- — Seattle Film Works
• RAF r-- — Fuji Photo RAW
• RAS rw+ — SUN Rasterfile
• RGB rw+ — Raw red, green, and blue samples
• RGBA rw+ — Raw red, green, blue, and matte samples
• RLA r-- — Alias/Wavefront image
• RLE r-- — Utah Run length encoded image
• SCT r-- — Scitex HandShake
• SFW r-- — Seattle Film Works
• SGI rw+ — Irix RGB image
• SHTML -w- — Hypertext Markup Language and a client-side image map
• STEGANO r-- — Steganographic image
• SUN rw+ — SUN Rasterfile
• SVG rw+ — Scalable Vector Gaphics (yes, "Gaphics" — that's literally how it's spelled here, don't come for me 🙈)
• TEXT rw+ — Raw text
• TGA rw+ — Truevision Targa image
• TIFF rw+ — Tagged Image File Format
• TILE r-- — Tile image with a texture
• TIM r-- — PSX TIM
• TOPOL r-- — TOPOL X Image
• TTF r-- — TrueType font
• TXT rw+ — Raw text
• UIL -w- — X-Motif UIL table
• UYVY rw- — 16bit/pixel interleaved YUV
• VDA rw+ — Truevision Targa image
• VICAR rw- — VICAR rasterfile format
• VID rw+ — Visual Image Directory
• VIFF rw+ — Khoros Visualization image
• VST rw+ — Truevision Targa image
• WBMP rw- — Wireless Bitmap (level 0) image
• WMF r-- — Windows Metafile
• WPG r-- — Word Perfect Graphics
• X rw- — X Image
• X3F r-- — Foveon X3 (Sigma/Polaroid) RAW
• XBM rw- — X Windows system bitmap (black and white)
• XC r-- — Constant image uniform color
• XCF r-- — GIMP image
• XMP rw- — Adobe XML metadata
• XPM rw- — X Windows system pixmap (color)
• XV rw+ — Khoros Visualization image
• XWD rw- — X Windows system window dump (color)
• YUV rw- — CCIR 601 4:1:1 or 4:2:2 (8-bit only)

---

💫 Modes

Okay so the mode letters — here's the decoder ring, keep it in your notes app bestie:

• r — Read
• w — Write
• + — Multi-image

---

Support for some of these formats require additional programs or libraries. See README in the source package for where to find optional additional software.

🫶 Real talk: "some of these need extra software installed" is doing a LOT of heavy lifting there. Check the README, it's your friend.

Note, a format delineated with + means that if more than one image is specified, frames are combined into a single multi-image file. Use +adjoin if you want a single image produced for each frame.

Your installation might not support all of the formats in the list. To get an accurate listing of the formats supported by your particular configuration, run gm convert -list format.

⭐ That's the move, periodt — don't trust this list blindly, ask your own install what it can actually do.

Raw images are expected to have one byte per pixel unless gm is compiled in 16-bit quantum mode or in 32-bit quantum mode. Here, the raw data is expected to be stored two or four bytes per pixel, respectively, in most-significant-byte-first order. For example, you can tell if gm was compiled in 16-bit mode by typing gm version without any options, and looking for Q:16 in the first line of output.

💀 This paragraph is the "check the quantum depth before you feed it raw bytes" warning, and it is NOT a vibe to ignore. Wrong byte order = cursed confetti image. You've been warned.

✨ FILES AND FORMATS ✨

Okay so bestie, here's how gm figures out what kind of image it's looking at 💅 By default, the image format is determined by its magic number — aka the first few bytes of the file. Sneaky but iconic.

Wanna force a specific format? Two ways:

• put an image format name and a colon in front of the filename (like ps:image)
• or just name the image type as the filename suffix (like image.ps)

The precedence rules are giving hierarchy, so pay attention 👑 The magic number takes precedence over the filename suffix, and the prefix takes precedence over the magic number and the suffix in input files. When a file is read, its magic number gets stored in the "image->magick" string.

For output files it flips a little: the prefix takes precedence over the filename suffix, and the filename suffix takes precedence over the "image->magick" string. Slay, consistent-ish 🎀

To read the "built-in" formats (GRANITE, H, LOGO, NETSCAPE, PLASMA, and ROSE) use a prefix (including the colon) without a filename or suffix. To read the XC format, follow the colon with a color specification. To read the CAPTION format, follow the colon with a text string or with a filename prefixed with the at symbol (@). Cute, right? 🦋

When you specify X as your image type, the filename has special meaning, bestie — it specifies an X window by id, name, or root. If no filename is specified, the window is selected by clicking the mouse in the desired window. Literally point-and-click energy 🖱️✨

Specify input_file as - for standard input, output_file as - for standard output. And the compression glow-up: if input_file has the extension .Z or .gz, the file is uncompressed with uncompress or gunzip respectively. If output_file has the extension .Z or .gz, the file is compressed using compress or gzip respectively. Auto-zip, no cap 🫶

Use an optional index enclosed in brackets after an input file name to specify a desired subimage of a multi-resolution image format like Photo CD (e.g. "img0001.pcd[4]") or a range for MPEG images (e.g. "video.mpg[50-75]"). A subimage specification can be disjoint (e.g. "image.tiff[2,7,4]"). For raw images, specify a subimage with a geometry (e.g. -size 640x512 "image.rgb[320x256+50+50]"). And surround the image name with quotation marks to prevent your shell from interpreting the square brackets — the shell is nosy, protect your brackets 🙈

Single images are written with the filename you specify. However, multi-part images (e.g., a multi-page PostScript document with +adjoin specified) may be written with the scene number included as part of the filename. To include the scene number in the filename, you gotta include a printf-style %d format specification in the file name and use the +adjoin option. For example:

image%02d.miff

writes files image00.miff, image01.miff, etc. Obsessed with a clean sequence 🌈

Only a single specification is allowed within an output filename. If more than one specification is present, it will be ignored. It is best to embed the scene number in the base part of the file name, not in the extension, because the extension will not be a recognizable image type. 💀 (learn from the chaos, don't put it in the extension)

When running a commandline utility, you can prepend an at sign @ to a filename to read a list of image filenames from that file. This is convenient in the event you have too many image filenames to fit on the command line. Game changer for the file-hoarders ✨

---

💅 OPTIONS 💅

Options are processed in command line order. Any option you specify on the command line remains in effect for the set of images that follows, until the set is terminated by the appearance of any option or -noop. Sticky, like a good setting should be 🍓

Some options only affect the decoding of images and others only the encoding. The latter can appear after the final group of input images. So yes, you can be a little chaotic with ordering — but only that specific way, iykyk.

This is a combined list of the command-line options used by the GraphicsMagick utilities (animate, compare, composite, convert, display, identify, import, mogrify and montage). One big shared closet, all the fits 👑

In this document, angle brackets ("<>") enclose variables and curly brackets ("{}") enclose optional parameters. For example, -fuzz <distance>{%} means you can use the option -fuzz 10 or -fuzz 2%. See? Not scary 🎀

• -adjoin — join images into a single multi-image file

By default, all images of an image sequence are stored in the same file. However, some formats (e.g. JPEG) do not support storing more than one image per file and only the first frame in an image sequence will be saved unless the result is saved to separate files. Use +adjoin to force saving multiple frames to multiple numbered files. If +adjoin is used, then the output filename must include a printf style formatting specification for the numeric part of the filename. For example,

image%02d.miff

• -affine <matrix> — drawing transform matrix

This option provides a transform matrix {sx,rx,ry,sy,tx,ty} for use by subsequent -draw or -transform options.

• -antialias — remove pixel aliasing

By default antialiasing algorithms are used when drawing objects (e.g. lines) or rendering vector formats (e.g. WMF and Postscript). Use +antialias to disable use of antialiasing algorithms. Reasons to disable antialiasing include avoiding increasing colors in the image, or improving rendering speed. Speed vs. smooth, pick your fighter ⭐

• -append — append a set of images

This option creates a single image where the images in the original set are stacked top-to-bottom. If they are not of the same width, any narrow images will be expanded to fit using the background color. Use +append to stack images left-to-right. The set of images is terminated by the appearance of any option. If the -append option appears after all of the input images, all images are appended. Stack 'em like a scrapbook 💖

• -asc-cdl <spec> — apply ASC CDL color transform

Applies ("bakes in") the ASC CDL, which is a format for the exchange of basic primary color grading information between equipment and software from different manufacturers. The format defines the math for three functions: slope, offset and power. Each function uses a number for the red, green, and blue color channels for a total of nine numbers comprising a single color decision. The tenth number (optional) is for chromiance (saturation) as specified by ASC CDL 1.2.

The argument string is comma delimited and is in the following form (but without invervening spaces or line breaks)

redslope,redoffset,redpower:
greenslope,greenoffset,greenpower:
blueslope,blueoffset,bluepower:
saturation

with the unity (no change) specification being:

"1.0,0.0,1.0:1.0,0.0,1.0:1.0,0.0,1.0:1.0"

• -authenticate <string> — decrypt image with this password

Use this option to supply a password for decrypting an image or an image sequence, if it is being read from a format such as PDF that supports encryption. Encrypting images being written is not supported. Read-only secrecy, bestie 🔐

• -auto-orient — orient (rotate) image so it is upright

Adjusts the image orientation so that it is suitable for viewing. Uses the orientation tag obtained from the image file or as supplied by the -orient option. Never post a sideways pic again 😭

• -average — average a set of images

The set of images is terminated by the appearance of any option. If the -average option appears after all of the input images, all images are averaged.

• -backdrop — display the image centered on a backdrop.

This backdrop covers the entire workstation screen and is useful for hiding other X window activity while viewing the image. The color of the backdrop is specified as the foreground color (X11 default is black). Refer to "X Resources", below, for details. Full-screen main character energy 👑

• -background <color> — the background color

The color is specified using the format described under the -fill option.

• -black-threshold red[,green][,blue][,opacity] — pixels below the threshold become black

Use -black-threshold to set pixels with values below the specified threshold to minimum value (black). If only one value is supplied, or the red, green, and blue values are identical, then intensity thresholding is used. If the color threshold values are not identical then channel-based thresholding is used, and color distortion will occur. Specify a negative value (e.g. -1) if you want a channel to be ignored but you do want to threshold a channel later in the list. If a percent (%) symbol is appended, then the values are treated as a percentage of maximum range. Thresholding is a vibe check for pixels, fr fr 🖤

• -blue-primary <x>,<y> — blue chromaticity primary point
• -blur <radius>{x<sigma>} — blur the image with a Gaussian operator

Blur with the given radius and standard deviation (sigma).

• -border <width>x<height> — surround the image with a border of color

See -geometry for details about the geometry specification.

• -bordercolor <color> — the border color

The color is specified using the format described under the -fill option.

• -borderwidth <geometry> — the border width
• -box <color> — set the color of the annotation bounding box

The color is specified using the format described under the -fill option.

See -draw for further details.

• -channel <type> — the type of channel

Choose from: Red, Green, Blue, Opacity, Matte, Cyan, Magenta, Yellow, Black, or Gray.

Use this option to extract a particular channel from the image. Opacity, for example, is useful for extracting the opacity values from an image. Channel your inner channel 🦋

• -charcoal <factor> — simulate a charcoal drawing
• -chop <width>x<height>{+-}<x>{+-}<y>{%} — remove pixels from the interior of an image

Width and height give the number of columns and rows to remove, and x and y are offsets that give the location of the leftmost column and topmost row to remove.

The x offset normally specifies the leftmost column to remove. If the -gravity option is present with NorthEast, East, or SouthEast gravity, it gives the distance leftward from the right edge of the image to the rightmost column to remove. Similarly, the y offset normally specifies the topmost row to remove, but if the -gravity option is present with SouthWest, South, or SouthEast gravity, it specifies the distance upward from the bottom edge of the image to the bottom row to remove.

The -chop option removes entire rows and columns, and moves the remaining corner blocks leftward and upward to close the gaps. Cutting out a whole strip and gluing the rest back together — arts and crafts era ✂️💅

• -clip — apply the clipping path, if one is present

If a clipping path is present, it will be applied to subsequent operations.

For example, if you type the following command:

gm convert -clip -negate cockatoo.tif negated.tif

Only the pixels inside the clipping path get negated, bestie. That’s the whole vibe. The -clip feature requires the XML library — and if the XML library isn’t there, the option is literally ignored. 😭

🌸 -coalesce

Merge a sequence of images. Each image N in the sequence after Image 0 is replaced with the image created by flattening images 0 through N.

The set of images is terminated by the appearance of any option. If the -coalesce option appears after all of the input images, all images are coalesced. Slay. ✨

🎀 -colorize <value>

Colorize the image with the pen color. Specify the amount of colorization as a percentage. You can apply separate colorization values to the red, green, and blue channels of the image with a colorization value list delimited with slashes (e.g. 0/0/50).

The -colorize option may be used in conjunction with -modulate to produce a nice sepia toned image like:

gm convert input.ppm -modulate 115,0,100 \
          -colorize 7,21,50 output.ppm.

🦋 -colormap <type>

Define the colormap type. Choose between shared or private.

This option only applies when the default X server visual is PseudoColor or GRAYScale. Refer to -visual for more details. By default, a shared colormap is allocated. The image shares colors with other X clients. Some image colors could be approximated, therefore your image may look very different than intended. Choose Private and the image colors appear exactly as they are defined. However, other clients may go technicolor when the image colormap is installed. 💀

💖 -colors <value>

Preferred number of colors in the image.

The actual number of colors in the image may be less than your request, but never more. Note, this is a color reduction option. Images with less unique colors than specified with this option will have any duplicate or unused colors removed. The ordering of an existing color palette may be altered. When converting an image from color to grayscale, convert the image to the gray colorspace before reducing the number of colors since doing so is most efficient. Refer to quantize for more details (the page lives at quantize.html).

Note, options -dither, -colorspace, and -treedepth affect the color reduction algorithm. 👑

🌈 -colorspace <value>

The type of colorspace.

Choices are: CineonLog, CMYK, GRAY, HSL, HWB, OHTA, RGB, Rec601Luma, Rec709Luma, Rec601YCbCr, Rec709YCbCr, Transparent, XYZ, YCbCr, YIQ, YPbPr, or YUV.

Color reduction, by default, takes place in the RGB color space. Empirical evidence suggests that distances in color spaces such as YUV or YIQ correspond to perceptual color differences more closely than do distances in RGB space. These color spaces may give better results when color reducing an image. Refer to quantize for more details.

Two gray colorspaces are supported. The Rec601Luma space is based on the recommendations for legacy NTSC television (ITU-R BT.601-5). The Rec709Luma space is based on the recommendations for HDTV (Rec. ITU-R BT.709-5) and is suitable for use with computer graphics, and for contemporary CRT displays. The GRAY colorspace currently selects the Rec601Luma colorspace by default for backwards compatibility reasons. This default may be re-considered in the future.

Two YCbCr colorspaces are supported. The Rec601YCbCr space is based on the recommendations for legacy NTSC television (ITU-R BT.601-5). The Rec709CbCr space is based on the recommendations for HDTV (Rec. ITU-R BT.709-5) and is suitable for use with computer graphics, and for contemporary CRT displays. The YCbCr colorspace specification is equivalent to Rec601YCbCr.

The Transparent color space behaves uniquely in that it preserves the matte channel of the image if it exists.

The -colors or -monochrome option, or saving to a file format which requires color reduction, is required for this option to take effect. 🫶

🍓 -comment <string>

Annotate an image with a comment.

Use this option to assign a specific comment to the image, when writing to an image format that supports comments. You can include the image filename, type, width, height, or other image attribute by embedding special format characters listed under the -format option. The comment is not drawn on the image, but is embedded in the image datastream via a "Comment" tag or similar mechanism. If you want the comment to be visible on the image itself, use the -draw option instead.

For example,

-comment "%m:%f %wx%h"

produces an image comment of MIFF:bird.miff 512x480 for an image titled bird.miff and whose width is 512 and height is 480.

If the first character of string is @, the image comment is read from a file titled by the remaining characters in the string. Please note that if the string comes from an untrusted source that it should be sanitized before use since otherwise the content of an arbitrary readable file could be incorporated in a comment in the output file (a security risk). 🚨

If the -comment option appears multiple times, only the last comment is stored.

In PNG images, the comment is stored in a tEXt or zTXt chunk with the keyword "comment".

💅 -compose <operator>

The type of image composition.

The description of composition uses abstract terminology in order to allow the description to be more clear, while avoiding constant values which are specific to a particular build configuration. Each image pixel is represented by red, green, and blue levels (which are equal for a gray pixel). MaxRGB is the maximum integral value which may be stored in the red, green, or blue channels of the image. Each image pixel may also optionally (if the image matte channel is enabled) have an associated level of opacity (ranging from opaque to transparent), which may be used to determine the influence of the pixel color when compositing the pixel with another image pixel. If the image matte channel is disabled, then all pixels in the image are treated as opaque. The color of an opaque pixel is fully visible while the color of a transparent pixel color is entirely absent (pixel color is ignored).

By definition, raster images have a rectangular shape. All image rows are of equal length, and all image columns have the same number of rows. By treating the opacity channel as a visual "mask" the rectangular image may be given a "shape" by treating the opacity channel as a cookie-cutter for the image. Pixels within the shape are opaque, while pixels outside the shape are transparent. Pixels on the boundary of the shape may be between opaque and transparent in order to provide antialiasing (visually smooth edges). The description of the composition operators use this concept of image "shape" in order to make the description of the operators easier to understand. While it is convenient to describe the operators in terms of "shapes" they are by no means limited to mask-style operations since they are based on continuous floating-point mathematics rather than simple boolean operations.

By default, the Over composite operator is used. The following composite operators are available:

Over
In
Out
Atop
Xor
Plus
Minus
Add
Subtract
Difference
Divide
Multiply
Bumpmap
Copy
CopyRed
CopyGreen
CopyBlue
CopyOpacity
CopyCyan
CopyMagenta
CopyYellow
CopyBlack

The behavior of each operator is described below.

• Over — The result will be the union of the two image shapes, with opaque areas of change-image obscuring base-image in the region of overlap.
• In — The result is simply change-image cut by the shape of base-image. None of the image data of base-image will be in the result.
• Out — The resulting image is change-image with the shape of base-image cut out.
• Atop — The result is the same shape as base-image, with change-image obscuring base-image where the image shapes overlap. Note this differs from over because the portion of change-image outside base-image's shape does not appear in the result.
• Xor — The result is the image data from both change-image and base-image that is outside the overlap region. The overlap region will be blank.
• Plus — The result is just the sum of the image data. Output values are cropped to MaxRGB (no overflow). This operation is independent of the matte channels.
• Minus — The result of change-image - base-image, with underflow cropped to zero. The matte channel is ignored (set to opaque, full coverage).
• Add — The result of change-image + base-image, with overflow wrapping around (mod MaxRGB+1).
• Subtract — The result of change-image - base-image, with underflow wrapping around (mod MaxRGB+1). The add and subtract operators can be used to perform reversible transformations.
• Difference — The result of abs(change-image - base-image). This is useful for comparing two very similar images.
• Divide — The result of change-image / base-image. This is useful for improving the readability of text on unevenly illuminated photos (by using a gaussian blurred copy of change-image as base-image).

Multiply

💫 (...the operator from the last part) — the result of change-image * base-image

Obsessed with this one for drop-shadows specifically, like it's literally the shadow girlie of the operator fam.

💫 Bumpmap

Your base-image, but shaded by change-image. Like contouring, but for pixels 🎀

💫 Copy

The resulting image is base-image replaced with change-image. Heads up though — here the matte information is ignored, okay? She said "no matte today" 🙈

💫 CopyRed

Resulting image = the red channel in base-image replaced with the red channel in change-image. The other channels are copied untouched, slay 💅

💫 CopyGreen

Total same energy but green: the green channel in base-image gets replaced with the green channel in change-image, and the other channels are copied untouched.

💫 CopyBlue

Blue edition bestie — the blue channel in base-image replaced with the blue channel in change-image. Other channels? Untouched, periodt.

💫 CopyOpacity

The opacity channel in base-image gets replaced with the opacity channel in change-image, and the other channels are copied untouched. Transparency glow-up ✨

💫 CopyCyan

The cyan channel in base-image gets swapped for the cyan channel in change-image, other channels copied untouched. ⚠️ But using this operator requires that base-image be in CMYK(A) colorspace — no exceptions, iykyk.

💫 CopyMagenta

Same deal, magenta era: magenta channel in base-image replaced with magenta channel in change-image, everything else copied untouched. Requires base-image in CMYK(A) colorspace 🌸

💫 CopyYellow

Yellow channel in base-image replaced with yellow channel in change-image, other channels copied untouched. Requires base-image be in CMYK(A) colorspace, no cap.

💫 CopyBlack

The black channel in base-image replaced with the black channel in change-image, other channels copied untouched. Requires base-image to be in CMYK(A) colorspace. And if change-image is not in CMYK space, then the change-image pixel intensities are used. That's the whole rule 🖤

---

🔥 -compress <type>

Literally just: what kind of image compression are we doing here.

The choices are: None, BZip, Fax, Group3, Group4, JPEG, Lossless, LZW, RLE, Zip, LZMA, JPEG2000, JPEG2000, JBIG, JBIG2, WebP, or ZSTD.

Slap +compress on there to store the binary image in an uncompressed format. The default is the compression type of the specified image file, mwah 💋

Lossless means lossless JPEG, which only exists if the JPEG library has been patched to support it. Use of lossless JPEG is generally not recommended — like, just don't, bestie 🫶

Use the -quality option to set the compression level used by the JPEG, JPEG-2000, PNG, MIFF, MPEG, and TIFF encoders. And use the -sampling-factor option to set the sampling factor that the DPX, JPEG, MPEG, and YUV encoders use for downsampling the chroma channels. Two different knobs, two different jobs ✨

---

⭐ -contrast

Enhance or reduce the image contrast.

This option enhances the intensity differences between the lighter and darker elements of the image. Use -contrast to enhance, or +contrast to reduce. Iconic behavior honestly.

Wanna go more pronounced? Just repeat the option, she doesn't mind:

gm convert rose: -contrast -contrast rose_c2.png

---

🌀 -convolve <kernel>

Convolve the image with the specified convolution kernel. Very math-girl of her.

The kernel is a comma-separated list of floating point values, ordered left to right, starting with the top row. The order of the kernel is determined by the square root of the number of entries. And rn, presently, only square kernels are supported — sorry to the rectangle truthers 😭

---

🗂️ -create-directories

Create the output directory if it's required.

Use this one with -output-directory when the input paths contain subdirectories and you want similar subdirectories created in the output directory. Without this option, mogrify will fail if the required output directory does not exist. So like... use it, bestie 💀

---

✂️ -crop <width>x<height>{+-}<x>{+-}<y>{%}

Preferred size and location of the cropped image.

See -geometry for details about the geometry specification — that's where the whole recipe lives.

The width and height give the size of the image that remains after cropping, and x and y are offsets that give the location of the top left corner of the cropped image with respect to the original image. To specify the amount to be removed, use -shave instead. Different vibe, different flag 🪒

If the x and y offsets are present, a single image is generated, consisting of the pixels from the cropping region. The offsets specify the location of the upper left corner of the cropping region measured downward and rightward with respect to the upper left corner of the image. If the -gravity option is present with NorthEast, East, or SouthEast gravity, it gives the distance leftward from the right edge of the image to the right edge of the cropping region. Similarly, if the -gravity option is present with SouthWest, South, or SouthEast gravity, the distance is measured upward between the bottom edges. Gravity really said "I'll decide which direction we're measuring from" 👑

If the x and y offsets are omitted, a set of tiles of the specified geometry, covering the entire input image, is generated. The rightmost tiles and the bottom tiles are smaller if the specified geometry extends beyond the dimensions of the input image — she just trims to fit, no drama 💅

---

🎡 -cycle <amount>

Displace the image colormap by amount.

Amount defines the number of positions each colormap entry is shifted. Round and round the color wheel we go 🌈

---

🐛 -debug <events>

Enable debug printout. Detective mode activated 🔍

The events parameter specifies which events are to be logged. It can be either None, All, or a comma-separated list consisting of one or more of the following domains: Annotate, Blob, Cache, Coder, Configure, Deprecate, Error, Exception, FatalError, Information, Locale, Option, Render, Resource, TemporaryFile, Transform, User, Warning, or X11. For example, to log cache and blob events, use:

gm convert -debug "Cache,Blob" rose: rose.png

The User domain is normally empty, but developers can log User events in their private copy of GraphicsMagick. Fancy 🎀

Use the -log option to specify the format for debugging output.

Use +debug to turn off all logging. Peace and quiet, finally 😌

An alternative to using -debug is to use the MAGICK_DEBUG environment variable. The allowed values for the MAGICK_DEBUG environment variable are the same as for the -debug option. So pick your fighter, either one works 💖

---

🧩 -deconstruct

Break down an image sequence into its constituent parts.

This option compares each image with the next in a sequence and returns the maximum bounding region of any pixel differences it discovers. This method can undo a coalesced sequence returned by the -coalesce option, and it's useful for removing redundant information from a GIF or MNG animation. Basically it's the un-coalesce button 🦋

The sequence of images is terminated by the appearance of any option. And if the -deconstruct option appears after all of the input images, all images are deconstructed. Order matters, bestie 🫶

---

📝 -define <key>{=<value>},...

Add coder/decoder specific options. This option creates one or more definitions for coders and decoders to use while reading and writing image data.

Definitions may be passed to coders and decoders to control options that are specific to certain image formats. If value is missing for a definition, an empty-valued definition of a flag will be created with that name. This is used to control on/off options. Use +define <key>,... to remove definitions previously created. Use +define "*" to remove all existing definitions. Clean slate energy 🧼

The following definitions may be created:

🖼️ bmp:allow-jpeg={true|false}

If the bmp:allow-jpeg value is set to true, then enable BMP files using JPEG compression, under control of the compression option (which may be JPEG already if the input file is JPEG). JPEG-compressed BMP files are very rare, so this option acts as a safeguard to assure that the current compression setting does not accidentally produce JPEG-compressed BMP files which most other software is unable to read. It's giving protective older sibling, honestly 🛡️

🎞️ cineon:colorspace={rgb|cineonlog}

Use the cineon:colorspace option when reading a Cineon file to specify the colorspace the Cineon file uses. This overrides the colorspace type implied by the DPX header (if any). Trust the flag, not the header 💅

🎬 dpx:bits-per-sample=<value>

If the dpx:bits-per-sample key is defined, GraphicsMagick will write DPX images with the specified bits per sample, overriding any existing depth value. If this option is not specified, then the value is based on the existing image depth value from the original image file. The DPX standard supports bits per sample values of 1, 8, 10, 12, and 16. Many DPX readers demand a sample size of 10 bits with type A padding (see below). Fussy readers, no shade 🙈

🎨 dpx:colorspace={rgb|cineonlog}

Use the dpx:colorspace option when reading a DPX file to specify the colorspace the DPX file uses. This overrides the colorspace type implied by the DPX header (if any).

📦 dpx:packing-method={packed|a|b|lsbpad|msbpad}

DPX samples are output within 32-bit words. They may be tightly packed end-to-end within the words (packed), padded with null bits to the right of the sample (a or lsbpad), or padded with null bits to the left of the sample (b or msbpad). This option only has an effect for sample sizes of 10 or 12 bits. If samples are not packed, the DPX standard recommends type A padding. Many DPX readers demand a sample size of 10 bits with type A padding. Read that twice, it's a whole thing 😭

🔀 dpx:pixel-endian={lsb|msb}

okay so this dpx:endian thing — you can literally tell it the endian order of the pixels when it's reading or writing DPX files. Sometimes that's lowkey useful if the file is written incorrectly (or has to be written incorrectly 💀) so that the file header and the pixels end up using different endianness. mess, but slay, we handle it.

• dpx:swap-samples={true|false} — 🎀
• dpx:swap-samples-read={true|false} — 🎀
• dpx:swap-samples-write={true|false} — GraphicsMagick tries to be a good little DPX standard citizen, but some parts of that standard are genuinely confusing, no cap. So some 10-bit DPX files come out with Red and Blue interchanged, or Cb and Cr interchanged, because somebody interpreted the standard differently or literally got their wires crossed 🫠. You drop swap-samples when you're reading or writing so it uses the sample order you actually need. And if you want swapping to happen only in the reader, it's swap-samples-read; if it should only happen in the writer, it's swap-samples-write. iykyk.
• heif:ignore-transformations={true|false} — gives you the HEIF (e.g. HEIC/AVIF) image back with zero transformations applied, so no rotation, nothing. In ping mode (that's the identify default) the default value is true, but otherwise it's false. Heads up bestie: when ignore-transformations is false, identify may be much slower 😭.
• heif:interleaved-rgb-decode={true|false} — asks libheif to hand back pixels in interleaved representation instead of planar representation. Should have zero influence on the decoded result, but the performance may be different ✨.
• heif:preserve-colorspace={true|false} — returns the HEIF image in its original colorspace, so you can get it back in the original YCbCr colorspace rather than RGB 💅.
• heif:tile-threads=number — if a HEIF image is made of multiple tiles, each tile may be decoded by one thread, and several tiles can be decoded at once if you use more threads 🔥. The maximum reasonable value is the min of (number of CPU cores available) or (number of tiles in the HEIF file), whichever is smaller. If OpenMP is available, the default is the number of OpenMP threads; otherwise the default is 4.
• heif:max-image-size-pixels=number — an image can be comprised of so many pixels, and if there's a huge number of them it may take a long time to return an image ⏳. Please refer to libheif source code (or documentation) for max_image_size_pixels. Also note: GraphicsMagick already computes a value for this parameter based on total pixel limits, or width/height limits — so supplying this parameter will override that. choose wisely 🎀.
• heif:max-number-of-tiles=number — an image can be comprised of many tiles, and a huge number of tiles means it may take a long time to return an image. Please refer to libheif source code (or documentation) for max_number_of_tiles 🦋.
• heif:max-bayer-pattern-pixels=number — please refer to libheif source code (or documentation) for max_bayer_pattern_pixels 🌸.
• heif:max-items=number — please refer to libheif source code (or documentation) for max_items 🌸.
• heif:max-color-profile-size=number — please refer to libheif source code (or documentation) for max_color_profile_size 🌸.
• heif:max-memory-block-size=number — please refer to libheif source code (or documentation) for max_memory_block_size 🌸.
• heif:max-components=number — please refer to libheif source code (or documentation) for max_components 🌸.
• heif:max-iloc-extents-per-item=number — please refer to libheif source code (or documentation) for max_iloc_extents_per_item 🌸.
• heif:max-size-entity-group=number — please refer to libheif source code (or documentation) for max_size_entity_group 🌸.
• heif:max-children-per-box=number — please refer to libheif source code (or documentation) for max_children_per_box 🌸.
• heif:max-total-memory=number — please refer to libheif source code (or documentation) for max_total_memory 🌸.
• heif:max-sample-description-box-entries=number — please refer to libheif source code (or documentation) for max_sample_description_box_entries 🌸.
• heif:max-sample-group-description-box-entries=number — please refer to libheif source code (or documentation) for max_sample_group_description_box_entries 🌸.

bestie that's a lot of heif knobs and honestly? iconic of libheif. anyway 💖

• gradient:direction={South|North|West|East|NorthWest|NorthEast|SouthWest|SouthEast} — by default the gradient coder makes a gradient from top to bottom, which is South. Since GraphicsMagick 1.3.35, you can specify the gradient direction and get gradient vectors per a gravity-like spec 🌈. The arguments are: South (Top to Bottom), North (Bottom to Top), West (Right to Left), East (Left to Right), NorthWest (Bottom-Right to Top-Left), NorthEast (Bottom-Left to Top-Right), SouthWest (Top-Right Bottom-Left), and SouthEast (Top-Left to Bottom-Right). literally art ✨
• jp2:rate=<value> — sets the compression factor when writing JPEG-2000 files. The compression factor is the reciprocal of the compression ratio 🧮. Valid range is 0.0 to 1.0, with 1.0 meaning lossless compression. If defined, this value overrides the -quality setting. The default quality setting of 75 results in a rate value of 0.06641.
• jpeg:arithmetic-coding={true|false} — turns arithmetic encoding on or off if the JPEG library supports it (default disabled). When this is enabled, optimize-coding is necessarily disabled — they don't share 🚫.
• jpeg:block-smoothing={true|false} — enables or disables block smoothing when reading a JPEG file (default enabled) 🫶.
• jpeg:dct-method=<value> — picks which IJG JPEG library DCT implementation to use. The encoding implementations vary in speed and encoding error. The available choices for value are islow, ifast, float, default and fastest. Note that fastest might not necessarily be fastest on your CPU, depending on the choices made when the JPEG library was built and how your CPU behaves 😭 classic.
• jpeg:fancy-upsampling={true|false} — enables or disables fancy upsampling when reading a JPEG file (default enabled) 💅.
• jpeg:data-precision=<value> — selects the JPEG sample storage size. Normal JPEG uses 8. Turbo-jpeg and some variants of IJG JPEG also support 12 and 16 bits.
• jpeg:lossless-precision=<value> — selects a degree of precision to use while performing DPCM-based compression. The available range is 0 to precision-1. For example, if the output depth is 16, then the available range is 0 to 15. When the precision is non-zero, the pixels are right-shifted to effectively support fewer bits (fewer colors), but offering an improved compression ratio. Use a value of 0 to achieve lossless compression — that's the one for the perfectionists 👑.
• jpeg:lossless-predictor=<value> — selects the predictor to use for lossless compression. Supported values are in the range of 1 to 7 and select a pre-defined predictor algorithm to use, which considers surrounding pixels. The best predictor for a particular image depends on the image 🦋.
• jpeg:max-scan-number=<value> — an integer value for the maximum number of progressive scans allowed in a JPEG file. The default maximum is 100 scans. This limit is imposed because of a weakness in the JPEG standard which allows small JPEG files to take many minutes or hours to be read 💀 yeah. not a vibe.
• jpeg:max-warnings=<value> — an integer value for how many warnings are allowed for any given error type before being promoted to a hard error. JPEG files producing excessive warnings indicate a problem with the file 🚩.
• jpeg:optimize-coding={true|false} — selects if huffman encoding should be used. Huffman encoding is enabled by default, but may be disabled for very large images since encoding requires that the entire image be buffered in memory 😳. Huffman encoding produces smaller JPEG files at the expense of added compression time and memory consumption.
• jpeg:preserve-settings — if this flag is defined, the JPEG encoder will use the same quality and sampling-factor settings that were found in the input file, if the input was in JPEG format 🌟. These settings are also preserved if the input is a JPEG file and the output is a JNG file. If the colorspace of the output file differs from that of the input file, the quality setting is preserved but the sampling-factors are not. sorry bestie, rules are rules.
• jxl:effort=<integer> — encoder effort. Higher values allow more computation, generally achieving smaller output at the same quality 💪. For lossy encoding, higher effort should achieve more accurate quality. Please refer to libjxl source code (or documentation) for JXL_ENC_FRAME_SETTING_EFFORT.
• jxl:decodingspeed=<integer> — decoding speed. Higher values improve decode speed at the expense of quality or density ⚡. Please refer to libjxl source code (or documentation) for JXL_ENC_FRAME_SETTING_DECODING_SPEED.
• pcl:fit-to-page
• pcl:fit-to-page — set this flag and the printer gets asked to scale the image so it fits the page size (width and/or height). slay, no more awkward cropping 💅
• png:chunk-malloc-max=<value> — this one caps the biggest chunk size libpng is allowed to read. libpng's own default is normally 8,000,000 bytes. very rarely, you'll meet a totally valid PNG where it just throws "chunk data is too large" — and that's your cue to bump the limit with this option. 🫶

⚠️ bestie, be careful raising this limit: an excessively large limit could let untrusted files eat excessive memory. no cap, don't go feral with it 😭

• mng:maximum-loops=<value> — sets the max number of loops a MNG LOOP chunk is allowed to request. Without a limit in place, a MNG file could ask for up to 2147483647 loops, which could run for a very long time (like, an unhinged amount of time 💀). The current default limit is 512 loops.
• pdf:use-cropbox={true|false} — set this to true and Ghostscript is requested to apply the PDF crop box. simple as that ✨
• pdf:stop-on-error={true|false} — set to true and Ghostscript is requested to stop processing the PDF the moment the first error shows up. Otherwise it'll try to process all requested pages anyway. stubborn queen behavior 💜
• ps:imagemask — define this flag and the PS3 and EPS3 coders will create Postscript files that render bilevel images with the Postscript imagemask operator instead of the image operator. iconic little switch 🔥
• ptif:minimum-geometry=<geometry> — define this key and GraphicsMagick uses it to figure out the minimum frame size to output when writing a pyramid TIFF file (that's a TIFF containing a succession of reduced versions of the first frame, iykyk). The default minimum geometry is 32x32. 🎀
• tiff:alpha={unspecified|associated|unassociated} — tells GraphicsMagick the TIFF alpha channel type when reading or writing TIFF files, overriding the normal value. The default alpha channel type for new files is unspecified alpha. Existing alpha settings get preserved when converting from one TIFF file to another. When a TIFF file uses associated alpha, the image pixels are pre-multiplied (i.e. altered) with the alpha channel. Files with "associated" alpha appear as if they were alpha composited on a black background when the matte channel is disabled. Pick the unassociated alpha type and the alpha channel is saved without altering the pixels. Photoshop recognizes associated alpha as transparency information; if the file is saved with unassociated alpha, the alpha information is loaded as an independent channel. 📸

🫠 fun lore drop: for many years, ImageMagick and GraphicsMagick marked TIFF files as using associated alpha, without properly pre-multiplying the pixels. so yeah. that happened.

• tiff:fill-order={msb2lsb|lsb2msb} — define this key and GraphicsMagick uses it to determine the bit fill order used while writing TIFF files. The normal default is msb2lsb, which matches the native bit order of all modern CPUs. The only exception is when Group3 or Group4 FAX compression is requested, since FAX machines send data in bit-reversed order and therefore RFC 2301 recommends using reverse order. fax machines really said "we do things differently here" 💀
• tiff:group-three-options=<value> — define this key and GraphicsMagick uses it to set the group3 options tag when writing group3-compressed TIFF. Please see the TIFF specification for the usage of this tag. The default value is 4. 📎
• tiff:ignore-tags=<tags> — define this key and it's used as a list of comma-delimited integer TIFF tag values to ignore while reading the TIFF file. Super useful when you need to read files that would otherwise fail to read because of problems with TIFF tags. 🩹

⚠️ note that some TIFF tags are required in order to be able to read the image data at all — so don't just ignore everything, bestie.

• tiff:lerc-maxzerror=<value> — when LERC compression is selected (via -compress lerc), this specifies the max allowable floating point error for lossy compression (e.g. 0.5). The default compression value is 0 (lossless). 🌸
• tiff:lerc-compression={1|deflate|zip|2|zstd} — after LERC compression has taken place, an optional second compression pass using Deflate or Zstd compression may be performed. Specify a value of 1, deflate, or zip to use Deflate compression, or 2 or zstd to use Zstd compression. The argument is case-insensitive. ✨
• tiff:lerc-zip-quality=<value> — when LERC compression is enabled and a second-pass compression using Deflate has been selected, hand it a value in the range of 1 to 12 to indicate the Deflate compression level (larger is more compression). The default level for Deflate compression is 6. 🎀
• tiff:lerc-zstd-level=<value> — when LERC compression is enabled and a second-pass compression using Zstd has been selected, hand it a value in the range of 1 and 19 (or 22 using a lot more memory) to indicate the Zstd compression level (larger is more compression). The default level for Zstd compression is 3. 💜
• tiff:photometric={minisblack|miniswhite} — for grayscale and bilevel images, GraphicsMagick normally chooses the photometric based on bit depth and compression. Define this key and GraphicsMagick will use it to determine the photometric used while writing grayscale and bilevel images. Normally bilevel images are miniswhite and grayscale images are minisblack. 🖤🤍
• tiff:report-warnings={false|true} — define this key and set it to true and TIFF warnings get reported as a warning exception rather than as a coder log message. These warnings show up after the image has been read or written. Most TIFF warnings are benign, but sometimes they help you deduce problems with the TIFF file, or help detect that the TIFF file needs a special application to read successfully because of proprietary or specialized extensions. lowkey useful 👀
• tiff:sample-format={unsigned|ieeefp} — define this key and GraphicsMagick uses it to determine the sample format used while writing TIFF files. The default is unsigned. Specify ieeefp to write floating-point TIFF files with float (32-bit) or double (64-bit) values. Use the tiff:bits-per-sample define to determine the type of floating-point value to use. 🔢
• tiff:max-sample-value=<value> — define this key and GraphicsMagick uses the assigned value as the maximum floating point value while reading or writing IEEE floating point TIFFs. Otherwise the max value is 1.0, or the value obtained from the file's SMaxSampleValue tag (if present). 🌟

💀 heads up: the floating point data is currently not scanned in advance to determine a best maximum sample value, so if the range isn't 1.0, or the SMaxSampleValue tag isn't present, it may be necessary to (intelligently) use this parameter to properly read a file. intelligently. that's the keyword, bestie.

• tiff:min-sample-value=<value> — define this key and GraphicsMagick uses the assigned value as the minimum floating point value while reading or writing IEEE floating point TIFFs. Otherwise the min value is 0.0, or the value obtained from the file's SMinSampleValue tag (if present). 🌈
• tiff:bits-per-sample=<value> — define this key and GraphicsMagick writes images with the specified bits per sample, overriding any existing depth value. Value may be any in the range of 1 to 32, or 64 when the default unsigned format is written, or 16/32/24/64 if IEEEFP format is written. 🎯

⚠️ please note that the baseline TIFF 6.0 specification only requires readers to handle certain powers of two, and the values to be handled depend on the nature of the image (e.g. colormapped, grayscale, RGB, CMYK). so pick wisely, fr fr.

• tiff:samples-per-pixel=<value> — define this key to a value and the TIFF coder will write TIFF images with the defined samples per pixel, overriding any value stored in the image. This option should not normally be used. 🙈
• tiff:rows-per-strip=<value> — lets you specify the number of rows per TIFF strip. Rounded up to a multiple of 16 when using JPEG compression. Ignored when using tiles. 📏
• tiff:strip-per-page=true — requests that the image is written in a single TIFF strip. This is normally the default when group3 or group4 compression is requested within reasonable limits. 🎁

⚠️ requesting a single strip for large images may result in failure due to resource consumption in the writer or reader. big image + one strip = chaos, no cap 😭

• tiff:tile — enables writing tiled TIFF (rather than stripped) using the default tile size. Tiled TIFF organizes the image as an array of smaller images (tiles) in order to enable random access. 🧩
• tiff:tile-geometry=<width>x<height> — ok so this one's deets are literally coming up in the next breath, hold that thought bestie 💅
• tiff:tile-geometry=<width>x<height> — the tile size to use while writing tiled TIFF 👑

Width and height should be a multiple of 16 ✨ If the value is not a multiple of 16, then it will be rounded down. Enables tiled TIFF if it has not already been enabled.

Lowkey important: GraphicsMagick does not use tiled storage internally, so tiles need to be converted back and forth from the internal scanline-oriented storage to tile-oriented storage 🔁 Testing with typical RGB images shows that useful square tile size values range from 128x128 to 1024x1024. Large images which require using a disk-based pixel cache benefit from large tile sizes while images which fit in memory work well with smaller tile sizes 💅

• tiff:tile-width=<width> — the tile width to use while writing tiled TIFF 🎀

The tile height is then defaulted to an appropriate size. Width should be a multiple of 16. If the value is not a multiple of 16, then it will be rounded down. Enables tiled TIFF if it has not already been enabled.

• tiff:tile-height=<height> — the tile height to use while writing tiled TIFF 🦋

The tile width is then defaulted to an appropriate size. Height should be a multiple of 16. If the value is not a multiple of 16, then it will be rounded down. Enables tiled TIFF if it has not already been enabled.

• tiff:webp-lossless={TRUE|FALSE} — set this to TRUE to enable lossless mode while writing WebP-compressed TIFF files 🌸

The WebP webp:lossless option may also be used. The quality factor set by the -quality option may be used to influence the level of effort expended while compressing.

• tiff:zstd-compress-level=<value> — the compression level to use while writing Zstd-compressed TIFF files 🔥

The valid range is 1 to 22. If this define is not specified, then the 'quality' value is used such that the default quality setting of 75 is translated to a compress level of 9 such that ´quality' has a useful range of 10-184 if used for this purpose.

• webp:exact={true|false} — preserve exact RGB values under transparent areas 🙈 Default false.
• webp:lossless={true|false} — enable lossless encoding. Also enables 'exact' mode ✨
• webp:method={0-6} — quality/speed trade-off ⚖️
• webp:image-hint={default,graph,photo,picture} — hint for image type 🖼️
• webp:target-size=<integer> — target size in bytes 📏
• webp:target-psnr=<float> — minimal distortion to try to achieve.
• webp:segments={1-4} — maximum number of segments to use.
• webp:sns-strength={0-100} — Spatial Noise Shaping 🌊
• webp:filter-strength={0-100} — filter strength.
• webp:filter-sharpness={0-7} — filter sharpness.
• webp:filter-type={0,1} — filtering type. 0 = simple, 1 = strong (only used if filter-strength > 0 or autofilter is enabled).
• webp:auto-filter={true|false} — auto adjust filter's strength 🪄
• webp:alpha-compression=<integer> — algorithm for encoding the alpha plane (0 = none, 1 = compressed with WebP lossless). Default is 1.
• webp:alpha-filtering=<integer> — predictive filtering method for alpha plane. 0: none, 1: fast, 2: best. Default is 1.
• webp:alpha-quality={0-100} — between 0 (smallest size) and 100 (lossless). Default is 100.
• webp:pass=[1..10] — number of entropy-analysis passes.
• webp:show-compressed={true|false} — export the compressed picture back. In-loop filtering is not applied.
• webp:preprocessing=[0,1,2] — 0=none, 1=segment-smooth, 2=pseudo-random dithering 🎲
• webp:partitions=[0-3] — log2(number of token partitions) in [0..3]. Default is 0 for easier progressive decoding.
• webp:partition-limit={0-100} — quality degradation allowed to fit the 512k limit on prediction modes coding (0: no degradation, 100: maximum possible degradation).
• webp:emulate-jpeg-size={true|false} — if true, compression parameters will be remapped to better match the expected output size from JPEG compression. Generally, the output size will be similar but the degradation will be lower 💜
• webp:thread-level=<integer> — if non-zero, try and use multi-threaded encoding 🧵
• webp:low-memory={true|false} — if set, reduce memory usage (but increase CPU use)
• webp:use-sharp-yuv={true|false} — if set, if needed, use sharp (and slow) RGB->YUV conversion

Okay so for example, to create a postscript file that will render only the black pixels of a bilevel image, use this:

gm convert bilevel.tif -define ps:imagemask eps3:stencil.ps

---

• -delay <1/100ths of a second> — display the next image after pausing ⏳

This option is useful for regulating the animation of image sequences — Delay/100 seconds must expire before the display of the next image. The default is no delay between each showing of the image sequence. The maximum delay is 65535.

You can specify a delay range (e.g. -delay 10-500) which sets the minimum and maximum delay 🎬

• -density <width>x<height> — horizontal and vertical resolution in pixels of the image 📐

This option specifies the image resolution to store while encoding a raster image or the canvas resolution while rendering (reading) vector formats such as Postscript, PDF, WMF, and SVG into a raster image. Image resolution provides the unit of measure to apply when rendering to an output device or raster image. The default unit of measure is in dots per inch (DPI). The -units option may be used to select dots per centimeter instead.

The default resolution is 72 dots per inch, which is equivalent to one point per pixel (Macintosh and Postscript standard). Computer screens are normally 72 or 96 dots per inch while printers typically support 150, 300, 600, or 1200 dots per inch. To determine the resolution of your display, use a ruler to measure the width of your screen in inches, and divide by the number of horizontal pixels (1024 on a 1024x768 display) 📏 If the file format supports it, this option may be used to update the stored image resolution.

Heads up bestie 🚨 Photoshop stores and obtains image resolution from a proprietary embedded profile. If this profile is not stripped from the image, then Photoshop will continue to treat the image using its former resolution, ignoring the image resolution specified in the standard file header.

The density option is an attribute and does not alter the underlying raster image. It may be used to adjust the rendered size for desktop publishing purposes by adjusting the scale applied to the pixels. To resize the image so that it is the same size at a different resolution, use the -resample option 💅

• -depth <value> — depth of the image 🌈

This is the number of bits of color to preserve in the image. Any value between 1 and QuantumDepth (build option) may be specified, although 8 or 16 are the most common values. Use this option to specify the depth of raw images whose depth is unknown such as GRAY, RGB, or CMYK, or to change the depth of any image after it has been read 👑

See the -operator option for a way (via '-operator All Depth N') to apply a depth limit to the pixels immediately rather than setting an image attribute ✨

• -descend — obtain image by descending window hierarchy 🪟
• -despeckle — reduce the speckles within an image 🧹
• -displace <horizontal scale>x<vertical scale> — shift image pixels as defined by a displacement map 🗺️

With this option, composite image is used as a displacement map. Black, within the displacement map, is a maximum positive displacement. White is a maximum negative displacement and middle gray is neutral. The displacement is scaled to determine the pixel shift. By default, the displacement applies in both the horizontal and vertical directions. However, if you specify mask, composite image is the horizontal X displacement and mask the vertical Y displacement.

• -display <host:display[.screen]> — specifies the X server to contact 🖥️

This option is used with convert for obtaining image or font from this X server. See X(1).

• -dispose <method> — GIF disposal method 🗑️

The Disposal Method indicates the way in which the graphic is to be treated after being displayed.

Here are the valid methods:

• Undefined — No disposal specified.
• None — Do not dispose between frames.
• Background — Overwrite the image area with the background color.
• Previous — Overwrite the image area with what was there prior to rendering the image.

• -dissolve <percent> — dissolve an image into another by the given percent 💫

The opacity of the composite image is multiplied by the given percent, then it is composited over the main image.

• -dither — apply Floyd/Steinberg error diffusion to the image 🎨

The basic strategy of dithering is to trade intensity resolution for spatial resolution by averaging the intensities of several neighboring pixels. Images which suffer from severe contouring when reducing colors can be improved with this option.

The -colors or -monochrome option is required for this option to take effect ⚠️

Use +dither to turn off dithering and to render PostScript without text or graphic aliasing. Disabling dithering often (but not always) leads to decreased processing time ⚡

• -draw <string> — annotate an image with one or more graphic primitives 💅

bestie this is THE main character option. use it to annotate an image with one or more graphic primitives. the primitives come in four flavors: shapes, text, transformations, and pixel operations. here's the shape squad:

     point           x,y
     line            x0,y0 x1,y1
     rectangle       x0,y0 x1,y1
     roundRectangle  x0,y0 x1,y1 wc,hc
     arc             x0,y0 x1,y1 a0,a1
     ellipse         x0,y0 rx,ry a0,a1
     circle          x0,y0 x1,y1
     polyline        x0,y0  ...  xn,yn
     polygon         x0,y0  ...  xn,yn
     Bezier          x0,y0  ...  xn,yn
     path            path specification
     image           operator x0,y0 w,h filename

the text primitive is this one, and only this one:

     text            x0,y0 string

and the text gravity primitive is:

     gravity         NorthWest, North, NorthEast, West, Center,
                     East, SouthWest, South, or SouthEast

ok so the gravity thing? it only affects where text gets placed and it does not interact with the other primitives at all. it's basically the same as using the -gravity commandline option, except its scope is limited to the -draw option it shows up in. cute, contained, no drama ✨

now the transformation primitives:

     rotate          degrees
     translate       dx,dy
     scale           sx,sy
     skewX           degrees
     skewY           degrees

and the pixel operation primitives:

     color           x0,y0 method
     matte           x0,y0 method

so here's how the drawing goes down: shape primitives get drawn in the color from the -stroke option that came before them. except for line and point — those two don't get filled — everything else is filled with the color from the preceding -fill option. want an unfilled shape? use -fill none and slay 🎀

coordinate requirements, one by one:

• point needs a single coordinate
• line needs a start and an end coordinate
• rectangle expects an upper left and a lower right coordinate
• roundRectangle takes the upper left and lower right coordinates plus the width and height of the corners
• circle takes a center coordinate and a coordinate for the outer edge
• use arc to inscribe an elliptical arc inside a rectangle — arcs need a start and end point plus the degree of rotation (e.g. 130,30 200,100 45,90)
• use ellipse to draw a partial ellipse centered at the given point with the x-axis and y-axis radius and start and end of arc in degrees (e.g. 100,100 100,150 0,360)
• and finally, polyline and polygon need three or more coordinates to define their boundaries

coordinates are integers separated by an optional comma. for example, to define a circle centered at 100,100 that extends to 150,150, use:

     -draw 'circle 100,100 150,150'

Paths (peep See Paths) represent an outline of an object, and they're defined in terms of moveto (set a new current point), lineto (draw a straight line), curveto (draw a curve using a cubic Bezier), arc (elliptical or circular arc) and closepath (close the current shape by drawing a line to the last moveto) elements. compound paths are totally a thing — that's a path with subpaths, each one being a single moveto followed by one or more line or curve operations — and they let you do effects like "donut holes" in objects 🍩

use image to composite an image with another image. follow the image keyword with the composite operator, image location, image size, and filename:

     -draw 'image Over 100,100 225,225 image.jpg'

you can use 0,0 for the image size, which means "use the actual dimensions found in the image header." otherwise it gets scaled to the given dimensions. see -compose for a description of the composite operators 👑

use text to annotate an image with text. follow the text coordinates with a string. if the string has embedded spaces, enclose it in single or double quotes. optionally you can include the image filename, type, width, height, or other image attribute by embedding special format characters — see -comment for details. for example,

     -draw 'text 100,100 "%m:%f %wx%h"'

annotates the image with MIFF:bird.miff 512x480 for an image titled bird.miff and whose width is 512 and height is 480. iconic ✨

🚨 ok important one: if the first character of string is @, the text is read from a file titled by the remaining characters in the string. please note that if the string comes from an untrusted source, it should be sanitized before use — that's a security risk, fr fr. do not skip this.

rotate rotates subsequent shape primitives and text primitives about the origin of the main image. if the -region option precedes the -draw option, the origin for transformations is the upper left corner of the region. translate translates them. scale scales them. skewX and skewY skew them with respect to the origin of the main image or the region 💫

the transformations modify the current affine matrix, which gets initialized from the initial affine matrix defined by the -affine option. transformations are cumulative within the -draw option. the initial affine matrix is not affected — that matrix only changes when another -affine option shows up. and if another -draw option appears, the current affine matrix gets reinitialized from the initial affine matrix. no carryover, no baggage 💅

use color to change the color of a pixel to the fill color (see -fill). follow the pixel coordinate with a method:

     point
     replace
     floodfill
     filltoborder
     reset

consider the target pixel as the one specified by your coordinate. the point method recolors the target pixel. the replace method recolors any pixel that matches the color of the target pixel. floodfill recolors any pixel that matches the color of the target pixel and is a neighbor, whereas filltoborder recolors any neighbor pixel that is not the border color. and finally, reset recolors all pixels. the whole canvas. gone. reborn 🌈

use matte to change the pixel matte value to transparent. follow the pixel coordinate with a method (see the color primitive for a description of methods). the point method changes the matte value of the target pixel. the replace method changes the matte value of any pixel that matches the color of the target pixel. floodfill changes the matte value of any pixel that matches the color of the target pixel and is a neighbor, whereas filltoborder changes the matte value of any neighbor pixel that is not the border color (-bordercolor). finally, reset changes the matte value of all pixels 🫶

you can set the primitive color, font, and font bounding box color with -fill, -font, and -box respectively. and listen — options are processed in command line order, so be sure to use these options before the -draw option. order matters, bestie 😭

---

• -edge <radius> — detect edges within an image. that's it, that's the vibe 🦋
• -emboss <radius> — emboss an image
• -encoding <type> — specify the text encoding

choose from AdobeCustom, AdobeExpert, AdobeStandard, AppleRoman, BIG5, GB2312, Latin 2, None, SJIScode, Symbol, Unicode, Wansung.

• -endian <type> — specify endianness (MSB, LSB, or Native) of image

MSB indicates big-endian (e.g. SPARC, Motorola 68K) while LSB indicates little-endian (e.g. Intel 'x86, VAX) byte ordering. Native indicates to use the normal ordering for the current CPU. this option currently only influences the CMYK, DPX, GRAY, RGB, and TIFF formats.

use +endian to revert to unspecified endianness 💫

• -enhance — apply a digital filter to enhance a noisy image. glow-up for grainy pics ✨
• -equalize — perform histogram equalization to the image
• -extent <width>x<height>{+-}<x>{+-}<y> — composite image on background color canvas image

this option composites the image onto a new background color (-background) canvas image of size <width>x<height>. the existing image content is composited at the position specified by geometry x and y offset and/or desired gravity (-gravity) using the current image compose (-compose) method. image content that falls outside the bounds of the new image dimensions is discarded — yeeted, gone, bye 💀

for example, this command creates a thumbnail of an image and centers it on a red color backdrop image, offsetting the canvas ten pixels to the left and five pixels up, with respect to the thumbnail:

    gm convert infile.jpg -thumbnail 120x80 -background red -gravity center \
              -extent 140x100-10-5 outfile.jpg

and this command reduces or expands a JPEG image to fit on an 800x600 display:

    gm convert -size 800x600 input.jpg \
              -resize 800x600 -background black \
              -compose Copy -gravity center \
              -extent 800x600 \
              -quality 92 output.jpg

if the aspect ratio of the input image isn't exactly 4:3, then the image is centered on an 800x600 black canvas. no stretching, no weirdness, just vibes 🖤

• -file <filename> — write annotated difference image to file

if -file is specified, then an annotated difference image is generated and written to the specified file. pixels which differ between the reference and compare images are modified from those in the compare image so that the changed pixels become more obvious. some images may require use of an alternative highlight style (see -highlight-style) or highlight color (see -highlight-color) before the changes are obvious 👀

• -fill <color> — color to use when filling a graphic primitive

numeric color specifications in GraphicsMagick are similar to RGB and RGBA color values as used by SVG and CSS Color Module Level 3 (https://www.w3.org/TR/css-color-3/#numerical). use gm convert -list color to list named colors:

...and here's the full menu of color formats you can hand it, bestie 👇

name                              (named color)
#RGB                              (hex numbers, 4 bits each)
#RRGGBB                           (8 bits each)
#RRRGGGBBB                        (12 bits each)
#RRRRGGGGBBBB                     (16 bits each)
#RRRRRRRRGGGGGGGGBBBBBBBB         (32 bits each)
#RGBA                             (4 bits each)
#RRGGBBAA                         (8 bits each)
#RRRRRRGGGGGGBBBBBB               (12 bits each)
#RRRRRRRRGGGGGGGGBBBBBBBBAAAAAAAA (32 bits each)
rgb(r,g,b)                        (r,g,b are decimal numbers, range 0 to QuantumDepth)
rgb(r%,g%,b%)                     (r,g,b are decimal numbers, range 0 to 100%)
rgba(r,g,b,a)                     (r,g,b,a are decimal numbers, range 0 to QuantumDepth)
rgba(r%,g%,b%,a%)                 (r,g,b,a are decimal numbers, range 0 to 100%)

pro tip that will save your whole afternoon: wrap the color spec in quotation marks so your shell doesn't eat the # or the parentheses and leave you confused and humbled. fr fr.

Like, for example:

gm convert -fill blue ...
gm convert -fill "#ddddff" ...
gm convert -fill "rgb(65000,65000,65535)" ...

The shorter forms get scaled up, and if they need to grow they do it by replication. So #3af, #33aaff, and #3333aaaaffff are all literally the same color 💅

See -draw for further details.

---

✨ -filter ✨

-filter <type> — use this type of filter when resizing an image

Use this option to affect the resizing operation of an image (see -geometry). Pick from these filters, which are listed in order of approximately increasing CPU time:

Point
Box
Triangle
Hermite
Hanning
Hamming
Blackman
Gaussian
Quadratic
Cubic
Catrom
Mitchell
Lanczos
Bessel
Sinc

The default filter gets auto-picked so you get the best quality while still spending a reasonable amount of time. Here's the tea: the Mitchell filter is used if the image supports a palette, supports a matte channel, or is being enlarged — otherwise it's Lanczos. iykyk 🧚

---

✨ -flatten ✨

-flatten — flatten a sequence of images

Okay so some file formats (hello, Photoshop's PSD bestie) represent complex images as "layers" — a bunch of independent images that have to be composited to get the final look. The -flatten option is the one that does that composition. The sequence of images gets replaced by a single image, made by compositing each image in turn while respecting composition operators and page offsets. And yes, even though -flatten is immediately useful for killing layers, it's also lowkey a great general-purpose composition tool 🫶

The sequence of images ends the moment any option shows up. So if -flatten comes after all the input images, then all the images get flattened. Also go look at -mosaic, which is similar to -flatten except it adds a suitably-sized canvas base image.

For example, this composites an image on top of a 640x400 transparent black canvas image:

gm convert -size 640x300 xc:transparent \
          -compose over -page +0-100 \
          frame.png -flatten output.png

and this one flattens a Photoshop PSD file:

gm convert input.psd -flatten output.png

---

✨ -flip ✨

-flip — create a "mirror image"

It reflects the scanlines in the vertical direction. ⬆️⬇️

---

✨ -flop ✨

-flop — create a "mirror image"

It reflects the scanlines in the horizontal direction. ⬅️➡️

---

✨ -font ✨

-font <name> — use this font when annotating the image with text

You can tag a font so it knows whether it's a PostScript, TrueType, or X11 font. So like, Arial.ttf is a TrueType font, ps:helvetica is PostScript, and x:fixed is X11. tag your fonts, bestie 🎀

---

✨ -foreground ✨

-foreground <color> — define the foreground color

The color is specified using the format described under the -fill option. 🌈

---

✨ -format ✨

-format <type> — the image format type

When you use this with the mogrify utility, the option will convert any image to the image format you specify. See gm(1) for the list of image format types supported by GraphicsMagick, or just peek at the output of gm convert -list format.

By default the file is written to its original name. But! If the filename extension matches a supported format, that extension gets replaced with the image format type you gave -format. So if you specify tiff as the format type and the input image filename is image.gif, your output image filename becomes image.tiff. Slay 💅

---

✨ -format ✨

-format <string> — output formatted image characteristics

When you use this with the identify utility, or with the convert utility with the output written to the 'info:-' file specification, use this option to print info about the image in a format of your own choosing. You can include the image filename, type, width, height, Exif data, or other image attributes by embedding these special format characters:

%b   file size
%c   comment
%d   directory
%e   filename extension
%f   filename
%g   page dimensions and offsets
%h   height
%i   input filename
%k   number of unique colors
%l   label
%m   magick
%n   number of scenes
%o   output filename
%p   page number
%q   image bit depth
%r   image type description
%s   scene number
%t   top of filename
%u   unique temporary filename
%w   width
%x   horizontal resolution
%y   vertical resolution
%A   transparency supported
%C   compression type
%D   GIF disposal method
%G   Original width and height
%H   page height
%M   original filename specification
%O   page offset (x,y)
%P   page dimensions (width,height)
%Q   compression quality
%T   time delay (in centi-seconds)
%U   resolution units
%W   page width
%X   page horizontal offset (x)
%Y   page vertical offset (y)
%@   trim bounding box
%#   signature
\n   newline
\r   carriage return
%%   %

For example,

-format "%m:%f %wx%h"

displays MIFF:bird.miff 512x480 for an image titled bird.miff whose width is 512 and height is 480. iconic, no notes 🌟

If the first character of string is @, the format gets read from a file titled by the remaining characters in the string.

🚨 SECURITY VIBE CHECK: if the string comes from an untrusted source, sanitize it before use, because this can be used to incorporate any readable file on the system (that's a security risk, bestie, not a drill).

The values of image type (%r) that may be returned include:

Bilevel
Grayscale
GrayscaleMatte
Palette
PaletteMatte
TrueColor
TrueColorMatte
ColorSeparation
ColorSeparationMatte
Optimize

You can also use this special formatting syntax to print the Exif information contained in the file:

%[EXIF:<tag>]

Where "<tag>" may be one of the following:

*  (print all Exif tags, in keyword=data format)
!  (print all Exif tags, in tag_number format)
#hhhh (print data for Exif tag #hhhh)
ImageWidth
ImageLength
BitsPerSample
Compression
PhotometricInterpretation
FillOrder
DocumentName
ImageDescription
Make
Model
StripOffsets
Orientation
SamplesPerPixel
RowsPerStrip
StripByteCounts
XResolution
YResolution
PlanarConfiguration
ResolutionUnit
TransferFunction
Software
DateTime
Artist
WhitePoint
PrimaryChromaticities
TransferRange
JPEGProc
JPEGInterchangeFormat
JPEGInterchangeFormatLength
YCbCrCoefficients
YCbCrSubSampling
YCbCrPositioning
ReferenceBlackWhite
CFARepeatPatternDim
CFAPattern
BatteryLevel
Copyright
ExposureTime
FNumber
IPTC/NAA
ExifOffset
InterColorProfile
ExposureProgram
SpectralSensitivity
GPSInfo
ISOSpeedRatings
OECF
ExifVersion
DateTimeOriginal
DateTimeDigitized
ComponentsConfiguration
CompressedBitsPerPixel
ShutterSpeedValue
ApertureValue
BrightnessValue
ExposureBiasValue
MaxApertureValue
SubjectDistance
MeteringMode
LightSource
Flash
FocalLength
MakerNote
UserComment
SubSecTime
SubSecTimeOriginal
SubSecTimeDigitized
FlashPixVersion
ColorSpace
ExifImageWidth
ExifImageLength
InteroperabilityOffset
FlashEnergy
SpatialFrequencyResponse
FocalPlaneXResolution
FocalPlaneYResolution
FocalPlaneResolutionUnit
SubjectLocation
ExposureIndex
SensingMethod
FileSource
SceneType

JPEG specific information (from reading a JPEG file) may be obtained like this:

%[JPEG-<tag>]

Where <tag> can be literally any of these cuties 👇

• * — every single JPEG-related tag, served up in keyword=data format
• Quality — the IJG JPEG "quality" estimate
• Colorspace — JPEG colorspace numeric ID
• Colorspace-Name — JPEG colorspace name
• Sampling-factors — JPEG sampling factors

Okay so real talk for a sec 🫶 — JPEG genuinely has no concept of "quality", like none, it's not a thing it knows about. The quality metric this software uses and estimates is based on the quality metric established by IJG JPEG 6b. Other encoders (e.g. the one Adobe Photoshop uses) run on different encoding metrics. So don't come at me comparing numbers across encoders, they're not the same scale, iykyk 💅

Wrap that format specification in quotation marks so your shell doesn't get any ideas about the spaces and square brackets, bestie. Trust. 🙈

---

• -frame <width>x<height>+<outer bevel width>+<inner bevel width> — surround the image with an ornamental border 🎀

Peep -geometry for the deets on the geometry specification. Also important: -frame is not affected by the -gravity option. Not even a little.

The color of the border comes from the -mattecolor command line option. That's the one that decides the vibe ✨

• -frame — include the X window frame in the imported image

(yes, there are two -frame entries on this page, that's just how it is, we move 💀)

• -fuzz <distance>{%} — colors within this Euclidean distance are considered equal 🌈

A bunch of algorithms go hunting for a target color, and by default that color has to be exact. This option is how you match colors that are merely close (in Euclidean distance) to the target color in RGB 3D space. The classic scenario: you wanna auto-trim the edges of an image with -trim, but the thing was scanned and the target background color might differ by a tiny amount. -fuzz accounts for those differences, slay 💖

The distance can be in absolute intensity units, or — by appending % — as a percentage of the maximum possible intensity (255, 65535, or 4294967295).

• -gamma <value> — level of gamma correction ⭐

Same color image, two different workstations, and it may look different because the display monitors just aren't the same. Gamma correction is how you adjust for that color difference. Reasonable values extend from 0.8 to 2.3. Gamma less than 1.0 darkens the image, gamma greater than 1.0 lightens it. Heads up though: large adjustments to image gamma may result in the loss of some image information if the pixel quantum size is only eight bits (quantum range 0 to 255) 😭

You can apply separate gamma values to the red, green, and blue channels with a gamma value list delimited with slashes (e.g., 1.7/2.3/1.2).

Use +gamma value to set the image gamma level without actually adjusting the image pixels. Useful if the image is of a known gamma but not set as an image attribute (e.g. PNG images). No cap, that's the move 👑

• -gaussian <radius>{x<sigma>} — blur the image with a Gaussian operator 🦋

Use the given radius and standard deviation (sigma). That's it, that's the whole thing.

• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — specify dimension, offset, and resize options 🎀

Okay so -geometry is the main character of options — it does a lot of different jobs depending on which utility you're using it with. Buckle up.

For the X11 commands (animate, display, and import), it specifies the preferred size and location of the Image window. By default, the window size is the image size and the location is chosen by you (or your window manager) when it is mapped. For the import, convert, mogrify utility commands it may be used to specify the desired size when resizing an image. In this case, symbols representing resize options may be appended to the geometry string to influence how the resize request is treated.

See later notes corresponding to usage by particular commands. The following notes apply to when -geometry is used to express a resize request, taking into account the current properties of the image.

By default, width and height are maximum values. That is, the image is expanded or contracted to fit the width and height value while maintaining the aspect ratio of the image.

Append a ^ to the geometry so the image aspect ratio is maintained when the image is resized, but the resulting width or height are treated as minimum values rather than maximum values.

Append a ! (exclamation point) to the geometry to force the image size to exactly the size you specify. For example, if you specify 640x480! the image width is set to 640 pixels and height to 480, changing the image aspect ratio if necessary. No negotiation 💅

If only the width is specified, without the trailing x, then height is set to width (e.g., -geometry 100 is the same as -geometry 100x100). If only the width is specified but with the trailing x, then width assumes the value and the height is chosen to maintain the aspect ratio of the image. Similarly, if only the height is specified prefixed by x (e.g., -geometry x256), the width is chosen to maintain the aspect ratio.

To specify a percentage width or height instead, append %. The image size is multiplied by the width and height percentages to obtain the final image dimensions. To increase the size of an image, use a value greater than 100 (e.g. 125%). To decrease an image's size, use a percentage less than 100.

Use @ to specify the maximum area in pixels of an image.

Use > to change the dimensions of the image only if its width or height exceeds the geometry specification. < resizes the image only if both of its dimensions are less than the geometry specification. For example, if you specify 640x480> and the image size is 256x256, the image size does not change. However, if the image is 512x512 or 1024x1024, it is resized to 480x480. Enclose the geometry specification in quotation marks so your shell doesn't interpret the < or > as a file redirection — that's a whole different kind of disaster 😭

When used with animate and display, offsets are handled in the same manner as in X(1) and the -gravity option is not used. If the x is negative, the offset is measured leftward from the right edge of the screen to the right edge of the image being displayed. Similarly, negative y is measured between the bottom edges. The offsets are not affected by %; they are always measured in pixels.

When used as a composite option, -geometry gives the dimensions of the image and its location with respect to the composite image. If the -gravity option is present with NorthEast, East, or SouthEast gravity, the x represents the distance from the right edge of the image to the right edge of the composite image. Similarly, if the -gravity option is present with SouthWest, South, or SouthEast gravity, y is measured between the bottom edges. Accordingly, a positive offset will never point in the direction outside of the image. The offsets are not affected by %; they are always measured in pixels. To specify the dimensions of the composite image, use the -resize option.

When used as a convert, import or mogrify option, -geometry is synonymous with -resize and specifies the size of the output image. The offsets, if present, are ignored. Bye offsets 👋

When used as a montage option, -geometry specifies the image size and border size for each tile; default is 256x256+0+0. Negative offsets (border dimensions) are meaningless. The -gravity option affects the placement of the image within the tile; the default gravity for this purpose is Center. If the % sign appears in the geometry specification, the tile size is the specified percentage of the original dimensions of the first tile. To specify the dimensions of the montage, use the -resize option.

• -gravity <type> — direction primitive gravitates to when annotating the image 🧚

Choices are: NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast.

The direction you choose specifies where to position the text when annotating the image. For example Center gravity forces the text to be centered within the image. By default, the image gravity is NorthWest. See -draw for more details about graphic primitives. Only the text primitive is affected by the -gravity option.

The -gravity option also works in concert with the -geometry option and other options that take <geometry> as a parameter, such as the -crop option. See -geometry for details of how -gravity interacts with the <x> and <y> parameters of a geometry specification.

When used as an option to composite, -gravity gives the direction that the image gravitates within the composite.

When used as an option to montage, -gravity gives the direction that an image gravitates within a tile. The default gravity for this purpose is Center.

• -green-primary <x>,<y> — green chromaticity primary point 💚

Short and sweet. That's the whole entry, bestie.

• -hald-clut <clut> — apply a Hald CLUT to the image 🎨

A Hald CLUT ("Color Look-Up Table") is a special square color image which contains a look-up table for red, green, and blue. The size of the Hald CLUT image is determined by its order. The width (and height) of a Hald CLUT is the cube of the order. For example, a Hald CLUT of order 8 is 512x512 pixels (262,144 colors) and of order 16 is 4096x4096 (16,777,216 colors). A special CLUT is the identity CLUT, which causes no change to the input image. In order to use the Hald CLUT, one takes an identity CLUT and adjusts its colors in some way. The modified CLUT can then be used to transform any number of images in an identical way. That's the whole flex of it, fr fr 🔥

GraphicsMagick contains a built-in identity CLUT generator via the IDENTITY coder. For example reading from the file name IDENTITY:8 returns an identity CLUT of order 8. Typical Hald CLUT identity images have an order of between 8 and 16. The default order for the IDENTITY CLUT generator is 8. Interpolation is used, so it's not usually necessary for CLUT images to be very large. The PNG file format is ideal for storing Hald CLUT images because it compresses them very well 🦋

• -help — prints the usage instructions, bestie. That's it, that's the whole vibe 💅
• -highlight-color <color> — the pixel annotation color ✨ Tells gm what color to paint with when it's annotating difference pixels (you know, the changed ones).
• -highlight-style <style> — the pixel annotation style 🎨 Picks how changed pixels get flagged so your eyes actually notice them. It's one of Assign, Threshold, Tint, or XOR. Quick rundown: Assign swaps the pixel out for the highlight color (that's the -highlight-color one), Threshold swaps the pixel for black or white depending on the intensity difference, Tint alpha-tints the pixel with the highlight color, and XOR does an XOR between the pixel and the highlight color. Pick your fighter 🫶
• -iconGeometry <geometry> — specify the icon geometry 🪟 If there are offsets in the geometry spec, they get handled the exact same way as the -geometry option does, using X11 style to deal with negative offsets.
• -iconic — iconic animation 👑
• -immutable — makes the image immutable, no cap 🔒
• -implode <factor> — implodes image pixels about the center 🌀
• -intent <type> — use this type of rendering intent when managing the image color 🌈 Use this to affect the color management operation of an image (peep -profile for the context). Your choices are: Absolute, Perceptual, Relative, Saturation.

The default intent is undefined. (Mysterious. Main character energy, honestly.)

• -interlace <type> — the type of interlacing scheme 🧚 Choices are: None, Line, Plane, or Partition. The default is None.

This one's for specifying the interlacing scheme on raw image formats like RGB or YUV. None means don't interlace at all, so it just goes RGBRGBRGBRGBRGBRGB...,

Line uses scanline interlacing (RRR...GGG...BBB...RRR...GGG...BBB...), and Plane uses plane interlacing (RRRRRR...GGGGGG...BBBBBB...).

Partition is like plane except the different planes get saved into individual files (like image.R, image.G, and image.B, slay).

Use Line to create an interlaced PNG or GIF or progressive JPEG image.

• -label <name> — assign a label to an image 🏷️ Use this option to assign a specific label to the image when you're writing to an image format that supports labels, such as TIFF, PNG, MIFF, or PostScript. You can include the image filename, type, width, height, or other image attribute by embedding special format characters. A label is not drawn on the image — it gets embedded in the image datastream via a "Label" tag or similar mechanism. If you want the label actually visible on the image itself, use the -draw option instead. See -comment for details.

For example,

-label "%m:%f %wx%h"

produces an image label of MIFF:bird.miff 512x480 for an image titled bird.miff whose width is 512 and height is 480. Iconic.

If the first character of the string is @, the image label is read from a file titled by the remaining characters in the string.

Ok so real talk: if the string comes from an untrusted source, sanitize it before use, because otherwise the content of an arbitrary readable file might get incorporated into the image label, and that's a security risk. Not the cute kind of surprise. 😭

If the -label option appears multiple times, only the last label is stored. (Last one wins, periodt.)

In PNG images, the label is stored in a tEXt or zTXt chunk with the keyword "label".

When converting to PostScript, use this option to specify a header string to print above the image. Specify the label font with -font.

When creating a montage, by default the label associated with an image is displayed with the corresponding tile in the montage. Use the +label option to suppress this behavior.

• -lat <width>x<height>{+-}<offset>{%} — perform local adaptive thresholding 🔍 Performs local adaptive thresholding using the specified width, height, and offset. The offset is a distance in sample space from the mean, as an absolute integer ranging from 0 to the maximum sample value, or as a percentage. If the percent option is supplied, then the offset is computed as a percentage of the quantum range.

It is strongly recommended to use the percent option so that results are not sensitive to pixel quantum depth. Fr fr, do it.

For example,

-colorspace gray -lat "10x10-5%"

will help clarify a scanned grayscale or color document, producing a bi-level equivalent. Chef's kiss 💋

• -level <black_point>{,<gamma>}{,<white_point>}{%} — adjust the level of image contrast 📊 Give one, two or three values delimited with commas: black-point, gamma, white-point (e.g. 10,1.0,250 or 2%,0.5,98%). The black and white points range from 0 to MaxRGB or from 0 to 100%; if the white point is omitted it's set to MaxRGB-black_point. If a "%" sign is present anywhere in the string, the black and white points are percentages of MaxRGB. Gamma is an exponent that ranges from 0.1 to 10.; if it's omitted, the default of 1.0 (no gamma correction) is assumed. This interface works similar to Photoshop's "Image->Adjustments->Levels..." "Input Levels" interface 🎚️
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit 🚦 By default, resource limits are estimated based on the available resources and capabilities of the system. Here's the whole squad:
  • Disk — maximum total disk space consumed
  • File — maximum number of file descriptors allowed to be open at once
  • Images — maximum images allowed to be handled by the whole program at one time
  • Map — maximum total number of file bytes which may be memory mapped
  • Memory — maximum total number of bytes of heap memory used for image storage
  • Pixels — maximum absolute image size (per image)
  • Width — maximum image pixels width
  • Height — maximum image pixels height
  • Read — maximum number of uncompressed bytes to read
  • Write — maximum number of uncompressed bytes to write
  • Threads — the maximum number of worker threads to use per OpenMP thread team

The Disk and Map resource limits are what decide if (for a given image) the decoded image ("pixel cache") should be stored in heap memory (RAM), in a memory-mapped disk file, or in a disk file accessed via read/write I/O.

The number of total pixels in one image (Pixels), and/or the width/height (Width/Height), may be limited in order to force the reading, or creation, of images larger than the limit (in pixels) to intentionally fail. The disk limit (Disk) establishes an overall limit since using the disk is the means of last resort. When the disk limit has been reached, no more images may be read. That's the final boss, basically.

The amount of uncompressed data read when reading one image file may be limited by the Read limit. Reading the image fails when the limit is hit. This option is useful if the data is read from a stream (pipe) or from a compressed file such as a gzipped file — some files are very compressible, so a small compressed file can decompress to a huge amount of data. This option also defends against files which produce seemingly endless loops while decoding by seeking backwards in the file. 💀

The amount of uncompressed data written when writing one image file may be limited by the Write limit. Writing the image fails (similar to disk full) when the limit is hit. If the output is to an ordinary file (rather than a pipe) then any truncated output file is removed. This option may be used to assure that unexpectedly large files are not written. It may also be used to simulate what will happen if the disk is full. Very "let's test the chaos" of it.

The value argument is an absolute value, but may have standard binary suffix characters applied ('K', 'M', 'G', 'T', 'P', 'E') to apply a scaling to the value (based on a multiplier of 1024). Any additional characters are ignored. For example, -limit Pixels 10MP limits the maximum image size to 10 megapixels and -limit memory 32MB -limit map 64MB limits memory and memory mapped files to 32 megabytes and 64 megabytes respectively.

Resource limits may also be set using environment variables, bestie. MAGICK_LIMIT_DISK, MAGICK_LIMIT_FILES, MAGICK_LIMIT_MAP, MAGICK_LIMIT_MEMORY, MAGICK_LIMIT_PIXELS, MAGICK_LIMIT_WIDTH, MAGICK_LIMIT_HEIGHT. MAGICK_LIMIT_READ, MAGICK_LIMIT_WRITE, and OMP_NUM_THREADS may be used to set the limits for disk space, open files, memory mapped size, heap memory, per-image pixels, image width, image height, bytes read from a file, bytes written to a file, and OpenMP threads respectively.

Use the option -list resource to list the current limits.

• -linewidth — the line width for subsequent draw operations ✏️
• -list <type> — the type of list 📋 Choices are: Color, Delegate, Format, Magic, Module, Resource, or Type. The Module option is only available if GraphicsMagick was built to support loadable modules.

This option lists information about the GraphicsMagick configuration.

• -log <string> — specify format for debug log 📝 This option specifies the format for the log printed when the -debug option is active.

You can display the following components by embedding special format characters:

%d   domain
%e   event
%f   function
%l   line
%m   module
%p   process ID
%r   real CPU time
%t   wall clock time
%u   user CPU time
%%   percent sign
\n   newline
\r   carriage return

For example:

gm convert -debug coders -log "%u %m:%l %e" in.gif out.png

The default behavior is to print all of the components. All of them. Every single one 💅

• -loop <iterations> — add Netscape loop extension to your GIF animation 🔁 A value other than zero forces the animation to repeat itself up to iterations times.
• -magnify — magnify the image 🔍 The image size is doubled using linear interpolation.
• -magnify <factor> — magnify the image 🔍 The displayed image is magnified by factor. (Yes, two -magnify entries, one takes a factor and one doesn't — gm said "why not both" 🫶)
• -map <filename> — pick a specific color palette straight out of this image file. [convert or mogrify]

Okay so by default, color reduction just vibes and picks an optimal set of colors that best represents the original image. But bestie, sometimes you want the exact palette from another image, and that's literally what this option is for. 💅

Use +map to reduce all images in the image sequence that follows down to a single optimal set of colors that best represents all of them. The sequence of images ends the moment any option shows up. And if +map appears after all the input images? Then all images get mapped. periodt.

• -map <type> — display the image using this type. [animate or display]

Choose from these Standard Colormap types:

• best
• default
• gray
• red
• green
• blue

The X server has to actually support the Standard Colormap you choose, otherwise you get an error — no cap, it will just fail on you. 😭 Use list as the type and display will search the list of colormap types in top-to-bottom order until it finds one that works. See xstdcmap(1) for one way of creating Standard Colormaps. iykyk ✨

• -mask <filename> — specify a clipping mask 🎭

The image read from the file is used as a clipping mask. It must have the same dimensions as the image being masked — that's non-negotiable, fr fr.

If the mask image has an opacity channel, the opacity of each pixel defines the mask. Otherwise, the intensity (gray level) of each pixel is used instead. Unmasked (black) pixels get modified, while masked pixels (anything not black) are protected from alteration. literally a bodyguard for your pixels. 🫶

Use +mask to remove the clipping mask.

You do not need to use -clip to activate the mask — -clip is implied by -mask. it's giving automatic.

• -matte — store the matte channel if the image has one

If the image doesn't have a matte channel, create an opaque one. 🎀

Use +matte to ignore the matte channel (treats it as opaque) and to avoid writing a matte channel in the output file.

For the compare command, -matte will add an opaque matte channel to images if they don't already have one, and matte will be enabled for both images. Likewise, if +matte is used, the matte channel is disabled for both images. This makes it easier to compare images regardless of whether they already have a matte channel — such a slay, honestly. ✨

• -mattecolor <color> — specify the color used with the -frame option

The color is specified using the format described under the -fill option. 💖

• -maximum-error <limit> — specifies the maximum amount of total image error

Specifies the maximum amount of total image error (based on comparison using a specified metric) before an error — "image difference exceeds limit" — gets reported. The error is reported via a non-zero command execution return status. so basically it's the vibe check for your images and it will yell at you if you fail it. 💀

• -median <radius> — apply a median filter to the image
• -metric <metric> — comparison metric (MAE, MSE, PAE, PSNR, RMSE)
• -minify <factor> — minify the image

The image size is halved using linear interpolation. tiny image, big energy. 🦋

• -mode <value> — mode of operation

The available montage modes are frame to place the images in a rectangular grid while adding a decorative frame with dropshadow, unframe to place undecorated images in a rectangular grid, and concatenate to pack the images closely together without any well-defined grid or decoration. pick your aesthetic bestie 🎀

• -modulate brightness[,saturation[,hue]] — vary the brightness, saturation, and hue of an image

Specify the percent change in brightness, color saturation, and hue, separated by commas. Default argument values are 100 percent, which results in no change. For example, to increase the color brightness by 20% and decrease the color saturation by 10% and leave the hue unchanged, use: -modulate 120,90.

Hue is the percentage of absolute rotation from the current position. For example 50 results in a counter-clockwise rotation of 90 degrees, 150 results in a clockwise rotation of 90 degrees, with 0 and 200 both resulting in a rotation of 180 degrees. the math is mathing 🌈

• -monitor — show progress indication

A simple command-line progress indication is shown while the command is running. The process indication shows the operation currently being performed and the percent completed. Commands using X11 may replace the command line progress indication with a graphical one once an image has been displayed. satisfying little progress bar moment 🫶

• -monochrome — transform the image to black and white
• -morph <frames> — morphs an image sequence

Both the image pixels and size are linearly interpolated to give the appearance of a meta-morphosis from one image to the next. absolute main character transformation arc ✨

The sequence of images is terminated by the appearance of any option. If the -morph option appears after all of the input images, all images are morphed.

• -mosaic — create a mosaic from an image or an image sequence

The -mosaic option provides a flexible way to composite one or more images onto a solid-color canvas image. It works similar to -flatten except that a base canvas image is automatically created with a suitable size given the image size, page dimensions, and page offsets of images to be composited. The color of the base canvas image may be set via the -background option. The default canvas color is 'white', but 'black' or 'transparent' may be more suitable depending on the composition algorithm requested. ngl this one eats 💅

The -compose option may be used to specify the composition algorithm to use when compositing the subsequent image on the base canvas.

The -page option can be used to establish the dimensions of the mosaic and to position the subsequent image within the mosaic. If the -page argument does not specify width and height, then the canvas dimensions are evaluated based on the image sizes and offsets.

The sequence of images is terminated by the appearance of any option. If the -mosaic option appears after all of the input images, all images are included in the mosaic.

The following is an example of composing an image based on red, green, and blue layers extracted from a sequence of images and pasted on the canvas image at specified offsets:

gm convert -background black \
          -compose CopyRed   -page +0-100 red.png \
          -compose CopyGreen -page +0+40  green.png \
          -compose CopyBlue  -page +0+180 blue.png \
          -mosaic output.png

• -motion-blur <radius>{x<sigma>}{+angle} — Simulate motion blur

Simulate motion blur by convolving the image with a Gaussian operator of the given radius and standard deviation (sigma). For reasonable results, radius should be larger than sigma. If radius is zero, then a suitable radius is automatically selected based on sigma. The angle specifies the angle that the object is coming from (the side which is blurred). blurry but make it fashion 💨

• -name — name an image
• -negate — replace every pixel with its complementary color

The red, green, and blue intensities of an image are negated. White becomes black, yellow becomes blue, etc. Use +negate to only negate the grayscale pixels of the image. 🖤🤍

• -noise <radius|type> — add or reduce noise in an image

The principal function of the noise peak elimination filter is to smooth the objects within an image without losing edge information and without creating undesired structures. The central idea of the algorithm is to replace a pixel with its next neighbor in value within a pixel window, if this pixel has been found to be noise. A pixel is defined as noise if and only if this pixel is a maximum or minimum within the pixel window.

Use radius to specify the width of the neighborhood.

Use +noise followed by a noise type to add noise to an image. The noise added modulates the existing image pixels. Choose from these noise types:

• Uniform
• Gaussian
• Multiplicative
• Impulse
• Laplacian
• Poisson
• Random (uniform distribution)

chaotic on purpose, we love to see it 🙈

• -noop — NOOP (no option)

The -noop option can be used to terminate a group of images and reset all options to their default values, when no other option is desired. it's literally the "do nothing, slay" option 💅

• -normalize — transform image to span the full range of color values

This is a contrast enhancement technique based on the image histogram. 🔥

When computing the contrast enhancement values, the histogram edges are truncated so that the majority of the image pixels are considered in the contrast enhancement, and outliers (e.g. random noise or minute details) are ignored. The default is that 0.1 percent of the histogram entries are ignored. The percentage of the histogram to ignore may be specified by using the -set option with the histogram-threshold parameter, similar to -set histogram-threshold 0.01 to specify 0.01 percent. Use 0 percent to use the entire histogram, with possibly diminished contrast enhancement. (yes bestie, "possibly diminished" — that's a real warning, don't come for me 😭)

• -opaque <color> — change this color to the pen color within the image

The color is specified using the format described under the -fill option. The color is replaced if it is identical to the target color, or close enough to the target color in a 3D space as defined by the Euclidean distance specified by -fuzz. geometry AND color theory, we contain multitudes 🦋

See -fill and -fuzz for more details.

• -operator channel operator rvalue[%] — apply a mathematical, bitwise, or value operator to an image channel

Apply a low-level mathematical, bitwise, or value operator to a selected image channel or all image channels. Operations which result in negative results are reset to zero, and operations which overflow the available range are reset to the maximum possible value. no underflowing, no overflowing, just boundaries. ⭐

So for the channel part of -operator, you've got these to pick from: Red, Green, Blue, Opacity, Matte, Cyan, Magenta, Yellow, Black, All, or Gray. Heads up bestie: All only touches the color channels and stays out of the Opacity channel's business. And except for the threshold operators, All operates on each channel independently, so it's all per-channel vibes.

Gray is the ✨aesthetic✨ one — it treats the color channels as a grayscale intensity and runs the operation you asked for on the equivalent pixel intensity, so the result is a gray image. iconic behavior tbh

And the operator menu? Take a deep breath:

Add, And, Assign, Depth, Divide, Gamma, Negate, LShift, Log, Max, Min, Multiply, Or, Pow, RShift, Subtract, Threshold, Threshold-White, Threshold-White-Negate, Threshold-Black, Threshold-Black-Negate, Xor, Noise-Gaussian, Noise-Impulse, Noise-Laplacian, Noise-Multiplicative, Noise-Poisson, Noise-Random, and Noise-Uniform.

---

Now, rvalue. It can be any floating point or integer value, fr fr. Normally rvalue lives in the range of 0 to MaxRGB, where MaxRGB is the largest quantum value your GraphicsMagick build supports — that's 255, 65535, or 4294967295 — but values outside that range are actually useful for some arithmetic operations, so don't be scared. 🫶

Arguments to logical or bit-wise operations get rounded to a positive integral value before they're used. And if you slap a percent (%) symbol on the end of the argument, then the argument has a range of 0 to 100 percent. cute, right?

Here's the whole operator roster, one by one:

💖 Add

Result is rvalue added to channel value.

💖 And

Result is the logical AND of rvalue with channel value.

💖 Assign

Result is rvalue. That's it. That's the whole thing.

💖 Depth

Result is channel value adjusted so that it may be (approximately) stored in the specified number of bits without additional loss.

💖 Divide

Result is channel value divided by rvalue.

💖 Gamma

Result is channel value gamma adjusted by rvalue.

💖 LShift

Result is channel value bitwise left shifted by rvalue bits.

💖 Log

Result is computed as log(value*rvalue+1)/log(rvalue+1).

💖 Max

Result is assigned to rvalue if rvalue is greater than value.

💖 Min

Result is assigned to rvalue if rvalue is less than value.

💖 Multiply

Result is channel value multiplied by rvalue.

💖 Negate

Result is the inverse of channel value (like a film negative 🎞️). An rvalue must be supplied but is currently not used. Inverting the image twice results in the original image — we stay symmetrical.

💖 Or

Result is the logical OR of rvalue with channel value.

💖 Pow

Result is computed as pow(value,rvalue). Similar to Gamma except that rvalue is not inverted.

💖 RShift

Result is channel value bitwise right shifted by rvalue bits.

💖 Subtract

Result is channel value minus rvalue.

💖 Threshold

Result is maximum (white) if channel value is greater than rvalue, or minimum (black) if it is less than or equal to rvalue. If all channels are specified, then thresholding is done based on computed pixel intensity.

💖 Threshold-white

Result is maximum (white) if channel value is greater than rvalue and is unchanged if it is less than or equal to rvalue. This can be used to remove apparent noise from the bright parts of an image. If all channels are specified, then thresholding is done based on computed pixel intensity.

💖 Threshold-White-Negate

Result is set to black if channel value is greater than rvalue and is unchanged if it is less than or equal to rvalue. If all channels are specified, then thresholding is done based on computed pixel intensity.

💖 Threshold-black

Result is minimum (black) if channel value is less than rvalue and is unchanged if it is greater than or equal to rvalue. This can be used to remove apparent noise from the dark parts of an image. If all channels are specified, then thresholding is done based on computed pixel intensity.

💖 Threshold-Black-Negate

Result is set to white if channel value is less than rvalue and is unchanged if it is greater than or equal to rvalue. If all channels are specified, then thresholding is done based on computed pixel intensity.

💖 Xor

Result is the logical XOR of rvalue with channel value. Fun little property of XOR: performing the same operation twice gives you back the original value. reversible king 👑

💖 Noise-Gaussian

Result is the current channel value modulated with gaussian noise according to the intensity specified by rvalue.

💖 Noise-Impulse

Result is the current channel value modulated with impulse noise according to the intensity specified by rvalue.

💖 Noise-Laplacian

Result is the current channel value modulated with laplacian noise according to the intensity specified by rvalue.

💖 Noise-Multiplicative

Result is the current channel value modulated with multiplicative gaussian noise according to the intensity specified by rvalue.

💖 Noise-Poisson

Result is the current channel value modulated with poisson noise according to the intensity specified by rvalue.

💖 Noise-Random

Result is the current channel value modulated with random (uniform distribution) noise according to the intensity specified by rvalue. The initial noise intensity (rvalue=1.0) is the range of one pixel quantum span.

💖 Noise-Uniform

Result is the channel value with uniform noise applied according to the intensity specified by rvalue.

---

Okay so like, a concrete example of all this — the Assign operator assigns a fixed value to a channel. For example, this command sets the red channel to the mid-range value:

gm convert in.bmp -operator red assign "50%" out.bmp

And this one applies 50% thresholding to the image and returns a gray image:

gm convert in.bmp -operator gray threshold "50%" out.bmp

go off, tiny command lines ✨

---

• -ordered-dither <channeltype> <NxN> — ordered dither the image

Whichever channel or channels you name in the channeltype argument get reduced to binary using an ordered dither method. Your channeltype choices are All, Intensity, Red, Green, Blue, Cyan, Magenta, Yellow, Black, and Opacity.

When channeltype is "All", the color samples are dithered into a gray level and then that gray level is stored in the three color channels. Separately, the opacity channel is dithered into a bilevel opacity value which is stored in the opacity channel. two jobs, one flag

When channeltype is "Intensity", only the color samples are dithered. When channeltype is "opacity" or "matte", only the opacity channel is dithered. When a color channel is specified, only that channel is dithered. Simple enough 💅

The choices for N are 2 through 7. The image gets divided into NxN pixel tiles, and in each tile, some or all pixels are turned to white depending on their intensity. For each N, (N**2)+1 levels of gray can be represented. For N == 2, 3, or 4, the pixels are turned to white in an order that maximizes dispersion (i.e., reduces granularity), while for N == 5, 6, and 7, they're turned to white in an order that creates a roughly circular black blob in the middle of each tile. blob era 🖤

Want that attractive "half-tone" look? First rotate the image 45 degrees, perform a 5x5 ordered-dither operation, then rotate it back to the original orientation and crop to the original image dimensions. And if the original image is gamma-encoded, it's advisable to convert it to linear intensity first, e.g., with the -gamma 0.45455 option. pro tip, but from the manual, not from me

• -output-directory <directory> — output files to directory

Use -output-directory to specify a directory under which to write the output files. Normally mogrify overwrites the input files, but with this option the output files may be written to a different directory tree so that the input files are preserved. no more accidental deletions, bestie 😭

The algorithm used preserves all of the input path specification in the output path, so that the user-specified input path (including any sub-directory part) is appended to the output path. If the input file lacks an extension, then a suitable extension is automatically added to the output file. The user is responsible for creating the output directory specified as an argument, but subdirectories will be created as needed if the -create-directories option is supplied.

This option may be used to apply transformations on files from one directory and write the transformed files to a different directory. In conjunction with -create-directories, this option is designed to support transforming whole directory trees of files — provided that the relative path of the input file is included as part the list of filenames. 🗂️

• -orient <orientation> — Set the image orientation attribute

Sets the image orientation attribute, which is compatible with the TIFF orientation tag (and the EXIF orientation tag, iykyk). Accepted values are undefined, TopLeft, TopRight, BottomRight, BottomLeft, LeftTop, RightTop, RightBottom, LeftBottom, and hyphenated versions thereof (e.g. left-bottom).

Please note that GraphicsMagick does not include an EXIF editor, so if an EXIF profile is written to the output image, the value in the EXIF profile might not match the image. It is possible for an image file to indicate its orientation in several different ways simultaneously. messy, but that's the situation 🙈

• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas

so here's the full menu of page sizes you can name by hand 🎀 — dots per inch if it's a PostScript page, pixels if it's a TEXT page:

     11x17         792  1224
     Ledger       1224   792
     Legal         612  1008
     Letter        612   792
     LetterSmall   612   792
     ArchE        2592  3456
     ArchD        1728  2592
     ArchC        1296  1728
     ArchB         864  1296
     ArchA         648   864
     A0           2380  3368
     A1           1684  2380
     A2           1190  1684
     A3            842  1190
     A4            595   842
     A4Small       595   842
     A5            421   595
     A6            297   421
     A7            210   297
     A8            148   210
     A9            105   148
     A10            74   105
     B0           2836  4008
     B1           2004  2836
     B2           1418  2004
     B3           1002  1418
     B4            709  1002
     B5            501   709
     C0           2600  3677
     C1           1837  2600
     C2           1298  1837
     C3            918  1298
     C4            649   918
     C5            459   649
     C6            323   459
     Flsa          612   936
     Flse          612   936
     HalfLetter    396   612

For convenience you can just say the page size by media name (like A4, Ledger, whatever, iykyk). Otherwise -page acts a whole lot like -geometry — for real, like -page letter+43+43>.

This same option is also the vibe-checker for placing subimages when you're writing to a multi-image format that supports offsets, like GIF89 and MNG 💅. When you're using it for that, the offsets are always measured from the top left corner of the canvas and they are not affected by the -gravity option — no cap. To position a GIF or MNG image, do -page{+-}<x>{+-}<y> (e.g. -page +100+200). When writing to a MNG file, a -page option that shows up ahead of the first image in the sequence with nonzero width and height defines the width and height values written into the MHDR chunk. Otherwise the MNG width and height get computed from the bounding box that contains all the images in the sequence. And when writing a GIF89 file, only the bounding box method is used to figure out its dimensions — that's just how it is, bestie.

For a PostScript page, the image is sized like -geometry and positioned relative to the lower left hand corner of the page by {+-}<xoffset>{+-}<y offset>. Use -page 612x792>, for example, to center the image within the page ✨. If the image size exceeds the PostScript page, it gets reduced to fit the page — it's giving considerate. The default gravity for the -page option is NorthWest, i.e. positive x and y offset are measured rightward and downward from the top left corner of the page, unless the -gravity option is present with a value other than NorthWest.

The default page dimensions for a TEXT image is 612x792.

This option is used in concert with -density 🫶.

Use +page to remove the page settings for an image.

• -paint radius — simulate an oil painting 🎨

Each pixel is replaced by the most frequent color in a circular neighborhood whose width is specified with radius. Iconic behaviour, honestly.

• -pause seconds — pause between animation loops [animate]

Pause for the specified number of seconds before repeating the animation.

• -pause seconds — pause between snapshots [import]

Pause for the specified number of seconds before taking the next snapshot. Same flag, different context — it's giving versatility.

• -pen color — (This option has been replaced by the -fill option)

Like, it's literally right there in the original: replaced. Use -fill, bestie 💀.

• -ping — efficiently determine image characteristics ⚡

Use this option to disable reading the image pixels so that image characteristics such as the image dimensions may be obtained very quickly. For identify, use +ping to force reading the image pixels so that the pixel read rate may be included in the displayed information 👑

• -pointsize value — pointsize of the PostScript, X11, or TrueType font
• -preview type — image preview type 🦋

Use this option to affect the preview operation of an image (e.g. convert file.png -preview Gamma Preview:gamma.png). Choose from these previews:

• Rotate
• Shear
• Roll
• Hue
• Saturation
• Brightness
• Gamma
• Spiff
• Dull
• Grayscale
• Quantize
• Despeckle
• ReduceNoise
• AddNoise
• Sharpen
• Blur
• Threshold
• EdgeDetect
• Spread
• Shade
• Raise
• Segment
• Solarize
• Swirl
• Implode
• Wave
• OilPaint
• CharcoalDrawing
• JPEG

The default preview is JPEG 🌸.

• -process command — process a sequence of images using a process module 🧚

The command argument has the form module=arg1,arg2,arg3,...,argN where module is the name of the module to invoke (e.g. "Analyze") and arg1,arg2,arg3,...,argN are an arbitrary number of arguments to pass to the process module. The sequence of images is terminated by the appearance of any option. Periodt.

If the -process option appears after all of the input images, all images are processed.

For example:

gm convert logo: -process Analyze= \
  -format "%[BrightnessMean],%[BrightnessStddev]" info:-
51952,23294

• -profile filename — add ICM, IPTC, or generic profile to image 💜

-profile filename adds an ICM (ICC color management), IPTC (newswire information), or a generic (including Exif) profile to the image.

Use +profile icm, +profile iptc, or +profile profile_name to remove the respective profile. Multiple profiles may be listed, separated by commas. Profiles may be excluded from subsequent listed matches by preceding their name with an exclamation point. For example, +profile '!icm,*' strips all profiles except for the ICM profile. Use identify -verbose to find out what profiles are in the image file. Use +profile "*" to remove all profiles. Writing the image to a format that does not support profiles will of course also cause all profiles to be removed. The JPEG and PNG formats will store any profiles that have been read and not removed. In JPEG they are stored in APP1 markers, and in PNG they are stored as hex-coded binary in compressed zTXt chunks, except for the iCC chunk which is stored in the iCCP chunk 👑

To extract a profile, the -profile option is not used. Instead, simply write the file to an image format such as APP1, 8BIM, ICM, or IPTC.

For example, to extract the Exif data (which is stored in JPEG files in the APP1 profile), use

gm convert cockatoo.jpg exifdata.app1

Note that GraphicsMagick does not attempt to update any profile to reflect changes made to the image, e.g., rotation from portrait to landscape orientation, so it is possible that the preserved profile may contain invalid data. That's a real caveat, not me being dramatic 😭

• -preserve-timestamp — preserve the original timestamps of the file ⭐

Use this option to preserve the original modification and access timestamps of the file, even if it has been modified. Obsessed with a file that remembers where it came from, honestly.

• +progress — disable progress monitor and busy cursor

By default, when an image is displayed, a progress monitor bar is shown in the top left corner of an existing image display window, and the current cursor is replaced with an hourglass cursor. Use +progress to disable the progress monitor and busy cursor during display operations. While the progress monitor is disabled for all operations, the busy cursor continues to be enabled for non-display operations such as image processing. This option is useful for non-interactive display operations, or when a "clean" look is desired ✨

• -quality value — JPEG/MIFF/PNG/TIFF compression level 🌈

For the JPEG and MPEG image formats, quality is 0 (lowest image quality and highest compression) to 100 (best quality but least effective compression). The default quality is 75. Use the -sampling-factor option to specify the factors for chroma downsampling. To use the same quality value as that found by the JPEG decoder, use the -define jpeg:preserve-settings flag.

For the MIFF image format, and the TIFF format while using ZIP compression, quality/10 is the zlib compression level, which is 0 (worst but fastest compression) to 9 (best but slowest). It has no effect on the image appearance, since the compression is always lossless.

For the JPEG-2000 image format, quality is mapped using a non-linear equation to the compression ratio required by the Jasper library. This non-linear equation is intended to loosely approximate the quality provided by the JPEG v1 format. The default quality value 75 results in a request for 16:1 compression. The quality value 100 results in a request for non-lossy compression 🔥

For the MNG and PNG image formats, the quality value sets the zlib compression level (quality / 10) and filter-type (quality % 10). Compression levels range from 0 (fastest compression) to 100 (best but slowest). For compression level 0, the Huffman-only strategy is used, which is fastest but not necessarily the worst compression.

If filter-type is 4 or less, the specified filter-type is used for all scanlines:

     0: none
     1: sub
     2: up
     3: average
     4: Paeth

If filter-type is 5, adaptive filtering is used when quality is greater than 50 and the image does not have a color map, otherwise no filtering is used 💅

okay so if filter-type is 6, you get adaptive filtering with minimum-sum-of-absolute-values. cute!

And only if the output is MNG, if filter-type is 7, you get the LOCO color transformation plus adaptive filtering with minimum-sum-of-absolute-values. very demure, very MNG-only.

The default is quality 75, which means nearly the best compression with adaptive filtering. And like — the quality setting has no effect on how PNG and MNG images actually look, because the compression is always lossless. so you're not sacrificing vibes for bytes bestie 💅

For further info, see the PNG specification. iykyk 📖

---

When you're writing a JNG image with transparency, you need two quality values — one for the main image and one for the grayscale image that conveys the opacity channel. They get written as a single integer equal to the main image quality plus 1000 times the opacity quality. So like, if you want quality 75 for the main image and quality 90 to compress the opacity data, you use -quality 90075. math girl summer 🧮✨

For the PNM family of formats (PNM, PGM, and PPM) specify a quality factor of zero in order to obtain the ASCII variant of the format. Little heads up tho: -compress none used to be used to trigger ASCII output but provided the opposite result of what was expected as compared with other formats. 🙈 so yeah, don't do that.

For the TIFF format, the JPEG, WebP, Zip, and Zstd compression algorithms are influenced by the quality value. JPEG and WebP provide lossy compression so higher quality produces a larger file with less degradation. The Zip and Zstd compression algorithms (and WebP in lossless mode) are lossless, and for these algorithms a higher 'quality' means to work harder to produce a smaller file, but with no difference in image quality. it's giving "effort ≠ glow-up" 😌

• -raise <width>x<height> — lighten or darken image edges. this creates a 3-D effect 🎀 See -geometry for the details about the geometry specification. Offsets are not used. Use -raise to create a raised effect, otherwise use +raise.
• -random-threshold <channeltype> <LOWxHIGH> — random threshold the image. The channel or channels specified in the channeltype argument are reduced to binary, using a random-threshold method. The choices for channeltype are All, Intensity, Red, Green, Blue, Cyan, Magenta, Yellow, Black, and Opacity.
  • When channeltype is "All", the color samples are thresholded into a graylevel and then that gray level is stored in the three color channels. Separately, the opacity channel is thresholded into a bilevel opacity value which is stored in the opacity channel. For each pixel, a new random number is used to establish the threshold to be used. The threshold never exceeds the specified maximum (HIGH) and is never less than the specified minimum (LOW). 🎲
  • When channeltype is "intensity", only the color samples are thresholded. When channeltype is "opacity" or "matte", only the opacity channel is thresholded. The other named channels only threshold the associated channel.
• -recolor <matrix> — apply a color translation matrix to image channels. A user supplied color translation matrix (expressed as a text string) is used to translate/blend the image channels based on weightings in a supplied matrix which may be of order 3 (color channels only), 4 (color channels plus opacity), or 5 (color channels plus opacity and offset). Values in the columns of the matrix (red, green, blue, opacity) are used as multipliers with the existing channel values and added together according to the rows of the matrix. Matrix values are floating point and may be negative. The offset column (column 5) is purely additive and is scaled such that 0.0 to 1.0 represents the maximum quantum range (but values are not limited to this range). The math for the color translation matrix is similar to that used by Adobe Flash except that the offset is scaled to 1.0 (divide Flash offset by 255 for use with GraphicsMagick) so that the results are independent of quantum depth. 🤓

An identity matrix exists for each matrix order which results in no change to the image. The translation matrix should be based on an alteration of the identity matrix.

Identity matrix of order 3

1 0 0
0 1 0
0 0 1

which may be formatted into a convenient matrix argument similar to (comma is treated as white space):

-recolor "1 0 0, 0 1 0, 0 0 1"

Identity matrix of order 4

1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1

Identity matrix of order 5. The last row is required to exist for the purpose of parsing, but is otherwise not used. (iconic honestly)

1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1

As an example, an image wrongly in BGR channel order may be converted to RGB using this matrix (blue->red, red->blue):

0 0 1
0 1 0
1 0 0

and an RGB image using standard Rec.709 primaries may be converted to grayscale using this matrix of standard weighting factors:

0.2126 0.7152 0.0722
0.2126 0.7152 0.0722
0.2126 0.7152 0.0722

and contrast may be reduced by scaling down by 80% and adding a 10% offset:

0.8 0.0 0.0 0.0 0.1
0.0 0.8 0.0 0.0 0.1
0.0 0.0 0.8 0.0 0.1
0.0 0.0 0.0 0.8 0.1
0.0 0.0 0.0 0.0 1.0

• -map <filename> — choose a particular set of colors from this image. [convert or mogrify] By default, color reduction chooses an optimal set of colors that best represent the original image. Alternatively, you can choose a particular set of colors from an image file with this option. 🎨

Use +map to reduce all images in the image sequence that follows to a single optimal set of colors that best represent all the images. The sequence of images is terminated by the appearance of any option. If the +map option appears after all of the input images, all images are mapped.

• -red-primary <x>,<y> — red chromaticity primary point. that's it, that's the whole thing 🍓
• -region <width>x<height>{+-}<x>{+-}<y> — apply options to a portion of the image. The x and y offsets are treated in the same manner as in -crop. ✂️
• -remote — perform a X11 remote operation. The -remote command sends a command to a "gm display" or "gm animate" which is already running. The only command recognized at this time is the name of an image file to load. This capability is very useful to load new images without needing to restart GraphicsMagick (e.g. for a slide-show or to use GraphicsMagick as the display engine for a different GUI). Also see the +progress option for a way to disable progress indication for a clean look while loading new images. 👑
• -render — render vector operations. Use +render to turn off rendering vector operations. This is useful when saving the result to vector formats such as MVG or SVG. 🦋
• -repage <width>x<height>+xoff+yoff[!] — adjust image page offsets. Adjust the current image page canvas and position based on a relative page specification. This option may be used to change the location of a subframe (e.g. part of an animation) prior to composition. If the geometry specification is absolute (includes a '!'), then the offset adjustment is absolute and there is no adjustment to page width and height, otherwise the page width and height values are also adjusted based on the current image dimensions. Use +repage to set the image page offsets to default. 📐
• -resample <horizontal>x<vertical> — resample image to specified horizontal and vertical resolution. Resize the image so that its rendered size remains the same as the original at the specified target resolution. Either the current image resolution units or the previously set with -units are used to interpret the argument. For example, if a 300 DPI image renders at 3 inches by 2 inches on a 300 DPI device, when the image has been resampled to 72 DPI, it will render at 3 inches by 2 inches on a 72 DPI device. Note that only a small number of image formats (e.g. JPEG, PNG, and TIFF) are capable of storing the image resolution. For formats which do not support an image resolution, the original resolution of the image must be specified via -density on the command line prior to specifying the resample resolution.

Note that Photoshop stores and obtains image resolution from a proprietary embedded profile. If this profile exists in the image, then Photoshop will continue to treat the image using its former resolution, ignoring the image resolution specified in the standard file header. 😭 photoshop bestie we talked about this

Some image formats (e.g. PNG) require use of metric or english units so even if the original image used a particular unit system, if it is saved to a different format prior to resampling, then it may be necessary to specify the desired resolution units using -units since the original units may have been lost. In other words, do not assume that the resolution units are restored if the image has been saved to a file. 💀

• -resize <width>x<height>{%}{@}{!}{<}{>} — resize an image. This is an alias for the -geometry option and it behaves in the same manner. If the -filter option precedes the -resize option, the specified filter is used. ✨

There are some exceptions:

• When used as a composite option, -resize conveys the preferred size of the output image, while -geometry conveys the size and placement of the composite image within the main image.
• When used as a montage option, -resize conveys the preferred size of the montage, while -geometry conveys information about the tiles.

• -roll {+-}<x>{+-}<y> — roll an image vertically or horizontally. See -geometry for the details of the geometry specification. The x and y offsets are not affected by the -gravity option. 🌈

A negative x offset rolls the image left-to-right. A negative y offset rolls the image top-to-bottom. 💅

• -rotate <degrees>{<}{>} — rotate the image 🔄

Positive angles rotate the image clockwise, negative angles rotate counter-clockwise (yes bestie, the sign literally matters).

Use > to rotate the image only if its width exceeds the height. < rotates the image only if its width is less than the height. So like, if you specify -rotate "-90>" and the image size is 480x640, the image is not rotated. But if the image is 640x480, it gets rotated by -90 degrees. Also!! if you use > or <, enclose it in quotation marks so it doesn't get misinterpreted as a file redirection (classic little gotcha, iykyk).

Empty triangles left over from rotating the image are filled with the color defined as background (class backgroundColor). That color is specified using the format described under the -fill option. ✨

• -sample <geometry> — scale image using pixel sampling 🧚

See -geometry for details about the geometry specification. -sample ignores the -filter selection if the -filter option is present. Offsets, if present in the geometry string, are ignored, and the -gravity option has no effect. Like, it's on its own little program fr.

• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors 🎀

This option specifies the sampling factors to be used by the DPX, JPEG, MPEG, or YUV encoders for chroma downsampling. The sampling factor must be specified while reading the raw YUV format since it is not preserved in the file header. Industry-standard video subsampling notation such as "4:2:2" may also be used to specify the sampling factors. "4:2:2" is equivalent to a specification of "2x1".

The JPEG decoder obtains the original sampling factors (and quality settings) when a JPEG file is read. To re-use the original sampling factors (and quality setting) when JPEG is output, use the -define jpeg:preserve-settings flag. 🫶

• -scale <geometry> — scale the image. 📏

See -geometry for details about the geometry specification. -scale uses a simpler, faster algorithm, and it ignores the -filter selection if the -filter option is present. Offsets, if present in the geometry string, are ignored, and the -gravity option has no effect.

• -scene <value> — set scene number 🎬

This option sets the scene number of an image or the first image in an image sequence.

• -scenes <value-value> — range of image scene numbers to read 🔢

Each image in the range is read with the filename followed by a period (.) and the decimal scene number. You can change this behavior by embedding a %d, %0Nd, %o, %0No, %x, or %0Nx printf format specification in the file name. For example,

gm montage -scenes 5-7 image.miff montage.miff

makes a montage of files image.miff.5, image.miff.6, and image.miff.7, and

gm animate -scenes 0-12 image%02d.miff

animates files image00.miff, image01.miff, through image12.miff. Iconic behavior fr fr.

• -screen — specify the screen to capture 🖥️

This option indicates that the GetImage request used to obtain the image should be done on the root window, rather than directly on the specified window. In this way, you can obtain pieces of other windows that overlap the specified window, and more importantly, you can capture menus or other popups that are independent windows but appear over the specified window. Sneaky but so valid.

• -set <attribute> <value> — set an image attribute 🏷️

Set a named image attribute. The attribute is set on the current (previously specified on command line) image.

• +set <attribute> — unset an image attribute 🚫

Unset a named image attribute. The attribute is removed from the current (previously specified on command line) image.

• -segment <cluster threshold>x<smoothing threshold> — segment an image 🧩

Segment an image by analyzing the histograms of the color components and identifying units that are homogeneous with the fuzzy c-means technique.

Segmentation is a very useful fast and approximate color quantization algorithm for scanned printed pages or scanned cartoons. It may also be used as a special effect. Specify cluster threshold as the minimum percentage of total pixels in a cluster before it is considered valid. For huge images containing small detail, this may need to be a tiny fraction of a percent (e.g. 0.015) so that important detail is not lost. Smoothing threshold eliminates noise in the second derivative of the histogram. As the value is increased, you can expect a smoother second derivative. The default is 1.5. Add the -verbose option to see a dump of cluster statistics given the parameters used. The statistics may be used as a guide to help fine tune the options. ngl this one is kind of a whole vibe.

• -shade <azimuth>x<elevation> — shade the image using a distant light source 💡

Specify azimuth and elevation as the position of the light source. Use +shade to return the shading results as a grayscale image.

• -shadow <radius>{x<sigma>} — shadow the montage 🌑
• -shared-memory — use shared memory 🧠

This option specifies whether the utility should attempt to use shared memory for pixmaps. GraphicsMagick must be compiled with shared memory support, and the display must support the MIT-SHM extension. Otherwise, this option is ignored. The default is True.

• -sharpen <radius>{x<sigma>} — sharpen the image 🔪

Use a Gaussian operator of the given radius and standard deviation (sigma).

• -shave <width>x<height>{%} — shave pixels from the image edges ✂️

Specify the width of the region to be removed from both sides of the image and the height of the regions to be removed from top and bottom.

• -shear <x degrees>x<y degrees> — shear the image along the X or Y axis 🌀

Use the specified positive or negative shear angle.

Shearing slides one edge of an image along the X or Y axis, creating a parallelogram. An X direction shear slides an edge along the X axis, while a Y direction shear slides an edge along the Y axis. The amount of the shear is controlled by a shear angle. For X direction shears, x degrees is measured relative to the Y axis, and similarly, for Y direction shears y degrees is measured relative to the X axis.

Empty triangles left over from shearing the image are filled with the color defined as background (class backgroundColor). The color is specified using the format described under the -fill option. Same filler energy as rotate, slay.

• -silent — operate silently 🤫
• -size <width>x<height>{+offset} — width and height of the image 📐

Use this option to specify the width and height of raw images whose dimensions are unknown such as GRAY, RGB, or CMYK. In addition to width and height, use -size with an offset to skip any header information in the image or tell the number of colors in a MAP image file, (e.g. -size 640x512+256).

For Photo CD images, choose from these sizes:

192x128
384x256
768x512
1536x1024
3072x2048

Finally, use this option to choose a particular resolution layer of a JBIG or JPEG image (e.g. -size 1024x768).

• -snaps <value> — number of screen snapshots 📸

Use this option to grab more than one image from the X server screen, to create an animation sequence.

• -solarize <factor> — negate all pixels above the threshold level ☀️

Specify factor as the percent threshold of the intensity (0 - 99.9%). Note the range carefully bestie, 99.9 is the ceiling.

This option produces a solarization effect seen when exposing a photographic film to light during the development process.

• -spread <amount> — displace image pixels by a random amount 🎲

Amount defines the size of the neighborhood around each pixel to choose a candidate pixel to swap.

• -stegano <offset> — hide watermark within an image 🕵️‍♀️

Use an offset to start the image hiding some number of pixels from the beginning of the image. Note this offset and the image size. You will need this information to recover the steganographic image (e.g. display -size 320x256+35 stegano:image.png).

• -stereo — composite two images to create a stereo anaglyph 🥽

The left side of the stereo pair is saved as the red channel of the output image. The right side is saved as the green channel. Red-green stereo glasses are required to properly view the stereo image.

• -strip — remove all profiles and text attributes from the image 🧼

All embedded profiles and text attributes are stripped from the image. This is useful for images used for the web, or when output files need to be as small as possible.

Be careful not to use this option to remove author, copyright, and license information that you are required to retain when redistributing an image. 😭 Like please do not accidentally delete legal stuff, that's not the vibe.

• -stroke <color> — color to use when stroking a graphic primitive 🖌️

The color is specified using the format described under the -fill option.

See -draw for further details.

• -strokewidth <value> — set the stroke width 📏

See -draw for further details.

• -swirl <degrees> — swirl image pixels about the center 🌪️

Degrees defines the tightness of the swirl.

• -text-font <name> — font for writing fixed-width text 🔤

Specifies the name of the preferred font to use in fixed (typewriter style) formatted text. The default is 14 point Courier.

You can tag a font to specify whether it is a PostScript, TrueType, or X11 font. For example, Courier.ttf is a TrueType font and x:fixed is X11.

• -texture <filename> — name of texture to tile onto the image background 🧵
• -threshold <value>{%} — threshold the image ⚖️
• -threshold — okay so picture this: every pixel sample with an intensity value greater than the threshold gets slammed to the maximum intensity (that's white, bestie), and everything else gets assigned the minimum intensity (black). If you slap a percent prefix on it, then the threshold is read as a percentage of the available range. 💅

Wanna make a black-and-white image out of a color one, efficiently? Do this:

gm convert -threshold 50% in.png out.png

The optimum threshold value? Highly image-dependent, ngl — it depends on the nature of the image. And if you want to threshold individual channels, use the -operator subcommand with its Threshold, Threshold-White, or Threshold-Black options. iykyk.

• -thumbnail <width>x<height>{%}{@}{!}{<}{>} — resize an image, but quickly ⚡

The -thumbnail command resizes the image as quickly as possible, with way more concern for speed than for resulting image quality. Regardless, the resulting quality should be acceptable for many uses. It's primarily intended for generating smaller versions of the image, but it may also be used to enlarge the image. The -thumbnail geometry argument observes the same syntax and rules as it does for -resize. 🫶

• -tile <filename> — tile image when filling a graphic primitive.
• -tile <geometry> — layout of images [montage].
• -title <string> — assign a title to the displayed image [animate, display, montage] 🏷️

Use this one to give an image a specific title. It gets assigned to the image window and is typically displayed in the window title bar. Optionally you can include the image filename, type, width, height, Exif data, or other image attribute by embedding special format characters described under the -format option.

For example:

-title "%m:%f %wx%h"

produces an image title of MIFF:bird.miff 512x480 for an image titled bird.miff whose width is 512 and height is 480. iconic honestly ✨

• -transform — transform the image.

This option applies the transformation matrix from a previous -affine option.

gm convert -affine 2,2,-2,2,0,0 -transform bird.ppm bird.jpg

• -transparent <color> — make this color transparent within the image 🫥

The color is specified using the format described under the -fill option.

• -treedepth <value> — tree depth for the color reduction algorithm 🌳

Normally this integer value is zero or one. A value of zero or one causes the use of an optimal tree depth for the color reduction algorithm.

An optimal depth generally allows the best representation of the source image with the fastest computational speed and the least amount of memory. However, the default depth is inappropriate for some images. To assure the best representation, try values between 2 and 8 for this parameter. Refer to quantize for more details.

Heads up: the -colors or -monochrome option — or writing to an image format which requires color reduction — is required for this option to take effect. No cap.

• -trim — trim an image ✂️

This option removes any edges that are exactly the same color as the corner pixels. Use -fuzz to make -trim remove edges that are nearly the same color as the corner pixels.

• -type <type> — the image type 🎭

Choose from: Bilevel, Grayscale, GrayscaleMatte, Palette, PaletteMatte, TrueColor, TrueColorMatte, ColorSeparation, ColorSeparationMatte, or Optimize.

Normally, when a format supports different subformats such as bilevel, grayscale, grayscale+alpha, palette, truecolor, and truecolor+alpha, the encoder will try to choose a suitable subformat based on the nature of the image. The -type option may be used to tailor the output subformat. By default the output subformat is based on readily available image information and is usually similar to the input format.

Specify -type Optimize in order to enable inspecting all pixels (if necessary) in order to find the most efficient subformat. Inspecting all of the pixels may be slow for very large images, particularly if they are stored in a disk cache. If an RGB image contains only gray pixels, then every pixel in the image must be inspected in order to decide that the image is actually grayscale! 😭 that's a lot of pixels.

Sometimes a specific subformat is what you want. For example, to force a JPEG image to be written in TrueColor RGB format even though only gray pixels are present, use:

gm convert bird.pgm -type TrueColor bird.jpg

Similarly, using -type TrueColorMatte will force the encoder to write an alpha channel even though the image is opaque, if the output format supports transparency.

Some pseudo-formats (e.g. the XC format) will respect the requested type if it occurs previously on the command line. For example, to obtain a DirectClass solid color canvas image rather than PsuedoClass, use:

gm convert -size 640x480 -type TrueColor xc:red red.miff

Likewise, specify -type Bilevel, Grayscale, TrueColor, or TrueColorMatte prior to reading a Postscript (or PDF file) in order to influence the type of image that Ghostcript returns. Reading performance will be dramatically improved for black/white Postscript if Bilevel is specified, and will be considerably faster if Grayscale is specified. slay that performance 💨

• -update <seconds> — detect when the image file is modified and redisplay. 👀

Suppose that while you are displaying an image, the file that is currently displayed is over-written. display will automatically detect that the input file has been changed and update the displayed image accordingly.

• -units <type> — the units of image resolution 📐

Choose from: Undefined, PixelsPerInch, or PixelsPerCentimeter. This option is normally used in conjunction with the -density option.

• -unsharp <radius>{x<sigma>}{+<amount>}{+<threshold>} — sharpen the image with an unsharp mask operator 🔪✨

The -unsharp option sharpens an image. The image is convolved with a Gaussian operator of the given radius and standard deviation (sigma). For reasonable results, radius should be larger than sigma. Use a radius of 0 to have the method select a suitable radius.

The parameters are:

• radius — the radius of the Gaussian, in pixels, not counting the center pixel (default 0).
• sigma — the standard deviation of the Gaussian, in pixels (default 1.0).
• amount — the percentage of the difference between the original and the blur image that is added back into the original (default 1.0).
• threshold — the threshold, as a fraction of MaxRGB, needed to apply the difference amount (default 0.05).

• -use-pixmap — use the pixmap. 🖼️
• -verbose — print detailed information about the image 🗣️

This information is printed: image scene number; image name; image size; the image class (DirectClass or PseudoClass); the total number of unique colors; and the number of seconds to read and transform the image. If the image is DirectClass, the total number of unique colors is not displayed unless -verbose is specified twice — since it may take quite a long time to compute, particularly for deep images. If the image is PseudoClass then its pixels are defined by indexes into a colormap. If the image is DirectClass then each pixel includes a complete and independent color specification. 😮‍💨

If -colors is also specified, the total unique colors in the image and color reduction error values are printed. Refer to quantize for a description of these values.

• -version — print detailed GraphicsMagick version/build information 🧾

Reports the release version, maximum sample-depth, web-site URL, copyright notice, feature support, and the options used while building the software. This is the same output as produced by the version sub-command.

• -version — print compact GNU-style GraphicsMagick version information 🧾✨

Reports the release version, maximum sample-depth, web-site URL, and copyright notice.

Yes bestie, -version shows up twice in this list — one glow-up is the detailed one, the other is the compact GNU-style one. Both real, both iconic.

• -view <string> — FlashPix viewing parameters.
• -virtual-pixel <method> — specify contents of "virtual pixels" 🔮

This option defines "virtual pixels" for use in operations that can access pixels outside the boundaries of an image.

Choose from these methods:

• Constant — use the image background color.
• Edge — extend the edge pixel toward infinity (default).
• Mirror — mirror the image.
• Tile — tile the image.

This option affects operations that use virtual pixels such as -blur, -sharpen, -wave, etc.

• -visual <type> — animate images using this X visual type 🎨

Choose from these visual classes:

StaticGray
GrayScale
StaticColor
PseudoColor
TrueColor
DirectColor
default
visual id

The X server must support the visual you choose, otherwise an error occurs. If a visual is not specified, the visual class that can display the most simultaneous colors on the default screen is chosen. periodt.

• -watermark <brightness>x<saturation> — percent brightness and saturation of a watermark 💧
• -wave <amplitude>x<wavelength> — alter an image along a sine wave 🌊

Specify amplitude and wavelength of the wave.

• -white-point <x>,<y> — chromaticity white point ⚪
• -white-threshold red[,green][,blue][,opacity] — pixels above the threshold become white 🤍

Use -white-threshold to set pixels with values above the specified threshold to maximum value (white). If only one value is supplied, or the red, green, and blue values are identical, then intensity thresholding is used. If the color threshold values are not identical then channel-based thresholding is used, and color distortion will occur. Specify a negative value (e.g. -1) if you want a channel to be ignored but you do want to threshold a channel later in the list. If a percent (%) symbol is appended, then the values are treated as a percentage of maximum range.

• -window <id> — make image the background of a window 🪟✨

So id can be a window id or a name. Specify root to select X's root window as the target window — yes bestie, the whole entire root window, main character energy fr.

By default the image is tiled onto the background of the target window. If backdrop or -geometry are specified, the image is surrounded by the background color. Refer to X RESOURCES for details 👀.

Heads up 💀 The image will not display on the root window if the image has more unique colors than the target window colormap allows. Use -colors to reduce the number of colors. No cap, that's the fix.

• -window-group — specify the window group 🫶 (that's it, that's the whole vibe)
• -write <filename> — write an intermediate image [convert, composite] 💾✨

The current image is written to the specified filename and then processing continues using that image. The following is an example of how several sizes of an image may be generated in one command (repeat as often as needed):

gm convert input.jpg -resize 50% -write input50.jpg \
          -resize 25% input25.jpg

• -write <filename> — write the image to a file [display] 🖼️

If filename already exists, you will be prompted as to whether it should be overwritten 🙈 (no silent clobbering, we love a considerate icon).

By default, the image is written in the format that it was read in as. To specify a particular image format, prefix filename with the image type and a colon (e.g., ps:image) or specify the image type as the filename suffix (e.g., image.ps). Specify file as - for standard output. If file has the extension .Z or .gz, the file size is compressed using compress or gzip respectively. Precede the image file name with | to pipe to a system command 🔥.

Use -compress to specify the type of image compression.

The equivalent X resource for this option is writeFilename (class WriteFilename). See "X Resources", below, for details ⭐.

---

🌍 ENVIRONMENT 🌍

okay bestie, this is where gm reads its vibe check from the shell 💅 everything here is an environment variable, and honestly some of these are lowkey iconic and some are highkey "handle with care" — let's get into it 🫶

• COLUMNS — the output screen width, used when it's formatting text for the screen. Many Unix systems keep this shell variable up to date, but ngl it might need to be explicitly exported before GraphicsMagick can actually see it 👀
• DISPLAY — the X11 display ID, in the form hostname:display.screen (so: host, display number, and screen) ✨
• HOME — where your home directory lives. For security reasons it's now only observed by "uninstalled" builds of GraphicsMagick, i.e. the ones that don't have their location hard-coded or set by an installer. When it is supported, GraphicsMagick searches for configuration files in $HOME/.magick if that directory exists. Need more flexibility? See MAGICK_CODER_MODULE_PATH, MAGICK_CONFIGURE_PATH, and MAGICK_FILTER_MODULE_PATH 🎀
• MAGICK_ACCESS_MONITOR — set it to TRUE and command line monitor mode (the one you turn on with -monitor) will also show files accessed (including temporary files) plus any external commands that get executed. Super useful for debugging, and it also illustrates the arguments made available to an access handler registered by the MagickSetConfirmAccessHandler() C library function 🔍
• MAGICK_CODER_STABILITY — the minimum coder stability level before a coder gets used. The available levels are PRIMARY, STABLE, UNSTABLE, and BROKEN. Default minimum level is UNSTABLE, which means all available working coders get used 💀 The whole point of this one is to reduce the security exposure (or the apparent complexity) that comes from the huge number of formats supported. Break it down:
  • PRIMARY coders are the commonly used formats with very well maintained implementations 👑
  • STABLE coders are reasonably well maintained, but they're less used formats 🌸
  • UNSTABLE coders either have weak implementations, or the file format itself is weak, or the probability the coder will even be needed is vanishingly small 🫠
  • BROKEN coders are known to often not work properly, or might not be useful in their current state at all 😭
• MAGICK_CODER_MODULE_PATH — the search path used when hunting for image format coder modules. This path lets you arbitrarily extend the image formats GraphicsMagick supports by dropping loadable modules in an arbitrary location instead of copying them into the GraphicsMagick installation directory. Formatting is like operating system search paths (colon delimited for Unix, semi-colon delimited for Microsoft Windows), and this user-specified path is used before trying the default search path 🦋
• MAGICK_CONFIGURE_PATH — same deal but for configuration (.mgk) files. Formatting is like operating system search paths (colon delimited for Unix, semi-colon delimited for Microsoft Windows), and again your user-specified path is used before trying the default search path ✨
• MAGICK_DEBUG — debug options (peep -debug for the details). Real talk: setting the configure debug option via an environment variable (e.g. MAGICK_DEBUG=configure) is necessary to see the complete initialization process, which includes searching for configuration files 👀
• MAGICK_FILTER_MODULE_PATH — the search path used when hunting for filter process modules (the ones invoked via -process). This path lets you arbitrarily extend GraphicsMagick's image processing functionality by adding loadable modules to an arbitrary location rather than copying them into the GraphicsMagick installation directory. Formatting is like operating system search paths (colon delimited for Unix, semi-colon delimited for Microsoft Windows), and the user-specified path is used before trying the default search path 💜
• MAGICK_GHOSTSCRIPT_PATH — for Microsoft Windows: specify the path to the Ghostscript installation rather than searching for it via the Windows registry. This helps in case Ghostscript wasn't installed via the Ghostscript Windows installer, or you just want more control over the Ghostscript being used 🎀
• MAGICK_HOME — path to the top of the GraphicsMagick installation directory. Only observed by "uninstalled" builds of GraphicsMagick which don't have their location hard-coded or set by an installer 🌈
• MAGICK_MMAP_READ — if it's set to TRUE, GraphicsMagick will attempt to memory-map the input file for reading. This usually substantially improves repeated read performance, since the file is already in memory after the first time it's been read 🔥 But — testing shows performance may be reduced for files accessed for the first time, since data is accessed via page-faults (upon first access) and many operating systems fail to do sequential read-ahead of memory mapped files, particularly if those files are accessed over a network. And if you read many large input files, enabling this option may harm performance by overloading the operating system's VM system, since it then needs to free unmapped pages and map new ones 😬
• MAGICK_IO_FSYNC — if set to TRUE, GraphicsMagick will request that the output file is fully flushed and synchronized to disk when it's closed. This incurs a performance penalty, but the benefit is that if the power fails or the system crashes, the file should be valid on disk. If image files are referenced from a database, this option helps assure that the files referenced by the database are valid 💖
• MAGICK_IOBUF_SIZE — the amount of I/O buffering (in bytes) to use when reading and writing encoded files. The default is 16384, which is observed to work well for many cases. The best value for a local filesystem is usually the native filesystem block size (e.g. 4096, 8192, or even 131,072 for ZFS) in order to minimize the number of physical disk I/O operations. I/O performance to files accessed over a network may benefit significantly by tuning this option. Larger values are not necessarily better (they may be slower!), and there is rarely any benefit from using values larger than 32768. Use convert's -verbose option in order to evaluate read and write rates in pixels per second, while keeping in mind that the operating system will try to cache files in RAM 🧚
• MAGICK_LIMIT_DISK — maximum amount of disk space allowed for use by the pixel cache 💾
• MAGICK_LIMIT_FILES — maximum number of open files 📂
• MAGICK_LIMIT_MAP — maximum size of a memory mapped file allocation. A memory mapped file consumes memory when the file is accessed, although the system may reclaim such memory when needed 🫶
• MAGICK_LIMIT_MEMORY — maximum amount of memory to allocate from the heap 🧠
• MAGICK_LIMIT_PIXELS — maximum number of total pixels (image rows times image columns) to allow for any image which is requested to be created or read. Useful for placing a limit on how large an image may be. If the input image file has image dimensions larger than the pixel limit, then the image memory allocation is denied and an error is returned immediately. Heads up: this is a per-image limit and does not limit the total number of pixels due to multiple image frames/pages (e.g. multi-page document or an animation) 👑
• MAGICK_LIMIT_READ — maximum number of uncompressed bytes which may be read while decoding an image. Every read by the software from the input file is counted against the total, even if it has been read before. Decoding fails when the limit is reached. This limit helps defend against highly compressed files (e.g. via gzip), or files which use complex looping structures, or when data is being read from a stream (pipe) 🛡️
• MAGICK_LIMIT_WIDTH — maximum pixel width of an image read, or created 📏
• MAGICK_LIMIT_HEIGHT — maximum pixel height of an image read, or created 📐
• MAGICK_TMPDIR — path to the directory where GraphicsMagick should write temporary files. The default is to use the system default, or the location set by TMPDIR.

⚠️ Warning: Linux and some other systems often mount the default temporary file directory (/tmp) as a memory-based filesystem (e.g. tmpfs), which is paged out to the swap partition when there is memory pressure. When GraphicsMagick writes large pixel-cache files, this causes memory contention with other processes, and GraphicsMagick (or another program!) may encounter a SIGBUS signal when swap becomes full 💀 The user may create a directory on a normal "disk" based filesystem, which has plenty of free space (e.g. using mktemp -d $HOME/tmp) and use MAGICK_TMPDIR to cause GraphicsMagick to use it. Note that while GraphicsMagick makes every effort to free temporary files it creates, a memory-based temporary file directory has the security advantage of forgetting everything after a reboot 🧠✨

• TMPDIR — for POSIX-compatible systems (Unix-compatible), the path to the directory where all applications should write temporary files. Overridden by MAGICK_TMPDIR if it is set 🐧
• TMP or TEMP — for Microsoft Windows, the path to the directory where applications should write temporary files. Overridden by MAGICK_TMPDIR if it is set 🪟
• OMP_NUM_THREADS — as per the OpenMP standard, this specifies the number of threads to use in parallel regions. Some compilers default the number of threads to use to the number of processor cores available, while others default to just one thread. See the OpenMP specification for other standard adjustments and your compiler's manual for vendor-specific settings ⚡

---

🎀 CONFIGURATION FILES 🎀

GraphicsMagick uses a number of XML format configuration files, and they’re literally giving organized bestie energy:

• colors.mgk — colors configuration file

<?xml version="1.0"?>
<colormap>
  <color name="AliceBlue" red="240" green="248" blue="255"
         compliance="SVG, X11, XPM" />
</colormap>

• delegates.mgk — delegates configuration file
• log.mgk — logging configuration file

<?xml version="1.0"?>
<magicklog>
  <log events="None" />
  <log output="stdout" />
  <log filename="Magick-%d.log" />
  <log generations="3" />
  <log limit="2000" />
  <log format="%t %r %u %p %m/%f/%l/%d:\n  %e"  />
</magicklog>

• modules.mgk — loadable modules configuration file

<?xml version="1.0"?>
<modulemap>
  <module magick="8BIM" name="META" />
</modulemap>

• type.mgk — master type (fonts) configuration file

<?xml version="1.0"?>
<typemap>
  <include file="type-windows.mgk" />
  <type
    name="AvantGarde-Book"
    fullname="AvantGarde Book"
    family="AvantGarde"
    foundry="URW"
    weight="400"
    style="normal"
    stretch="normal"
    format="type1"
    metrics="/usr/local/share/ghostscript/fonts/a010013l.afm"
    glyphs="/usr/local/share/ghostscript/fonts/a010013l.pfb"
  />
</typemap>

---

💫 GM ANIMATE 💫

Okay so animate displays a sequence of images on any workstation display running an X server. First, animate determines the hardware capabilities of the workstation. If the number of unique colors in an image is less than or equal to the number the workstation can support, the image is displayed in an X window. Otherwise, the number of colors in the image is first reduced to match the color resolution of the workstation before it is displayed.

This means a continuous-tone 24 bits-per-pixel image can display on an 8 bit pseudo-color device or monochrome device. In most instances the reduced color image closely resembles the original. Alternatively, a monochrome or pseudo-color image sequence can display on a continuous-tone 24 bits-per-pixel device.

To help prevent color flashing on X server visuals that have colormaps, animate creates a single colormap from the image sequence. This can be rather time consuming. You can speed this operation up by reducing the colors in the image before you “animate” them. Use mogrify to color reduce the images to a single colormap. See mogrify for details. Alternatively, you can use a Standard Colormap; or a static, direct, or true color visual. You can define a Standard Colormap with xstdcmap. See xstdcmap(1) for details. This method is recommended for colormapped X server because it eliminates the need to compute a global colormap.

---

🌸 EXAMPLES 🌸

To animate a set of images of a cockatoo, use:

gm animate cockatoo.*

To animate a cockatoo image sequence while using the Standard Colormap best, use:

xstdcmap -best
gm animate -map best cockatoo.*

To animate an image of a cockatoo without a border centered on a backdrop, use:

gm animate +borderwidth -backdrop cockatoo.*

---

✨ OPTIONS ✨

For a more detailed description of each option, see Options, above.

• -authenticate <string> — decrypt image with this password
• -backdrop — display the image centered on a backdrop.
• -background <color> — the background color
• -bordercolor <color> — the border color
• -borderwidth <geometry> — the border width
• -chop <width>x<height>{+-}<x>{+-}<y>{%} — remove pixels from the interior of an image
• -colormap <type> — define the colormap type
• -colors <value> — preferred number of colors in the image
• -colorspace <value> — the type of colorspace
• -crop <width>x<height>{+-}<x>{+-}<y>{%} — preferred size and location of the cropped image
• -debug <events> — enable debug printout
• -define <key>{=<value>},... — add coder/decoder specific options
• -delay <1/100ths of a second> — display the next image after pausing
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image
• -depth <value> — depth of the image
• -display <host:display[.screen]> — specifies the X server to contact
• -dispose <method> — GIF disposal method
• -dither — apply Floyd/Steinberg error diffusion to the image
• -font <name> — use this font when annotating the image with text
• -foreground <color> — define the foreground color
• -gamma <value> — level of gamma correction
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — Specify dimension, offset, and resize options.
• -help — print usage instructions
• -iconGeometry <geometry> — specify the icon geometry
• -iconic — iconic animation
• -interlace <type> — the type of interlacing scheme
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit
• -log <string> — Specify format for debug log
• -map <type> — display image using this type.
• -matte — store matte channel if the image has one
• -mattecolor <color> — specify the color to be used with the -frame option
• -monitor — show progress indication
• -monochrome — transform the image to black and white
• -name — name an image
• -noop — NOOP (no option)
• -pause <seconds> — pause between animation loops [animate]
• -remote — perform a X11 remote operation
• -rotate <degrees>{<}{>} — rotate the image
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors
• -scenes <value-value> — range of image scene numbers to read
• -shared-memory — use shared memory
• -size <width>x<height>{+offset} — width and height of the image
• -text-font <name> — font for writing fixed-width text
• -title <string> — assign title to displayed image [animate, display, montage]
• -treedepth <value> — tree depth for the color reduction algorithm
• -trim — trim an image
• -type <type> — the image type
• -verbose — print detailed information about the image
• -version — print detailed GraphicsMagick version/build information
• -version — print compact GNU-style GraphicsMagick version information
• -visual <type> — animate images using this X visual type
• -window <id> — make image the background of a window

For a more detailed description of each option, see Options, above.

Any option you specify on the command line remains in effect for the group of images following it, until the group is terminated by the appearance of any option or -noop. For example, to animate three images, the first with 32 colors, the second with an unlimited number of colors, and the third with only 16 colors, use:

gm animate -colors 32 cockatoo.1 -noop cockatoo.2
         -colors 16 cockatoo.3

Animate options can appear on the command line or in your X resources file. See X(1). Options on the command line supersede values specified in your X resources file. Image filenames may appear in any order on the command line if the image format is MIFF (refer to miff(5) and the scene keyword is specified in the image). Otherwise the images will display in the order they appear on the command line.

---

🖱️ MOUSE BUTTONS 🖱️

Press any button to map or unmap the Command widget. See the next section for more information about the Command widget.

---

🎛️ COMMAND WIDGET 🎛️

The Command widget lists a number of sub-menus and commands. They are:

• Animate
  • Open
  • Play
  • Step
  • Repeat
  • Auto Reverse
• Speed
  • Faster
  • Slower
• Direction
  • Forward
  • Reverse
• Image Info
• Help
• Quit

Menu items with an indented triangle have a sub-menu. They are represented above as the indented items. To access a sub-menu item, move the pointer to the appropriate menu and press a button and drag. When you find the desired sub-menu item, release the button and the command is executed. Move the pointer away from the sub-menu if you decide not to execute a particular command.

---

⌨️ KEYBOARD ACCELERATORS ⌨️

• Ctl+O — Press to load an image from a file.
• space — Press to display the next image in the sequence.
• < — Press to speed-up the display of the images. Refer to -delay for more information.
• > — Press to slow the display of the images. Refer to -delay for more information.
• ? — Press to display information about the image. Press any key or button to erase the information. This information is printed: image name; image size; and the total number of unique colors in the image.
• F1 — Press to display helpful information about animate.
• Ctl-q — Press to discard all images and exit program.
• background (class Background) — the color you want behind your image in the Image window. Default is #ccc. 🎨
• borderColor (class BorderColor) — the color of the Image window's border. Default is #ccc.
• borderWidth (class BorderWidth) — how thick that border is, in pixels. Default is 2.
• font (class Font or FontList) — the font name you want for normal formatted text. Default is 14 point Helvetica. ✨
• foreground (class Foreground) — the color for text inside the Image window. Default is black.
• geometry (class geometry) — the preferred size and position of the image window. Lowkey important: it is not necessarily obeyed by all window managers. Offsets, if present, are handled in X(1) style — a negative x offset is measured from the right edge of the screen to the right edge of the icon, and a negative y offset is measured from the bottom edge of the screen to the bottom edge of the icon.
• iconGeometry (class IconGeometry) — preferred size and position of the application when it's iconified. Again, not necessarily obeyed by all window managers. Offsets, if present, are handled in the same manner as in class Geometry.
• iconic (class Iconic) — set this to say "bestie I'd prefer my windows start out invisible," like you'd already iconified them immediately. Window managers may choose not to honor that request. 🙈
• matteColor (class MatteColor) — the color of windows. It's used for the backgrounds of windows, menus, and notices. That 3D effect comes from highlight and shadow colors derived from this color. Default value: #ddd.
• name (class Name) — which name the app's resources should be found under. Genuinely useful in shell aliases to tell different invocations of an application apart, without resorting to creating links to alter the executable file name. Default is the application name. 🏷️
• sharedMemory (class SharedMemory) — whether animate should attempt to use shared memory for pixmaps. ImageMagick must be compiled with shared memory support, and the display must support the MIT-SHM extension — otherwise this resource is ignored. Default is True.
• text_font (class textFont) — the font name for fixed (typewriter style) formatted text. Default is 14 point Courier. 🎀
• title (class Title) — the title used for the Image window. Window managers sometimes use it to give you a header identifying the window. Default is the image file name.

---

🫶 GM BATCH 🫶

DESCRIPTION

batch runs an arbitrary number of the utility commands (e.g. convert) as a simple linear batch script, and the whole point is to improve execution efficiency, and/or to let it run as a subordinate co-process under the control of an arbitrary script or program. Periodt. 💅

EXAMPLES

To drive gm batch using a shell script (or a program written in any language), have the script/program send commands to gm batch via its standard input. You specify standard input by using - as the file name. This example converts all files matching *.jpg to TIFF format while rotating each file by 90 degrees and stripping all embedded profiles. The shell script syntax is standard Unix shell:

for file in *.jpg
do
  outfile=`basename $file .jpg`.tiff
  echo convert -verbose "'$file'" -rotate 90 \
  +profile "'*'" "'$outfile'"
done | gm batch -echo on -feedback on -

We can accomplish the same as the previous example by putting all the commands in a text file and then specifying the name of the text file as the script to execute:

for file in *.jpg
do
  outfile=`basename $file .jpg`.tiff
  echo convert -verbose "'$file'" -rotate 90 \
  +profile "'*'" "'$outfile'"
done > script.txt
gm batch -echo on -feedback on script.txt

OPTIONS

Options are processed from left to right and must appear before any filename argument. 🚨

• -echo on|off — command echo on or off. Say on to enable echoing commands to standard output as they're read, or off to disable. Default is off.
• -escape unix|windows — parse using unix or windows syntax. Commands have to be parsed from the input stream, and escaping needs to be used to protect spaces or quoting characters in the input. Say unix for unix-style command line parsing, or windows for Microsoft Windows command shell style parsing. The default depends on whether the software is compiled for Microsoft Windows or for a Unix-type system (including Cygwin on Microsoft Windows). It is recommended to use unix syntax because it's more powerful and more portable. 💯
• -fail text — text to print if a command fails. When feedback is enabled, this is the text printed when the command fails. Default text is FAIL.
• -feedback on|off — enable error feedback. Print the text feedback (see the -pass and -fail options) after each command to show the result. Default is off.
• -help — prints batch command help.
• -pass text — text to print if a command passes. When feedback is enabled, this is the text printed when the command passes. Default text is PASS. ✅
• -prompt text — prompt text to use for the command line. If no filename argument was specified, a simple command prompt appears where you may enter GraphicsMagick commands. Default prompt is GM>. Use this option to change it to something else.
• -stop-on-error on|off — specify if command processing stops on error. Normally command processing continues if a command hits an error. Specify -stop-on-error on to make processing quit immediately on error.

---

🔥 GM BENCHMARK 🔥

DESCRIPTION

benchmark executes an arbitrary gm utility command (e.g. convert) for one or more loops, and/or a specified execution time, and then reports a whole bunch of execution metrics. For builds using OpenMP, there's a mode that runs the benchmark with an increasing number of threads and reports speedup and multi-thread execution efficiency. If benchmark is used to execute a command without any additional benchmark options, then the command is run once. ⭐

EXAMPLES

To obtain benchmark information for a single execution of a command:

gm benchmark convert input.ppm -gaussian 0x1 output.ppm

To obtain benchmark information from 100 iterations of the command:

gm benchmark -iterations 100 convert input.ppm \
  -gaussian 0x1 output.ppm

To obtain benchmark information by iterating the command until a specified amount of time (in seconds) has been consumed:

gm benchmark -duration 30 convert input.ppm \
  -gaussian 0x1 output.ppm

To obtain a full performance report with an increasing number of threads (1-32 threads, stepping the number of threads by four each time):

gm benchmark -duration 3 -stepthreads 4 convert \
  input.ppm -gaussian 0x2 output.ppm

Here's how to read that output, bestie:

• threads - number of threads used.
• iter - number of command iterations executed.
• user - total user time consumed.
• total - total elapsed time consumed.
• iter/s - number of command iterations per second.
• iter/cpu - amount of CPU time consumed per iteration.
• speedup - speedup compared with one thread.
• karp-flatt - Karp-Flatt measure of speedup efficiency.

Please note that the reported "speedup" is based on the execution time of just one thread. A preliminary warm-up pass is used before timing the first loop, to make sure the CPU is brought out of power-saving modes and that system caches are warmed up. Most modern CPUs provide a "turbo" mode where the CPU clock speed is increased (e.g. by a factor of two) when only one or two cores are active. If the CPU grows excessively hot (due to insufficient cooling), then it may dial back its clock rates as a form of thermal management. These factors result in an under-reporting of speedup compared to if "turbo" mode was disabled and the CPU didn't need to worry about thermal management. The powertop utility available under Linux and Solaris gives you a way to observe CPU core clock rates while a benchmark is running. 🌡️

OPTIONS

Options are processed from left to right and must appear before any argument.

• -duration duration — duration to run benchmark. Specify the number of seconds to run the benchmark. The command is executed repeatedly until the specified amount of time has elapsed. ⏱️
• -help — prints benchmark command help.
• -iterations loops — number of command iterations. Specify the number of iterations to run the benchmark. The command is executed repeatedly until the specified number of iterations has been reached.
• -rawcsv — print results in CSV format. Output comes as comma-separated value (CSV), which is easy to parse for plotting or importing into a spreadsheet or database. The values reported are threads, iterations, user_time, and elapsed_time. 🍓
• -stepthreads step — execute a per-thread benchmark ramp. Increment the number of threads at each step by the specified value. The maximum number of threads is taken from the standard OMP_NUM_THREADS environment variable. 🦋

💅 GM COMPARE 💅

okay so compare is the one that looks at two images and is like "bestie… what changed 👀". It compares two similar images using a statistical method you pick (see -metric) and/or by writing out a difference image (-file), where the altered pixels get annotated using a method you also pick (see -highlight-style) plus a color (see -highlight-color). reference-image is the original, and compare-image is the (possibly) altered version — and fr fr it should have the same dimensions as reference-image. 🫶

---

✨ EXAMPLES ✨

To compare two images using Mean Square Error (MSE) statistical analysis use:

gm compare -metric mse original.miff compare.miff

To create an annotated difference image use:

gm compare -highlight-style assign -highlight-color purple \
  -file diff.miff original.miff compare.miff

---

🎀 OPTIONS 🎀

Options are processed in command line order. Any option you specify on the command line remains in effect only for the image that follows. All options are reset to their default values after each image is read. (yes, every single time, no carrying over, that's the vibe 💅)

For a more detailed description of each option, see Options, above.

• -authenticate <string> — decrypt image with this password 🔐
• -auto-orient — orient (rotate) image so it is upright
• -colorspace <value> — the type of colorspace
• -compress <type> — the type of image compression
• -debug <events> — enable debug printout
• -define <key>{=<value>},... — add coder/decoder specific options
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image
• -depth <value> — depth of the image
• -display <host:display[.screen]> — specifies the X server to contact
• -endian <type> — specify endianness (MSB, LSB, or Native) of image
• -file <filename> — write annotated difference image to file 📝
• -help — print usage instructions
• -highlight-color <color> — pixel annotation color
• -highlight-style <style> — pixel annotation style
• -interlace <type> — the type of interlacing scheme
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit
• -log <string> — Specify format for debug log
• -matte — store matte channel if the image has one
• -maximum-error <limit> — specifies the maximum amount of total image error
• -metric <metric> — comparison metric (MAE, MSE, PAE, PSNR, RMSE)
• -monitor — show progress indication
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors
• -size <width>x<height>{+offset} — width and height of the image
• -type <type> — the image type
• -verbose — print detailed information about the image
• -version — print detailed GraphicsMagick version/build information
• -version — print compact GNU-style GraphicsMagick version information

For a more detailed description of each option, see Options, above.

---

🦋 GM COMPOSITE 🦋

composite composites (combines) images to create new images, iykyk. base-image is the base image and change-image contains the changes. ouput-image is the result, and normally has the same dimensions as base-image.

The optional mask-image can be used to provide opacity information for change-image when it has none, or if you want a different mask. A mask image is typically grayscale and the same size as base-image. If mask-image is not grayscale, it is converted to grayscale and the resulting intensities are used as opacity information. 🌈

🌸 EXAMPLES 🌸

To composite an image of a cockatoo with a perch, use:

gm composite cockatoo.miff perch.ras composite.miff

To compute the difference between images in a series, use:

gm composite -compose difference series.2 series.1
          difference.miff

To composite an image of a cockatoo with a perch starting at location (100,150), use:

gm composite -geometry +100+150 cockatoo.miff
          perch.ras composite.miff

To tile a logo across your image of a cockatoo, use

gm convert +shade 30x60 cockatoo.miff mask.miff
gm composite -compose bumpmap -tile logo.png
          cockatoo.miff mask.miff composite.miff

To composite a red, green, and blue color plane into a single composite image, try

gm composite -compose CopyGreen green.png red.png
          red-green.png
gm composite -compose CopyBlue blue.png red-green.png
          gm composite.png

🧚 OPTIONS 🧚

Options are processed in command line order. Any option you specify on the command line remains in effect only for the image that follows. All options are reset to their default values after each image is read.

For a more detailed description of each option, see Options, above.

• -authenticate <string> — decrypt image with this password
• -background <color> — the background color
• -blue-primary <x>,<y> — blue chromaticity primary point
• -colors <value> — preferred number of colors in the image
• -colorspace <value> — the type of colorspace
• -comment <string> — annotate an image with a comment
• -compose <operator> — the type of image composition
• -compress <type> — the type of image compression
• -debug <events> — enable debug printout
• -define <key>{=<value>},... — add coder/decoder specific options
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image
• -depth <value> — depth of the image
• -displace <horizontal scale>x<vertical scale> — shift image pixels as defined by a displacement map
• -display <host:display[.screen]> — specifies the X server to contact
• -dispose <method> — GIF disposal method
• -dissolve <percent> — dissolve an image into another by the given percent
• -dither — apply Floyd/Steinberg error diffusion to the image
• -encoding <type> — specify the text encoding
• -endian <type> — specify endianness (MSB, LSB, or Native) of image
• -filter <type> — use this type of filter when resizing an image
• -font <name> — use this font when annotating the image with text
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — Specify dimension, offset, and resize options.
• -gravity <type> — direction primitive gravitates to when annotating the image.
• -green-primary <x>,<y> — green chromaticity primary point
• -help — print usage instructions
• -interlace <type> — the type of interlacing scheme
• -label <name> — assign a label to an image
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit
• -log <string> — Specify format for debug log
• -matte — store matte channel if the image has one
• -monitor — show progress indication
• -monochrome — transform the image to black and white
• -negate — replace every pixel with its complementary color
• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas
• -profile <filename> — add ICM, IPTC, or generic profile to image
• -quality <value> — JPEG/MIFF/PNG/TIFF compression level
• -recolor <matrix> — apply a color translation matrix to image channels
• -red-primary <x>,<y> — red chromaticity primary point
• -render — render vector operations
• -repage <width>x<height>+xoff+yoff[!] — Adjust image page offsets
• -resize <width>x<height>{%}{@}{!}{<}{>} — resize an image
• -rotate <degrees>{<}{>} — rotate the image
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors
• -scene <value> — set scene number
• -set <attribute> <value> — set an image attribute
• +set <attribute> — unset an image attribute
• -sharpen <radius>{x<sigma>} — sharpen the image
• -size <width>x<height>{+offset} — width and height of the image
• -stegano <offset> — hide watermark within an image
• -stereo — composite two images to create a stereo anaglyph
• -strip — remove all profiles and text attributes from the image
• -thumbnail <width>x<height>{%}{@}{!}{<}{>} — resize an image (quickly)
• -treedepth <value> — tree depth for the color reduction algorithm
• -trim — trim an image
• -type <type> — the image type
• -units <type> — the units of image resolution
• -unsharp <radius>{x<sigma>}{+<amount>}{+<threshold>} — sharpen the image with an unsharp mask operator
• -verbose — print detailed information about the image
• -version — print detailed GraphicsMagick version/build information
• -version — print compact GNU-style GraphicsMagick version information
• -watermark <brightness>x<saturation> — percent brightness and saturation of a watermark
• -white-point <x>,<y> — chromaticity white point
• -write <filename> — write an intermediate image [convert, composite]

For a more detailed description of each option, see Options, above.

---

🔥 GM CONJURE 🔥

The Magick scripting language (MSL) will primarily benefit those that want to accomplish custom image processing tasks but do not wish to program, or those that do not have access to a Perl interpreter or a compiler. The interpreter is called conjure and here is an example script, bestie, look at it, it's iconic:

<?xml version="1.0" encoding="UTF-8"?>
<image size="400x400" >
  <read filename="image.gif" />
  <get width="base-width" height="base-height" />
  <resize geometry="%[dimensions]" />
  <get width="width" height="height" />
  <print output=
    "Image sized from %[base-width]x%[base-height]
     to %[width]x%[height].\n" />
  <write filename="image.png" />
</image>

invoked with

gm conjure -dimensions 400x400 incantation.msl

All operations will closely follow the key/value pairs defined in PerlMagick, unless otherwise noted. ✨

✨ OPTIONS ✨

Options get processed in command line order. Periodt. Any option you drop on the command line stays in effect until it's explicitly changed — either by specifying the option again with a different effect, or by a statement in the scripting language swooping in and changing it. 💅

You can also define your own keyword/value pairs right there on the command line, and then your script can use that info when setting values by including %[keyword] in the string. Example bestie: if you included -dimensions 400x400 on the command line (like the illustration above shows), then any string containing %[dimensions] would get 400x400 substituted in. The %[string] can be the entire string, like geometry="%[dimensions]", or just a part of a string, like filename="%[basename].png". 🦋

The keyword can be literally any string except these reserved ones (in any upper, lower, or mixed case variant): debug, help, and verbose — their usage is described below, iykyk.

The value can be any string too. If either the keyword or the value contains white space or any symbols that have special meanings to your shell — like #, |, or % — enclose the string in quotation marks or use \ to escape the white space and special symbols. 🙈

Keywords and values are case dependent. Key, key, and KEY are three different keywords, no cap. 👑

For a more detailed description of each option, see Options, above.

• -debug <events> — enable debug printout 🔍
• -define <key>{=<value>},... — add coder/decoder specific options
• -help — print usage instructions
• -log <string> — Specify format for debug log
• -verbose — print detailed information about the image
• -version — print detailed GraphicsMagick version/build information
• -version — print compact GNU-style GraphicsMagick version information

---

🔮 MAGICK SCRIPTING LANGUAGE 🔮

The Magick Scripting Language (MSL) presently defines the following elements and their attributes — here's the whole lineup, bestie, don't skip a single one:

<image>

background, color, id, size

Define a new image object. </image> destroys it. Because of this, if you wish to reference multiple "subimages" (aka pages or layers), you can embed one image element inside of another. For example:

<image>
<read filename="input.png" />
<get width="base-width" height="base-height" />
<image height="base-height" width="base-width">
<image />
<write filename="output.mng" />
</image>

<image size="400x400" />

<group>

Define a new group of image objects. By default, images are only valid for the life of their <image> element.

<image>   -- creates the image
.....     -- do stuff with it
</image>  -- dispose of the image

However, in a group, all images in that group will stay around for the life of the group:

<group>                           -- start a group
    <image>                       -- create an image
    ....                          -- do stuff
    </image>                      -- NOOP
    <image>                       -- create another image
    ....                          -- do more stuff
    </image>                      -- NOOP
    <write filename="image.mng" />  -- output
</group>                          -- dispose of both images

<read>

filename

Read a new image from a disk file.

<read filename="image.gif" />

To read two images use

<read filename="image.gif" />
<read filename="image.png />

<write>

filename

Write the image(s) to disk, either as a single multiple-image file or multiple ones if necessary.

 <write filename=image.tiff" />

<get>

Get any attribute recognized by PerlMagick's GetAttribute() and stores it as an image attribute for later use. Currently only width and height are supported.

<get width="base-width" height="base-height" />
<print output="Image size is %[base-width]x%[base-height].\n" />

<set>

background, bordercolor, clip-mask, colorspace, density, magick, mattecolor, opacity. Set an attribute recognized by PerlMagick's GetAttribute().

<profile>

[profilename]

Read one or more IPTC, ICC or generic profiles from file and assign to image

<profile iptc="profile.iptc" generic="generic.dat" />

To remove a specified profile use ! as the filename, eg

<profile icm="!" iptc="profile.iptc" />

<border>

fill, geometry, height, width

<blur>

radius, sigma

<charcoal>

radius, sigma

<chop>

geometry, height, width, x, y

<crop>

geometry, height, width, x, y

<composite>

compose, geometry, gravity, image, x, y

<?xml version="1.0" encoding="UTF-8"?>
<group>
    <image id="image_01">
        <read filename="cloud3.gif"/>
        <resize geometry="250x90"/>
    </image>
    <image id="image_02">
        <read filename="cloud4.gif"/>
        <resize geometry="190x100"/>
    </image>
    <image>
        <read filename="background.jpg"/>
        <composite image="image_01" geometry="+740+470"/>
        <composite image="image_02" geometry="+390+415"/>
    </image>
    <write filename="result.png"/>
</group>

<despeckle>

<emboss>

radius, sigma

<enhance>

<equalize>

<edge>

radius

<flip>

<flop>

<frame>

fill, geometry, height, width, x, y, inner, outer

<flatten>

<get>

height, width

<gamma>

red, green, blue

<image>

background, color, id, size

<implode>

amount

<magnify>

<minify>

<medianfilter>

radius

<normalize>

<oilpaint>

radius

<print>

output

<profile>

[profilename]

<read>

<resize>

blur, filter, geometry, height, width

<roll>

geometry, x, y

<rotate>

degrees

<reducenoise>

radius

<sample>

geometry, height, width

<scale>

geometry, height, width

<sharpen>

radius, sigma

<shave>

geometry, height, width

<shear>

x, y

<solarize>

threshold

<spread>

radius

<stegano>

image

<stereo>

image

<swirl>

degrees

<texture>

image

<threshold>

threshold

<transparent>

color

<trim>

---

🎀 GM CONVERT 🎀

Convert converts an input file using one image format to an output file with a differing image format. Iconic behavior honestly. ✨ On top of that, various types of image processing can be performed on the converted image during the conversion process — so it's not just a glow-up, it's a whole makeover. 💖 Convert recognizes the image formats listed in gm(1).

✨ EXAMPLES ✨

okay bestie, time for the glow-up examples 💅 these are the little recipes that make gm convert feel like main character energy.

To make a thumbnail of a JPEG image, use:

gm convert -size 120x120 cockatoo.jpg -resize 120x120
        +profile "*" thumbnail.jpg

In this example, -size 120x120 gives a hint to the JPEG decoder that the image is going to be downscaled to 120x120, allowing it to run faster by avoiding returning full-resolution images to GraphicsMagick for the subsequent resizing operation. The -resize 120x120 specifies the desired dimensions of the output image. It will be scaled so its largest dimension is 120 pixels. The +profile "*" removes any ICM, EXIF, IPTC, or other profiles that might be present in the input and aren't needed in the thumbnail.

To convert a MIFF image of a cockatoo to a SUN raster image, use:

gm convert cockatoo.miff sun:cockatoo.ras

To convert a multi-page PostScript document to individual FAX pages, use:

gm convert -monochrome document.ps fax:page

To convert a TIFF image to a PostScript A4 page with the image in the lower left-hand corner, use:

gm convert -page A4+0+0 image.tiff document.ps

To convert a raw Gray image with a 128 byte header to a portable graymap, use:

gm convert -depth 8 -size 768x512+128 gray:raw
        image.pgm

In this example, "raw" is the input file. Its format is "gray" and it has the dimensions and number of header bytes specified by the -size option and the sample depth specified by the -depth option. The output file is "image.pgm". The suffix ".pgm" specifies its format.

To convert a Photo CD image to a TIFF image, use:

gm convert -size 1536x1024 img0009.pcd image.tiff
gm convert img0009.pcd[4] image.tiff

To create a visual image directory of all your JPEG images, use:

gm convert 'vid:*.jpg' directory.miff

To annotate an image with blue text using font 12x24 at position (100,100), use:

gm convert -font helvetica -fill blue
        -draw "text 100,100 Cockatoo"
        bird.jpg bird.miff

To tile a 640x480 image with a JPEG texture with bumps use:

gm convert -size 640x480 tile:bumps.jpg tiled.png

To surround an icon with an ornamental border to use with a web browser, use:

gm convert -mattecolor "#697B8F" -frame 6x6 bird.jpg
        icon.png

To create a MNG animation from a DNA molecule sequence, use:

gm convert -delay 20 dna.* dna.mng

---

🎀 OPTIONS 🎀

Options are processed in command line order. Any option you specify on the command line remains in effect for the set of images that follows, until the set is terminated by the appearance of any option or -noop. Some options only affect the decoding of images and others only the encoding. The latter can appear after the final group of input images.

For a more detailed description of each option, see Options, above. 💖

• -adjoin — join images into a single multi-image file.
• -affine <matrix> — the drawing transform matrix.
• -antialias — remove pixel aliasing.
• -append — append a set of images.
• -asc-cdl <spec> — apply ASC CDL color transform.
• -authenticate <string> — decrypt image with this password.
• -auto-orient — orient (rotate) image so it is upright.
• -average — average a set of images.
• -background <color> — the background color.
• -black-threshold red[,green][,blue][,opacity] — pixels below the threshold become black.
• -blue-primary <x>,<y> — blue chromaticity primary point.
• -blur <radius>{x<sigma>} — blur the image with a Gaussian operator.
• -border <width>x<height> — surround the image with a border of color.
• -bordercolor <color> — the border color.
• -box <color> — set the color of the annotation bounding box.
• -channel <type> — the type of channel.
• -charcoal <factor> — simulate a charcoal drawing.
• -chop <width>x<height>{+-}<x>{+-}<y>{%} — remove pixels from the interior of an image.
• -clip — apply the clipping path, if one is present.
• -coalesce — merge a sequence of images.
• -colorize <value> — colorize the image with the pen color.
• -colors <value> — preferred number of colors in the image.
• -colorspace <value> — the type of colorspace.
• -comment <string> — annotate an image with a comment.
• -compose <operator> — the type of image composition.
• -compress <type> — the type of image compression.
• -contrast — enhance or reduce the image contrast.
• -convolve <kernel> — convolve image with the specified convolution kernel.
• -crop <width>x<height>{+-}<x>{+-}<y>{%} — preferred size and location of the cropped image.
• -cycle <amount> — displace image colormap by amount.
• -debug <events> — enable debug printout.
• -deconstruct — break down an image sequence into constituent parts.
• -define <key>{=<value>},... — add coder/decoder specific options.
• -delay <1/100ths of a second> — display the next image after pausing.
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image.
• -depth <value> — depth of the image.
• -despeckle — reduce the speckles within an image.
• -display <host:display[.screen]> — specifies the X server to contact.
• -dispose <method> — GIF disposal method.
• -dither — apply Floyd/Steinberg error diffusion to the image.
• -draw <string> — annotate an image with one or more graphic primitives.
• -edge <radius> — detect edges within an image.
• -emboss <radius> — emboss an image.
• -encoding <type> — specify the text encoding.
• -endian <type> — specify endianness (MSB, LSB, or Native) of image.
• -enhance — apply a digital filter to enhance a noisy image.
• -equalize — perform histogram equalization to the image.
• -extent <width>x<height>{+-}<x>{+-}<y> — composite image on background color canvas image.
• -fill <color> — color to use when filling a graphic primitive.
• -filter <type> — use this type of filter when resizing an image.
• -flatten — flatten a sequence of images.
• -flip — create a "mirror image".
• -flop — create a "mirror image".
• -font <name> — use this font when annotating the image with text.
• -format <string> — output formatted image characteristics.
• -frame <width>x<height>+<outer bevel width>+<inner bevel width> — surround the image with an ornamental border.
• -fuzz <distance>{%} — colors within this Euclidean distance are considered equal.
• -gamma <value> — level of gamma correction.
• -gaussian <radius>{x<sigma>} — blur the image with a Gaussian operator.
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — Specify dimension, offset, and resize options.
• -gravity <type> — direction primitive gravitates to when annotating the image.
• -green-primary <x>,<y> — green chromaticity primary point.
• -hald-clut <clut> — apply a Hald CLUT to the image.
• -help — print usage instructions.
• -implode <factor> — implode image pixels about the center.
• -intent <type> — use this type of rendering intent when managing the image color.
• -interlace <type> — the type of interlacing scheme.
• -label <name> — assign a label to an image.
• -lat <width>x<height>{+-}<offset>{%} — perform local adaptive thresholding.
• -level <black_point>{,<gamma>}{,<white_point>}{%} — adjust the level of image contrast.
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit.
• -list <type> — the type of list.
• -log <string> — Specify format for debug log.
• -loop <iterations> — add Netscape loop extension to your GIF animation.
• -magnify — magnify the image.
• -map <filename> — choose a particular set of colors from this image.
• -mask <filename> — Specify a clipping mask.
• -matte — store matte channel if the image has one.
• -mattecolor <color> — specify the color to be used with the -frame option.
• -median <radius> — apply a median filter to the image.
• -minify <factor> — minify the image.
• -modulate brightness[,saturation[,hue]] — vary the brightness, saturation, and hue of an image.
• -monitor — show progress indication.
• -monochrome — transform the image to black and white.
• -morph <frames> — morphs an image sequence.
• -mosaic — create a mosaic from an image or an image sequence.
• -motion-blur <radius>{x<sigma>}{+angle} — Simulate motion blur.
• -negate — replace every pixel with its complementary color.
• -noise <radius|type> — add or reduce noise in an image.
• -noop — NOOP (no option).
• -normalize — transform image to span the full range of color values.
• -opaque <color> — change this color to the pen color within the image.
• -operator channel operator rvalue[%] — apply a mathematical, bitwise, or value operator to an image channel.
• -ordered-dither <channeltype> <NxN> — ordered dither the image.
• -orient <orientation> — Set the image orientation attribute.
• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas.
• -paint <radius> — simulate an oil painting.
• -pen <color> — (This option has been replaced by the -fill option).
• -ping — efficiently determine image characteristics.
• -pointsize <value> — pointsize of the PostScript, X11, or TrueType font.
• -preview <type> — image preview type.
• -process <command> — process a sequence of images using a process module.
• -profile <filename> — add ICM, IPTC, or generic profile to image.
• -quality <value> — JPEG/MIFF/PNG/TIFF compression level.
• -raise <width>x<height> — lighten or darken image edges.
• -random-threshold <channeltype> <LOWxHIGH> — random threshold the image.
• -recolor <matrix> — apply a color translation matrix to image channels.
• -map <filename> — choose a particular set of colors from this image.
• -red-primary <x>,<y> — red chromaticity primary point.
• -region <width>x<height>{+-}<x>{+-}<y> — apply options to a portion of the image.
• -render — render vector operations.
• -repage <width>x<height>+xoff+yoff[!] — Adjust image page offsets.
• -resample <horizontal>x<vertical> — resamples the image to exactly the horizontal and vertical resolution you type in, no cap ✨
• -resize <width>x<height>{%}{@}{!}{<}{>} — resizes an image 💅 the geometry suffixes are all there for you to play with, they're iconic
• -roll {+-}<x>{+-}<y> — rolls the image vertically or horizontally, like a little scroll wheel for your pixels 🌀
• -rotate <degrees>{<}{>} — rotates the image by however many degrees you want 🎀
• -sample <geometry> — scales the image using pixel sampling, simple and speedy fr fr
• -sampling-factor <horizontal_factor>x<vertical_factor> — sets the chroma subsampling factors 🌈
• -scale <geometry> — scales the image. that's it, that's the vibe 🦋
• -scene <value> — sets the scene number, lowkey important for multi-image files
• -set <attribute> <value> — sets an image attribute 💖
• +set <attribute> — unsets an image attribute (note the + bestie, don't come for me)
• -segment <cluster threshold>x<smoothing threshold> — segments an image into color clusters, very main character energy 👑
• -shade <azimuth>x<elevation> — shades the image using a distant light source, it's giving dramatic lighting 💡
• -sharpen <radius>{x<sigma>} — sharpens the image 🔥
• -shave <width>x<height>{%} — shaves pixels off the image edges, lil haircut for your pic ✂️
• -shear <x degrees>x<y degrees> — shears the image along the X or Y axis, slanted and serving
• -size <width>x<height>{+offset} — the width and height of the image 📐
• -solarize <factor> — negates all pixels above the threshold level, that blown-out artsy look 🌞
• -spread <amount> — displaces image pixels by a random amount, chaos but make it aesthetic 🎲
• -strip — removes all profiles and text attributes from the image 🧼
• -stroke <color> — the color to use when stroking a graphic primitive 🖌️
• -strokewidth <value> — sets the stroke width, thick or thin, your call
• -swirl <degrees> — swirls image pixels about the center, trippy bestie 🌀💜
• -texture <filename> — the name of the texture to tile onto the image background 🧵
• -threshold <value>{%} — thresholds the image, black and white decisions only ⚫⚪
• -thumbnail <width>x<height>{%}{@}{!}{<}{>} — resizes an image (quickly) ⚡ that's literally what the parenthetical says, fast fast
• -tile <filename> — tiles the image when filling a graphic primitive 🧱
• -transform — transforms the image ✨
• -transparent <color> — makes this color transparent within the image 🫥
• -treedepth <value> — tree depth for the color reduction algorithm 🌳
• -trim — trims an image, bye bye boring empty edges ✂️
• -type <type> — the image type 🏷️
• -units <type> — the units of image resolution 📏
• -unsharp <radius>{x<sigma>}{+<amount>}{+<threshold>} — sharpens the image with an unsharp mask operator, yes the name is confusing, no I don't make the rules 💅
• -use-pixmap — uses the pixmap 🖼️
• -verbose — prints detailed information about the image, yapper mode activated 🗣️
• -version — prints detailed GraphicsMagick version/build information 📋
• -version — prints compact GNU-style GraphicsMagick version information 📋✨ (yes there are two -version lines in here, that's how the page is, not me being messy iykyk)
• -view <string> — FlashPix viewing parameters
• -virtual-pixel <method> — specifies the contents of "virtual pixels" 👻
• -wave <amplitude>x<wavelength> — alters an image along a sine wave 🌊
• -white-point <x>,<y> — chromaticity white point ⚪
• -white-threshold red[,green][,blue][,opacity] — pixels above the threshold become white 🤍
• -write <filename> — writes an intermediate image [convert, composite] 💾

For a more detailed description of each option, see Options, above.

---

🖥️ GM DISPLAY 💖

Display is a machine architecture independent image processing and display program ✨ It can display an image on any workstation screen running an X server. Display can read and write many of the more popular image formats (e.g. JPEG, TIFF, PNM, Photo CD, etc.) — so many formats bestie, it's fluent 💅

With display, you can perform these functions on an image:

• load an image from a file 📂
• display the next image
• display the former image
• display a sequence of images as a slide show 🎞️
• write the image to a file
• print the image to a PostScript printer 🖨️
• delete the image file
• create a Visual Image Directory
• select the image to display by its thumbnail rather than name 🖼️
• undo last image transformation
• copy a region of the image
• paste a region to the image
• restore the image to its original size
• refresh the image
• half the image size
• double the image size
• resize the image
• crop the image ✂️
• cut the image
• flop image in the horizontal direction
• flip image in the vertical direction 🙃
• rotate the image 90 degrees clockwise
• rotate the image 90 degrees counter-clockwise
• rotate the image
• shear the image
• roll the image
• trim the image edges
• invert the colors of the image 🎨
• vary the color brightness
• vary the color saturation
• vary the image hue 🌈
• gamma correct the image
• sharpen the image contrast 🔥
• dull the image contrast
• perform histogram equalization on the image
• perform histogram normalization on the image
• negate the image colors
• convert the image to grayscale 🩶
• set the maximum number of unique colors in the image
• reduce the speckles within an image
• eliminate peak noise from an image
• detect edges within the image
• emboss an image
• segment the image by color 🎯
• simulate an oil painting 🖼️✨
• simulate a charcoal drawing ✏️
• annotate the image with text
• draw on the image
• edit an image pixel color
• edit the image matte information
• composite an image with another
• add a border to the image
• surround image with an ornamental border 🎀
• apply image processing techniques to a region of interest
• display information about the image
• zoom a portion of the image 🔍
• show a histogram of the image 📊
• display image to background of a window
• set user preferences
• display information about this program
• discard all images and exit program
• change the level of magnification
• display images specified by a World Wide Web (WWW) uniform resource locator (URL) 🌐

---

🌸 EXAMPLES 🌸

To scale an image of a cockatoo to exactly 640 pixels in width and 480 pixels in height and position the window at location (200,200), use:

gm display -geometry 640x480+200+200! cockatoo.miff

To display an image of a cockatoo without a border centered on a backdrop, use:

gm display +borderwidth -backdrop cockatoo.miff

To tile a slate texture onto the root window, use:

gm display -size 1280x1024 -window root slate.png

To display a visual image directory of all your JPEG images, use:

gm display 'vid:*.jpg'

To display a MAP image that is 640 pixels in width and 480 pixels in height with 256 colors, use:

gm display -size 640x480+256 cockatoo.map

To display an image of a cockatoo specified with a World Wide Web (WWW) uniform resource locator (URL), use:

gm display ftp://wizards.dupont.com/images/cockatoo.jpg

To display histogram of an image, use:

gm convert file.jpg HISTOGRAM:- | gm display -

🫶 and that's the display tea bestie, but we are so not done yet —

Options get processed in command line order, bestie — literally in the order you type them. Any option you drop on the command line stays in effect until you explicitly change it by specifying it again with a different effect. So if you want three images where the first has 32 colors, the second has an unlimited number of colors, and the third has only 16 — watch this:

gm display -colors 32 cockatoo.miff -noop duck.miff
             -colors 16 macaw.miff

Display options can live on the command line or in your X resources file. See X(1). And the command line always wins — options on the command line supersede values specified in your X resources file. 👑

For a more detailed description of each option, see Options, above.

• -authenticate <string> — decrypt the image with this password 🔐
• -backdrop — display the image centered on a backdrop
• -background <color> — the background color
• -border <width>x<height> — surround the image with a border of color
• -bordercolor <color> — the border color
• -borderwidth <geometry> — the border width
• -colormap <type> — define the colormap type
• -colors <value> — preferred number of colors in the image
• -colorspace <value> — the type of colorspace
• -comment <string> — annotate an image with a comment
• -compress <type> — the type of image compression
• -contrast — enhance or reduce the image contrast
• -crop <width>x<height>{+-}<x>{+-}<y>{%} — preferred size and location of the cropped image ✂️
• -debug <events> — enable debug printout
• -define <key>{=<value>},... — add coder/decoder specific options
• -delay <1/100ths of a second> — display the next image after pausing ⏱️
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image
• -depth <value> — depth of the image
• -despeckle — reduce the speckles within an image
• -display <host:display[.screen] — specifies the X server to contact
• -dispose <method> — GIF disposal method
• -dither — apply Floyd/Steinberg error diffusion to the image
• -edge <radius> — detect edges within an image
• -endian <type> — specify endianness (MSB, LSB, or Native) of image
• -enhance — apply a digital filter to enhance a noisy image
• -filter <type> — use this type of filter when resizing an image
• -flip — create a "mirror image" 🪞
• -flop — create a "mirror image" 🪞
• -font <name> — use this font when annotating the image with text
• -foreground <color> — define the foreground color
• -frame <width>x<height>+<outer bevel width>+<inner bevel width> — surround the image with an ornamental border 🖼️
• -gamma <value> — level of gamma correction
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — Specify dimension, offset, and resize options.
• -help — print usage instructions
• -iconGeometry <geometry> — specify the icon geometry
• -iconic — iconic animation
• -immutable — make image immutable
• -interlace <type> — the type of interlacing scheme
• -label <name> — assign a label to an image
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit
• -log <string> — Specify format for debug log
• -magnify <factor> — magnify the image 🔍
• -map <type> — display image using this type.
• -matte — store matte channel if the image has one
• -mattecolor <color> — specify the color to be used with the -frame option
• -monitor — show progress indication
• -monochrome — transform the image to black and white
• -name — name an image
• -negate — replace every pixel with its complementary color
• -noop — NOOP (no option) 💀
• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas
• +progress — disable progress monitor and busy cursor
• -quality <value> — JPEG/MIFF/PNG/TIFF compression level
• -raise <width>x<height> — lighten or darken image edges
• -remote — perform a X11 remote operation
• -roll {+-}<x>{+-}<y> — roll an image vertically or horizontally
• -rotate <degrees>{<}{>} — rotate the image 🔄
• -sample <geometry> — scale image using pixel sampling
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors
• -scenes <value-value> — range of image scene numbers to read
• -set <attribute> <value> — set an image attribute
• +set <attribute> — unset an image attribute
• -segment <cluster threshold>x<smoothing threshold> — segment an image
• -shared-memory — use shared memory
• -sharpen <radius>{x<sigma>} — sharpen the image
• -size <width>x<height>{+offset} — width and height of the image
• -text-font <name> — font for writing fixed-width text
• -texture <filename> — name of texture to tile onto the image background
• -title <string> — assign title to displayed image [animate, display, montage]
• -treedepth <value> — tree depth for the color reduction algorithm
• -trim — trim an image
• -type <type> — the image type
• -update <seconds> — detect when image file is modified and redisplay.
• -use-pixmap — use the pixmap
• -verbose — print detailed information about the image
• -version — print detailed GraphicsMagick version/build information
• -version — print compact GNU-style GraphicsMagick version information
• -visual <type> — animate images using this X visual type
• -window <id> — make image the background of a window
• -window-group — specify the window group
• -write <filename> — write the image to a file [display]

yes, -version really is listed twice up there — that's exactly how it's written, don't come for me 🙈

For a more detailed description of each option, see Options, above.

---

🖱️ MOUSE BUTTONS 🖱️

Okay so the effects of each button press is described below. Three buttons are required. If you have a two button mouse, button 1 and 3 are returned. Press ALT and button 3 to simulate button 2. iykyk.

• 1 — Press this button to map or unmap the Command widget. See the next section for more information about the Command widget.
• 2 — Press and drag to define a region of the image to magnify. 🔍
• 3 — Press and drag to choose from a select set of display commands.

Button 3 behaves differently if the image being displayed is a visual image directory. Choose a particular tile of the directory and press this button and drag to select a command from a pop-up menu. Choose from these menu items:

• Open
• Next
• Former
• Delete
• Update

If you choose Open, the image represented by the tile is displayed. To return to the visual image directory, choose Next from the Command widget (refer to Command Widget). Next and Former moves to the next or former image respectively. Choose Delete to delete a particular image tile. Finally, choose Update to synchronize all the image tiles with their respective images. See montage and miff for more details. ✨

---

🎛️ COMMAND WIDGET 🎛️

The Command widget lists a number of sub-menus and commands. They are:

• File
  • Open...
  • Next
  • Former
  • Select...
  • Save...
  • Print...
  • Delete...
  • Canvas...
  • Visual Directory...
  • Quit
• Edit
  • Undo
  • Redo
  • Cut
  • Copy
  • Paste
• View
  • Half Size
  • Original Size
  • Double Size
  • Resize...
  • Apply
  • Refresh
  • Restore
• Transform
  • Crop
  • Chop
  • Flop
  • Flip
  • Rotate Right
  • Rotate Left
  • Rotate...
  • Shear...
  • Roll...
  • Trim Edges
• Enhance
  • Hue...
  • Saturation...
  • Brightness...
  • Gamma...
  • Spiff...
  • Dull
  • Equalize
  • Normalize
  • Negate
  • GRAYscale
  • Quantize...
• Effects
  • Despeckle
  • Emboss
  • Reduce Noise
  • Add Noise
  • Sharpen...
  • Blur...
  • Threshold...
  • Edge Detect...
  • Spread...
  • Shade...
  • Raise...
  • Segment...
• F/X
  • Solarize...
  • Swirl...
  • Implode...
  • Wave...
  • Oil Paint...
  • Charcoal Draw...
• Image Edit
  • Annotate...
  • Draw...
  • Color...
  • Matte...
  • Composite...
  • Add Border...
  • Add Frame...
  • Comment...
  • Launch...
  • Region of Interest...
• Miscellany
  • Image Info
  • Zoom Image
  • Show Preview...
  • Show Histogram
  • Show Matte
  • Background...
  • Slide Show
  • Preferences...
• Help
  • Overview
  • Browse Documentation
  • About Display

Menu items with a indented triangle have a sub-menu. They are represented above as the indented items. To access a sub-menu item, move the pointer to the appropriate menu and press button 1 and drag. When you find the desired sub-menu item, release the button and the command is executed. Move the pointer away from the sub-menu if you decide not to execute a particular command. easy as that 💅

💅 KEYBOARD ACCELERATORS 💅

Okay so accelerators are one or two key presses that effect a particular command. The keyboard accelerators that display understands are, like, the whole lineup:

• Ctl+O — press to load an image from a file. 🖼️
• space — press to display the next image. ⏭️

If the image is a multi-paged document such as a PostScript document, you can skip ahead several pages by preceding this command with a number. For example, to display the fourth page beyond the current page, press 4space. 💅

• backspace — press to display the former image. ⏮️

If the image is a multi-paged document such as a PostScript document, you can skip behind several pages by preceding this command with a number. For example, to display the fourth page preceding the current page, press 4n. 🌸

• Ctl-S — press to save the image to a file. 💾
• Ctl-P — press to print the image to a PostScript printer. 🖨️
• Ctl-D — press to delete an image file. 🗑️
• Ctl-N — press to create a blank canvas. 🎨
• Ctl-Q — press to discard all images and exit program. 👋
• Ctl+Z — press to undo last image transformation. ↩️
• Ctl+R — press to redo last image transformation. ↪️
• Ctl-X — press to cut a region of the image. ✂️
• Ctl-C — press to copy a region of the image. 📋
• Ctl-V — press to paste a region to the image. 🫶
• < — press to halve the image size. 🔍
• . — press to return to the original image size. 🎀
• > — press to double the image size. 🔍
• % — press to resize the image to a width and height you specify. 📐
• Cmd-A — press to make any image transformations permanent. By default, any image size transformations are applied to the original image to create the image displayed on the X server. However, the transformations are not permanent (i.e. the original image does not change size — only the X image does). For example, if you press > the X image will appear to double in size, but the original image will in fact remain the same size. To force the original image to double in size, press > followed by Cmd-A. 👑
• @ — press to refresh the image window. 🔄
• C — press to crop the image. ✂️
• [ — press to chop the image. 🪓
• H — press to flop image in the horizontal direction. ↔️
• V — press to flip image in the vertical direction. ↕️
• / — press to rotate the image 90 degrees clockwise. 🔄
• \ — press to rotate the image 90 degrees counter-clockwise. 🔄
• * — press to rotate the image the number of degrees you specify. 🌀
• S — press to shear the image the number of degrees you specify. 📐
• R — press to roll the image. 🎞️
• T — press to trim the image edges. ✂️
• Shft-H — press to vary the color hue. 🌈
• Shft-S — press to vary the color saturation. 🎨
• Shft-L — press to vary the image brightness. ☀️
• Shft-G — press to gamma correct the image. ✨
• Shft-C — press to spiff up the image contrast. 🔥
• Shft-Z — press to dull the image contrast. 💀
• = — press to perform histogram equalization on the image. 📊
• Shft-N — press to perform histogram normalization on the image. 📈
• Shft-~ — press to negate the colors of the image. 🎨
• . — press to convert the image colors to gray. 🩶
• Shft-# — press to set the maximum number of unique colors in the image. 🌈
• F2 — press to reduce the speckles in an image. ✨
• F2 — press to emboss an image. 🪞
• F4 — press to eliminate peak noise from an image. 🧹
• F5 — press to add noise to an image. 🌪️
• F6 — press to sharpen an image. 🔪
• F7 — press to blur an image. 🌫️
• F8 — press to threshold the image. ⚫⚪
• F9 — press to detect edges within an image. 🕵️
• F10 — press to displace pixels by a random amount. 🎲
• F11 — press to shade the image using a distant light source. 💡
• F12 — press to lighten or darken image edges to create a 3-D effect. 🧊
• F13 — press to segment the image by color. 🎨
• Meta-S — press to swirl image pixels about the center. 🌀
• Meta-I — press to implode image pixels about the center. 💥
• Meta-W — press to alter an image along a sine wave. 🌊
• Meta-P — press to simulate an oil painting. 🖼️
• Meta-C — press to simulate a charcoal drawing. ✏️
• Alt-X — press to composite the image with another. 🧩
• Alt-A — press to annotate the image with text. 📝
• Alt-D — press to draw a line on the image. 📏
• Alt-P — press to edit an image pixel color. 🎯
• Alt-M — press to edit the image matte information. 🎭
• Alt-X — press to composite the image with another. 🧩
• Alt-A — press to add a border to the image. 🖼️
• Alt-F — press to add a ornamental frame to the image. 🖼️
• Alt-Shft-! — press to add an image comment. 💬
• Ctl-A — press to apply image processing techniques to a region of interest. 🎯
• Shft-? — press to display information about the image. ℹ️
• Shft-+ — press to map the zoom image window. 🔍
• Shft-P — press to preview an image enhancement, effect, or f/x. 👀
• F1 — press to display helpful information about the display utility. 🆘
• Find — press to browse documentation about GraphicsMagick. 📚
• 1-9 — press to change the level of magnification. 🔎

Use the arrow keys to move the image one pixel up, down, left, or right within the magnify window. Be sure to first map the magnify window by pressing button 2. 🖱️

Press ALT and one of the arrow keys to trim off one pixel from any side of the image. ✂️

✨ X RESOURCES ✨

okay so bestie, listen — you can put your display options on the command line or in your X resource file. And if you do both? Command line wins, every time, it supersedes whatever you put in your X resource file. Peep X(1) if you wanna know more about X resources, iykyk 💅

Most display options have a matching X resource. Plus, display also uses these extra ones down here:

• background (class Background) — the preferred color for the Image window background. Default is #ccc. 🎀
• borderColor (class BorderColor) — the preferred color for the Image window border. Default is #ccc. ✨
• borderWidth (class BorderWidth) — how wide the image window border is, in pixels. Default is 2. 📏
• browseCommand (class browseCommand) — the name of the preferred browser for when you're looking at GraphicsMagick documentation. Default is xdg-open %s, which just rolls with your system default. 🌐
• confirmExit (class ConfirmExit) — Display pops up a dialog box to confirm you actually wanna exit the program when you exit the program. Set this resource to False to dip out with no confirmation. 🙈
• displayGamma (class DisplayGamma) — specifies the gamma of the X server. You can hit the red, green, and blue channels separately with a gamma value list separated by slashes (i.e. 1.7/2.3/1.2). Default is 2.2. 🌈
• displayWarnings (class DisplayWarnings) — Display pops up a dialog box anytime a warning message happens. Set this resource to False to straight up ignore warning messages. 💀
• font (class FontList) — the name of the preferred font for normal formatted text. Default is 14 point Helvetica. 🖋️
• font[1-9] (class Font[1-9]) — the name of the preferred font for annotating the image window with text. The default fonts are fixed, variable, 5x8, 6x10, 7x13bold, 8x13bold, 9x15bold, 10x20, and 12x24. 🔤
• foreground (class Foreground) — the preferred color for text inside the image window. Default is black. 🖤
• gammaCorrect (class gammaCorrect) — if true, this resource will lighten or darken an image of known gamma to match the gamma of the display (check the displayGamma resource, bestie). Default is True. ⭐
• geometry (class Geometry) — the preferred size and position of the image window. Not necessarily obeyed by all window managers, ngl. Offsets, if present, are handled in X(1) style. A negative x offset is measured from the right edge of the screen to the right edge of the icon, and a negative y offset is measured from the bottom edge of the screen to the bottom edge of the icon. 📐
• iconGeometry (class IconGeometry) — the preferred size and position of the application when it's iconified. Not necessarily obeyed by all window managers. Offsets, if present, are handled the same way as in class Geometry. 🪟
• iconic (class Iconic) — this resource says you'd prefer the application's windows to initially not be visible, like the windows had been immediately iconified by you. Window managers may choose not to honor the application's request, so… vibe check 🫶
• magnify (class Magnify) — specifies an integral factor by which the image should be enlarged. Default is 3. This value only affects the magnification window, which gets invoked with button number 3 after the image is displayed. 🔍
• matteColor (class MatteColor) — the color of windows. It's used for the backgrounds of windows, menus, and notices. A 3D effect is achieved by using highlight and shadow colors derived from this color. Default value: #697B8F. 🎨
• name (class Name) — this resource specifies the name under which resources for the application should be found. Super useful in shell aliases to tell apart different invocations of an application without resorting to creating links to alter the executable file name. The default is the application name. 👑
• pen[1-9] (class Pen[1-9]) — the color of the preferred font to use when annotating the image window with text. The default colors are black, blue, green, cyan, gray, red, magenta, yellow, and white. 🖍️
• printCommand (class PrintCommand) — this command is executed whenever Print is issued. In general, it's the command to print PostScript to your printer. Default value: lp -c -s %i. 🖨️
• sharedMemory (class SharedMemory) — whether display should attempt to use shared memory for pixmaps. GraphicsMagick must be compiled with shared memory support, and the display must support the MIT-SHM extension. Otherwise, this resource is ignored. Default is True. 🧠
• textFont (class textFont) — the name of the preferred font for fixed (typewriter style) formatted text. Default is 14 point Courier. ✍️
• title (class Title) — this resource specifies the title to be used for the image window. This info is sometimes used by a window manager to give a header identifying the window. The default is the image file name. 🏷️
• undoCache (class UndoCache) — specifies, in mega-bytes, the amount of memory in the undo edit cache. Every time you modify the image it gets saved in the undo edit cache as long as memory is available. You can subsequently undo one or more of these transformations. The default is 16 Megabytes. 😭
• usePixmap (class UsePixmap) — images are maintained as a XImage by default. Set this resource to True to utilize a server Pixmap instead. This option is useful if your image exceeds the dimensions of your server screen and you intend to pan the image. Panning is much faster with Pixmaps than with a XImage. Pixmaps are considered a precious resource, so use them with discretion, bestie.

fr fr, to set the geometry of the Magnify or Pan or window, use the geometry resource. For example, to set the Pan window geometry to 256x256, use:

gm display.pan.geometry: 256x256

---

🦋 IMAGE LOADING 🦋

To select an image to display, pick Open of the File sub-menu from the Command widget. A file browser shows up. To choose a particular image file, move the pointer to the filename and press any button. The filename gets copied to the text window. Next, press Open or hit the RETURN key. Alternatively, you can just type the image file name directly into the text window. To descend directories, choose a directory name and press the button twice quickly. A scrollbar lets a big list of filenames be moved through the viewing area if it exceeds the size of the list area. 📂

You can trim the list of file names by using shell globbing characters. For example, type *.jpg to list only files that end with .jpg. 🌸

To select your image from the X server screen instead of from a file, choose Grab of the Open widget. 📸

---

🌈 VISUAL IMAGE DIRECTORY 🌈

To create a Visual Image Directory, choose Visual Directory of the File sub-menu from the Command widget. A file browser shows up. To create a Visual Image Directory from all the images in the current directory, press Directory or press the RETURN key. Alternatively, you can select a set of image names by using shell globbing characters. For example, type *.jpg to include only files that end with .jpg. To descend directories, choose a directory name and press the button twice quickly. A scrollbar lets a big list of filenames be moved through the viewing area if it exceeds the size of the list area. 🗂️

After you select a set of files, they get turned into thumbnails and tiled onto a single image. Now move the pointer to a particular thumbnail and press button 3 and drag. Finally, select Open. The image represented by the thumbnail is displayed at its full size. Choose Next from the File sub-menu of the Command widget to return to the Visual Image Directory. 🖼️

---

✂️ IMAGE CUTTING ✂️

⚠️ okay important: cut information for the image window is not retained for colormapped X server visuals (e.g. StaticColor, StaticColor, GRAYScale, PseudoColor). Correct cutting behavior may require a TrueColor or DirectColor visual or a Standard Colormap.

To begin, press choose Cut of the Edit sub-menu from the Command widget. Alternatively, press F3 in the image window. 🎀

A small window appears showing where the cursor is in the image window. You are now in cut mode. In cut mode, the Command widget has these options:

Help
Dismiss

To define a cut region, press button 1 and drag. The cut region is defined by a highlighted rectangle that expands or contracts as it follows the pointer. Once you're satisfied with the cut region, release the button. You are now in rectify mode. In rectify mode, the Command widget has these options:

Cut
Help
Dismiss

You can make adjustments by moving the pointer to one of the cut rectangle corners, pressing a button, and dragging. Finally, press Cut to commit your copy region. To exit without cutting the image, press Dismiss. 💅

---

💖 IMAGE COPYING 💖

To begin, press choose Copy of the Edit sub-menu from the Command widget. Alternatively, press F4 in the image window. 🦋

A small window appears showing where the cursor is in the image window. You are now in copy mode. In copy mode, the Command widget has these options:

Help
Dismiss

To define a copy region, press button 1 and drag. The copy region is defined by a highlighted rectangle that expands or contracts as it follows the pointer. Once you're satisfied with the copy region, release the button. You are now in rectify mode. In rectify mode, the Command widget has these options:

Copy
Help
Dismiss

You can make adjustments by moving the pointer to one of the copy rectangle corners, pressing a button, and dragging. Finally, press Copy to commit your copy region. To exit without copying the image, press Dismiss. 👑

🎀 IMAGE PASTING 🎀

okay so to start, bestie, pick Paste from the Edit sub-menu in the Command widget. Or just hit F5 while you're in the image window. ✨

A tiny window pops up showing where your cursor is in the image window. You're in Paste mode now. Wanna bail immediately? Press Dismiss. In Paste mode, the Command widget has these options:

• Operators — opens the composite operator sub-menu.
• Help — the help option.
• Dismiss — exits Paste mode.

Inside Operators, you've got all of these:

• over
• in
• out
• atop
• xor
• plus
• minus
• add
• subtract
• difference
• multiply
• bumpmap
• replace

Pick a composite operator from the Operators sub-menu of the Command widget. Here's what each one does, fr fr. Just so we're clear: image window is the image currently displayed on your X server, and image is the image obtained with the File Browser widget. 💖

• over — the result is the union of the two image shapes, with image obscuring image window in the region of overlap.
• in — the result is simply image cut by the shape of image window. None of the image data of image window is in the result.
• out — the resulting image is image with the shape of image window cut out.
• atop — the result is the same shape as image window, with image obscuring image window where the image shapes overlap. Note this differs from over because the portion of image outside image window's shape does not appear in the result.
• xor — the result is the image data from both image and image window that is outside the overlap region. The overlap region is blank.
• plus — the result is just the sum of the image data. Output values are cropped to the maximum value (no overflow). This operation is independent of the matte channels.
• minus — the result of image - image window, with underflow cropped to zero. The matte channel is ignored (set to opaque, full coverage).
• add — the result of image + image window, with overflow wrapping around (mod MaxRGB+1).
• subtract — the result of image - image window, with underflow wrapping around (mod MaxRGB+1). The add and subtract operators can be used to perform reversible transformations.
• difference — the result of abs(image - image window). This is useful for comparing two very similar images.
• multiply — the result of image * image window. This is useful for the creation of drop-shadows.
• bumpmap — the result of image window shaded by window.
• replace — the resulting image is image window replaced with image. Here the matte information is ignored.

The image compositor requires a matte, or alpha channel in the image for some operations. This extra channel usually defines a mask which represents a sort of a cookie-cutter for the image. This is the case when matte is 255 (full coverage) for pixels inside the shape, zero outside, and between zero and 255 on the boundary. If image does not have a matte channel, it is initialized with 0 for any pixel matching in color to pixel location (0,0), otherwise 255. See Matte Editing for a method of defining a matte channel.

⚠️ heads up, bestie ⚠️ Note that matte information for image window is not retained for colormapped X server visuals (e.g. StaticColor, StaticColor, GrayScale, PseudoColor). Correct compositing behavior may require a TrueColor or DirectColor visual or a Standard Colormap.

Choosing a composite operator is optional. The default operator is replace. However, you must choose a location to composite your image and press button 1. Press and hold the button before releasing and an outline of the image will appear to help you identify your location. 🫶

The actual colors of the pasted image is saved. However, the color that appears in image window may be different. For example, on a monochrome screen image window will appear black or white even though your pasted image may have many colors. If the image is saved to a file it is written with the correct colors. To assure the correct colors are saved in the final image, any PseudoClass image is promoted to DirectClass. To force a PseudoClass image to remain PseudoClass, use -colors. 🌈

---

💅 IMAGE CROPPING 💅

To begin, choose Crop of the Transform submenu from the Command widget. Alternatively, press C in the image window.

A small window appears showing the location of the cursor in the image window. You are now in crop mode. In crop mode, the Command widget has these options:

• Help — the help option.
• Dismiss — exits crop mode.

To define a cropping region, press button 1 and drag. The cropping region is defined by a highlighted rectangle that expands or contracts as it follows the pointer. Once you are satisfied with the cropping region, release the button. You are now in rectify mode. In rectify mode, the Command widget has these options:

• Crop — commits the cropping region.
• Help — the help option.
• Dismiss — exits without cropping.

You can make adjustments by moving the pointer to one of the cropping rectangle corners, pressing a button, and dragging. Finally, press Crop to commit your cropping region. To exit without cropping the image, press Dismiss. ✂️

---

🦋 IMAGE CHOPPING 🦋

An image is chopped interactively. There is no command line argument to chop an image. To begin, choose Chop of the Transform sub-menu from the Command widget. Alternatively, press [ in the Image window.

You are now in Chop mode. To exit immediately, press Dismiss. In Chop mode, the Command widget has these options:

• Direction — sub-menu with:
  • horizontal
  • vertical
• Help — the help option.
• Dismiss — exits Chop mode.

If you choose the horizontal direction (this is the default), the area of the image between the two horizontal endpoints of the chop line is removed. Otherwise, the area of the image between the two vertical endpoints of the chop line is removed.

Select a location within the image window to begin your chop, press and hold any button. Next, move the pointer to another location in the image. As you move a line will connect the initial location and the pointer. When you release the button, the area within the image to chop is determined by which direction you choose from the Command widget.

To cancel the image chopping, move the pointer back to the starting point of the line and release the button. 🍓

---

🌈 IMAGE ROTATION 🌈

Press the / key to rotate the image 90 degrees or </kbd> to rotate -90 degrees. To interactively choose the degree of rotation, choose Rotate... of the Transform submenu from the Command Widget. Alternatively, press * in the image window.

A small horizontal line is drawn next to the pointer. You are now in rotate mode. To exit immediately, press Dismiss. In rotate mode, the Command widget has these options:

• Pixel Color — sub-menu with:
  • black
  • blue
  • cyan
  • green
  • gray
  • red
  • magenta
  • yellow
  • white
  • Browser...
• Direction — sub-menu with:
  • horizontal
  • vertical
• Help — the help option.
• Dismiss — exits rotate mode.

Choose a background color from the Pixel Color sub-menu. Additional background colors can be specified with the color browser. You can change the menu colors by setting the X resources pen1 through pen9.

If you choose the color browser and press Grab, you can select the background color by moving the pointer to the desired color on the screen and press any button.

Choose a point in the image window and press this button and hold. Next, move the pointer to another location in the image. As you move a line connects the initial location and the pointer. When you release the button, the degree of image rotation is determined by the slope of the line you just drew. The slope is relative to the direction you choose from the Direction sub-menu of the Command widget.

To cancel the image rotation, move the pointer back to the starting point of the line and release the button. 🌀

🎨 IMAGE ANNOTATION 🎨

Okay so annotating an image is a whole interactive thing, bestie — there's literally no command line argument to annotate an image. To start, pick Annotate from the Image Edit sub-menu in the Command widget. Or, honestly easier, just press a in the image window. 💅

A lil' window pops up showing you where your cursor is sitting in the image window, and boom — you're in annotate mode now. Wanna bail instantly? Press Dismiss. While you're in annotate mode, the Command widget hands you these options:

• Font Name — pick your typeface vibe:
  • fixed
  • variable
  • 5x8
  • 6x10
  • 7x13bold
  • 8x13bold
  • 9x15bold
  • 10x20
  • 12x24
  • Browser...
• Font Color — the ink, the moment:
  • black
  • blue
  • cyan
  • green
  • gray
  • red
  • magenta
  • yellow
  • white
  • transparent
  • Browser...
• Box Color — same energy, but for the box:
  • black
  • blue
  • cyan
  • green
  • gray
  • red
  • magenta
  • yellow
  • white
  • transparent
  • Browser...
• Rotate Text — tilt it, queen:
  • -90
  • -45
  • -30
  • 0
  • 30
  • 45
  • 90
  • 180
  • Dialog...
• Help — when you need a minute.
• Dismiss — exit, stage left.

Pick a font name from the Font Name sub-menu. Want more font names? The font browser has you — additional ones can be specified there. And yes, you can rename those menu entries by setting the X resources font1 through font9. ✨

Pick a font color from the Font Color sub-menu, and again, extra colors can be specified with the color browser. You can change the menu colors by setting the X resources pen1 through pen9.

If you open the color browser and press Grab, you can pick the font color by moving the pointer to whatever color on screen you're obsessed with and pressing any button. 🌈

Now, rotating the text: choose Rotate Text from the menu and select an angle. Typically you'll only want to rotate one line of text at a time — depending on the angle you choose, subsequent lines may end up overwriting each other. 😭 So like, keep it cute and keep it single-line.

Choosing a font and its color is optional, no pressure. The default font is fixed and the default color is black. But — and this is the mandatory part — you must choose a location to begin entering text and press a button. An underscore character appears at the location of the pointer, and the cursor changes to a pencil to show you're in text mode. To exit immediately, press Dismiss.

In text mode, any key presses display the character at the location of the underscore and advance the underscore cursor. Type out your text, and once you're done press Apply to finish your image annotation. Messed up? Press BACK SPACE to correct errors. Wanna nuke a whole line of text? Press DELETE. And any text that exceeds the boundaries of the image window is automatically continued onto the next line — she just keeps going, no cap. 🦋

The actual color you request for the font is saved in the image. However, the color that appears in your Image window may be different. For example, on a monochrome screen the text will appear black or white even if you choose red as the font color. But the image saved to a file with -write is written with red lettering. 🔥

To assure the correct color text in the final image, any PseudoClass image is promoted to DirectClass (see miff(5)). To force a PseudoClass image to remain PseudoClass, use -colors.

---

🧩 IMAGE COMPOSITING 🧩

An image composite is created interactively too — and again, there's no command line argument to composite an image. To begin, choose Composite of the Image Edit from the Command widget. Alternatively, press x in the Image window. 🫶

First, a popup window shows up asking you to enter an image name. Press Composite, Grab, or just type a file name. Press Cancel if you decide you don't wanna create a composite image after all. When you choose Grab, move the pointer to the desired window and press any button.

If the Composite image doesn't have any matte information, you get informed about it and the file browser appears again. Enter the name of a mask image. That image is typically grayscale and the same size as the composite image. If it's not grayscale, it gets converted to grayscale and the resulting intensities are used as matte information.

A small window appears showing the location of the cursor in the image window — congrats, you're in composite mode. To exit immediately, press Dismiss. In composite mode, the Command widget has these options:

• Operators — the whole menu of blend behavior:
  • over
  • in
  • out
  • atop
  • xor
  • plus
  • minus
  • add
  • subtract
  • difference
  • bumpmap
  • replace
• Blend
• Displace
• Help
• Dismiss

Choose a composite operation from the Operators sub-menu of the Command widget. How each operator behaves is described below — for the record, image window is the image currently displayed on your X server, and image is the image obtained. 💖

• over — The result is the union of the two image shapes, with image obscuring image window in the region of overlap.
• in — The result is simply image cut by the shape of image window. None of the image data of image window is in the result.
• out — The resulting image is image with the shape of image window cut out.
• atop — The result is the same shape as image window, with image obscuring image window where the image shapes overlap. Note this differs from over because the portion of image outside image window's shape does not appear in the result.
• xor — The result is the image data from both image and image window that is outside the overlap region. The overlap region is blank.
• plus — The result is just the sum of the image data. Output values are cropped to 255 (no overflow). This operation is independent of the matte channels.
• minus — The result of image - image window, with underflow cropped to zero. The matte channel is ignored (set to 255, full coverage).
• add — The result of image + image window, with overflow wrapping around (mod 256).
• subtract — The result of image - image window, with underflow wrapping around (mod 256). The add and subtract operators can be used to perform reversible transformations.
• difference — The result of abs(image - image window). This is useful for comparing two very similar images.
• bumpmap — The result of image window shaded by window.
• replace — The resulting image is image window replaced with image. Here the matte information is ignored.

The image compositor requires a matte, or alpha channel in the image for some operations. This extra channel usually defines a mask which represents a sort of a cookie-cutter for the image. That's the case when matte is 255 (full coverage) for pixels inside the shape, zero outside, and between zero and 255 on the boundary. If image doesn't have a matte channel, it's initialized with 0 for any pixel matching in color to pixel location (0,0), otherwise 255. See Matte Editing for a method of defining a matte channel. 👑

If you choose blend, the composite operator becomes over. The image matte channel percent transparency is initialized to factor. The image window is initialized to (100-factor). Where factor is the value you specify in the Dialog widget.

Displace shifts the image pixels as defined by a displacement map. With this option, image is used as a displacement map. Black, within the displacement map, is a maximum positive displacement. White is a maximum negative displacement and middle gray is neutral. The displacement is scaled to determine the pixel shift. By default, the displacement applies in both the horizontal and vertical directions. However, if you specify mask, image is the horizontal X displacement and mask the vertical Y displacement. 🌸

Important heads-up: matte information for image window is not retained for colormapped X server visuals (e.g. StaticColor, StaticColor, GrayScale, PseudoColor). Correct compositing behavior may require a TrueColor or DirectColor visual or a Standard Colormap.

Choosing a composite operator is optional — the default operator is replace. But you must choose a location to composite your image and press button 1. Press and hold the button before releasing and an outline of the image will appear to help you identify your location. 🎀

The actual colors of the composite image is saved. However, the color that appears in image window may be different. For example, on a monochrome screen Image window will appear black or white even though your composited image may have many colors. If the image is saved to a file it is written with the correct colors. To assure the correct colors are saved in the final image, any PseudoClass image is promoted to DirectClass (see miff). To force a PseudoClass image to remain PseudoClass, use -colors. 💜

🎨 COLOR EDITING ✨

Okay bestie, so recoloring a set of pixels? It's fully interactive, no cap. There is no command line argument to edit a pixel — you gotta do it with your actual hands.

To start: pick Color from the Image Edit submenu of the Command widget. Or honestly? Just press c in the image window. Same energy.

A little window pops up showing where your cursor is in the image window, and boom — you are now in color edit mode 🫶. Wanna bounce immediately? Hit Dismiss.

While you're in color edit mode, the Command widget is serving these options:

🌈 Method

• point
• replace
• floodfill
• reset

🌸 Pixel Color

• black
• blue
• cyan
• green
• gray
• red
• magenta
• yellow
• white
• Browser...

🎀 Border Color

• black
• blue
• cyan
• green
• gray
• red
• magenta
• yellow
• white
• Browser...

💜 Fuzz

• 0
• 2
• 4
• 8
• 16
• Dialog...

Plus, sitting right there in the widget:

• Undo
• Help
• Dismiss

---

Pick your color editing method from the Method sub-menu of the Command widget. Here's the tea ☕:

• point — recolors any pixel you select with the pointer, and keeps doing it unless the button is released.
• replace — recolors any pixel that matches the color of the pixel you select with a button press.
• floodfill — recolors any pixel that matches the color of the pixel you pick with a button press and is a neighbor.
• filltoborder — meanwhile this one changes the matte value of any neighbor pixel that is not the border color.
• And finally reset — changes the entire image to the designated color. Whole thing. Main character move.

Next up, choose a pixel color from the Pixel Color sub-menu. Want extra colors beyond the menu? That's what the color browser is for 💅. And psst — you can change the menu colors by setting the X resources pen1 through pen9.

Now press button 1 to select a pixel inside the Image window and change its color. More pixels may get recolored depending on the method you chose — and you can get additional pixels by increasing the Delta value.

If the Magnify widget is mapped, it's genuinely so helpful for positioning your pointer inside the image (refer to button 2). Alternatively you can select a pixel to recolor from inside the Magnify widget: move the pointer to the Magnify widget, position the pixel with the cursor control keys, then press a button to recolor the selected pixel (or pixels). Slay 🦋

Okay important vibe check: the actual color you request for the pixels is what's saved in the image. But the color that shows up in your Image window might be different 😭. Like, on a monochrome screen the pixel will appear black or white even if you chose red. BUT — the image saved to a file with -write is written with red pixels. To assure the correct color text in the final image, any PseudoClass image is promoted to DirectClass. And to force a PseudoClass image to remain PseudoClass, use -colors.

---

🫥 MATTE EDITING 🎭

Matte information inside an image is lowkey useful for some operations, like image compositing. This extra channel usually defines a mask — think of it as a cookie-cutter for the image 🍪. That's the case when matte is 255 (full coverage) for pixels inside the shape, zero outside, and between zero and 255 on the boundary.

Setting matte info in an image is done interactively. There is no command line argument to edit a pixel. To begin, choose Matte of the Image Edit sub-menu from the Command widget.

Alternatively? Press m in the image window. Easy.

A small window appears showing the location of the cursor in the image window. You are now in matte edit mode. To exit immediately, press Dismiss. In matte edit mode, the Command widget has these options:

🌈 Method

• point
• replace
• floodfill
• reset

🎀 Border Color

• black
• blue
• cyan
• green
• gray
• red
• magenta
• yellow
• white
• Browser...

💜 Fuzz

• 0
• 2
• 4
• 8
• 16
• Dialog...

✨ And also

• Matte
• Undo
• Help
• Dismiss

---

Choose a matte editing method from the Method sub-menu of the Command widget. The breakdown:

• point — changes the matte value of any pixel selected with the pointer, until the button is released.
• replace — changes the matte value of any pixel that matches the color of the pixel you select with a button press.
• floodfill — changes the matte value of any pixel that matches the color of the pixel you select with a button press and is a neighbor.
• filltoborder — meanwhile this one recolors any neighbor pixel that is not the border color.
• reset — finally, changes the entire image to the designated matte value.

Choose Matte Value and a dialog appears asking for a matte value. Enter a value between 0 and 255. That value gets assigned as the matte value of the selected pixel or pixels 💖

Now press any button to select a pixel within the Image window to change its matte value. You can change the matte value of additional pixels by increasing the Delta value. The Delta value is first added then subtracted from the red, green, and blue of the target color — and any pixels within that range also have their matte value updated.

If the Magnify widget is mapped, it's super helpful for positioning your pointer within the image (refer to button 2). Alternatively you can select a pixel to change the matte value from inside the Magnify widget. Move the pointer to the Magnify widget and position the pixel with the cursor control keys. Finally, press a button to change the matte value of the selected pixel (or pixels).

Big caveat energy 🚨: matte information is only valid in a DirectClass image. Therefore, any PseudoClass image is promoted to DirectClass. Note that matte information for PseudoClass is not retained for colormapped X server visuals (e.g. StaticColor, StaticColor, GrayScale, PseudoColor) unless you immediately save your image to a file (refer to Write). Correct matte editing behavior may require a TrueColor or DirectColor visual or a Standard Colormap.

---

🖌️ IMAGE DRAWING 🎨

An image gets drawn upon interactively. There is no command line argument to draw on an image. To begin, choose Draw of the Image Edit sub-menu from the Command widget. Alternatively, press d in the image window ✍️

The cursor changes to a crosshair to let you know you're in draw mode. To exit immediately, press Dismiss. In draw mode, the Command widget has these options:

🔷 Primitive

• point
• line
• rectangle
• fill rectangle
• circle
• fill circle
• ellipse
• fill ellipse
• polygon
• fill polygon

🌈 Color

• black
• blue
• cyan
• green
• gray
• red
• magenta
• yellow
• white
• transparent
• Browser...

🎀 Stipple

• Brick
• Diagonal
• Scales
• Vertical
• Wavy
• Translucent
• Opaque
• Open...

💜 Width

• 1
• 2
• 4
• 8
• 16
• Dialog...

Plus a little trio at the bottom:

• Undo
• Help
• Dismiss

---

Choose a drawing primitive from the Primitive sub-menu. Iconic, honestly.

Next, choose a color from the Color sub-menu. Additional colors can be specified with the color browser. You can change the menu colors by setting the X resources pen1 through pen9. The transparent color updates the image matte channel and is useful for image compositing ✨

If you choose the color browser and press Grab, you can select the primitive color by moving the pointer to the desired color on the screen and pressing any button. (Yes — the transparent color updates the image matte channel and is useful for image compositing. Told you twice, because it matters 💅)

Choose a stipple, if appropriate, from the Stipple sub-menu. Additional stipples can be specified with the file browser. Stipples obtained from the file browser must be on disk in the X11 bitmap format 🧚

Choose a line width, if appropriate, from the Width sub-menu. Want a specific width? Select the Dialog widget.

Now the fun part: choose a point in the image window, press button 1, and hold. Next, move the pointer to another location in the image. As you move, a line connects the initial location and the pointer — very satisfying, 10/10. When you release the button, the image is updated with the primitive you just drew. For polygons, the image is updated when you press and release the button without moving the pointer.

To cancel image drawing, move the pointer back to the starting point of the line and release the button. Abort mission 🫡

---

🔍 REGION OF INTEREST 🌸

To begin, press choose Region of Interest of the Pixel Transform sub-menu from the Command widget. Alternatively, press R in the image window.

A small window appears showing the location of the cursor in the image window. You are now in region of interest mode. In region of interest mode, the Command widget has these options:

• Help
• Dismiss

To define a region of interest, press button 1 and drag. The region of interest is defined by a highlighted rectangle that expands or contracts as it follows the pointer. Once you're satisfied with your region, release the button. You are now in apply mode 🎉

In apply mode, the Command widget has these options:

📁 File

• Save...
• Print...

✏️ Edit

• Undo
• Redo

🔄 Transform

• Flip
• Flop
• Rotate Right
• Rotate Left
• Enhance
• Hue...
• Saturation...
• Brightness...
• Gamma...
• Spiff
• Dull
• Equalize
• Normalize
• Negate
• GRAYscale
• Quantize...

🌪️ Effects

• Despeckle
• Emboss
• Reduce Noise
• Add Noise
• Sharpen...
• Blur...
• Threshold...
• Edge Detect...
• Spread...
• Shade...
• Raise...
• Segment...

🌈 F/X

• Solarize...
• Swirl...
• Implode...
• Wave...
• Oil Paint
• Charcoal Draw...

🧚 Miscellany

• Image Info
• Zoom Image
• Show Preview...
• Show Histogram
• Show Matte

Plus the classics:

• Help
• Dismiss

---

You can make adjustments to the region of interest by moving the pointer to one of the rectangle corners, pressing a button, and dragging 👑 Finally, choose an image processing technique from the Command widget. And here's the cute part: you can choose more than one image processing technique to apply to an area. Alternatively, you can move the region of interest before applying another image processing technique. When you're done, press Dismiss. Periodt. 💖

🖼️ IMAGE PANNING 🖼️

okay so this is the situation: when an image is bigger than the width or height of the X server screen, display lowkey maps a lil panning icon for you. the rectangle inside that panning icon shows the area that's currently displayed in the image window — like a mini map but make it iconic.

to pan around the image, press any button and drag the pointer inside the panning icon. the pan rectangle moves with the pointer and the image window updates to reflect where the rectangle is inside the panning icon. when you've picked the area of the image you wanna see, release the button. that's it, that's the whole vibe 💅

you can also use the arrow keys to pan the image one pixel up, down, left, or right within the image window. precision queen behavior fr fr 🌸

the panning icon gets withdrawn if the image becomes smaller than the dimensions of the X server screen. it simply ✨ leaves ✨.

---

💖 USER PREFERENCES 💖

Preferences affect the default behavior of display. each preference is either true or false and they're stored in your home directory as .displayrc. the girlies love a config file 🎀

• display image centered on a backdrop" — this backdrop covers the entire workstation screen and is useful for hiding other X window activity while you're viewing the image. the color of the backdrop is specified as the background color. refer to X Resources for the details, iykyk.
• confirm on program exit" — asks for a confirmation before exiting the display program. safety net energy 🫶
• correct image for display gamma" — if the image has a known gamma, the gamma is corrected to match that of the X server (see the X Resource displayGamma). match made in heaven ngl.
• display warning messages" — displays any warning messages. yes bestie, we wanna know.
• apply Floyd/Steinberg error diffusion to image" — okay so the basic strategy of dithering is to trade intensity resolution for spatial resolution by averaging the intensities of several neighboring pixels. images that suffer from severe contouring when reducing colors can be improved with this preference. it's giving glow-up ✨
• use a shared colormap for colormapped X visuals" — this option only applies when the default X server visual is PseudoColor or GRAYScale. refer to -visual for more details. by default, a shared colormap is allocated. the image shares colors with other X clients. some image colors could be approximated, therefore your image may look very different than intended. otherwise the image colors appear exactly as they are defined — but other clients may go full technicolor when the image colormap is installed 😭 pick your fighter.
• display images as an X server pixmap" — images are maintained as a XImage by default. set this resource to True to utilize a server Pixmap instead. this option is useful if your image exceeds the dimensions of your server screen and you intend to pan the image — panning is much faster with Pixmaps than with a XImage. but Pixmaps are considered a precious resource, so use them with discretion. treat them like the last slice of cake 🍓

---

👑 GM IDENTIFY 👑

Identify describes the format and characteristics of one or more image files as internally supported by the software. it will also report if an image is incomplete or corrupt — snitching, but make it helpful 💀

the info displayed includes:

• the scene number
• the file name
• the width and height of the image
• whether the image is colormapped or not
• the number of colors in the image
• the number of bytes in the image
• the format of the image (JPEG, PNM, etc.)
• and finally the number of seconds in both user time and elapsed time it took to read and process the image

if -verbose or +ping are provided as an option, the pixel read rate is also displayed. an example line output from identify follows:

images/aquarium.miff 640x480 PseudoClass 256c
       308135b MIFF 0.000u 0:01

if -verbose is set, expect additional output including any image comment:

Image: images/aquarium.miff
class: PseudoClass
colors: 256
signature: eb5dca81dd93ae7e6ffae99a527eb5dca8...
matte: False
geometry: 640x480
   depth: 8
bytes: 308135
format: MIFF
comments:
Imported from MTV raster image: aquarium.mtv

for some formats, additional format-specific information about the file will be written if the -debug coder or -debug all option is used. extra lore, no cap 🌈

---

🌸 IDENTIFY OPTIONS 🌸

options are processed in command line order. any option you specify on the command line remains in effect for the set of images immediately following, until the set is terminated by the appearance of any option or -noop. that's the rule, periodt.

for a more detailed description of each option, see Options, above.

• -authenticate <string> — decrypt image with this password 🔐
• -debug <events> — enable debug printout
• -define <key>{=<value>},... — add coder/decoder specific options
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image
• -depth <value> — depth of the image
• -format <string> — output formatted image characteristics
• -help — print usage instructions
• -interlace <type> — the type of interlacing scheme
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit
• -log <string> — Specify format for debug log
• -ping — efficiently determine image characteristics ⚡
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors
• -size <width>x<height>{+offset} — width and height of the image
• -verbose — print detailed information about the image
• -version — print detailed GraphicsMagick version/build information
• -version — print compact GNU-style GraphicsMagick version information

yes bestie, -version is listed twice in here. we don't question it, we just copy it faithfully 🫡

for a more detailed description of each option, see Options, above.

---

🧚 GM IMPORT 🧚

Import reads an image from any visible window on an X server and outputs it as an image file. you can capture a single window, the entire screen, or any rectangular portion of the screen. then use display for redisplay, printing, editing, formatting, archiving, image processing, etc. of the captured image — because we don't just screenshot and let it rot in a folder, we curate 🎀

the target window can be specified by id, name, or may be selected by clicking the mouse in the desired window. if you press a button and then drag, a rectangle will form which expands and contracts as the mouse moves. to save the portion of the screen defined by the rectangle, just release the button. main character selection energy ✨

the keyboard bell is rung once at the beginning of the screen capture and twice when it completes 🔊🔊

---

💜 EXAMPLES 💜

to select an X window or an area of the screen with the mouse and save it in the MIFF image format to a file entitled window.miff, use:

gm import window.miff

to select an X window or an area of the screen with the mouse and save it in the Encapsulated PostScript format to include in another document, use:

gm import figure.eps

to capture the entire X server screen in the JPEG image format in a file entitled root.jpeg, without using the mouse, use:

gm import -window root root.jpeg

to capture the 512x256 area at the upper right corner of the X server screen in the PNG image format in a well-compressed file entitled corner.png, without using the mouse, use:

gm import -window root -crop 512x256-0+0 -quality 90
       corner.png

go off bestie, your screenshots are gonna be so crispy 🔥

✨ OPTIONS ✨

Okay bestie, options are processed in command line order — like, first come, first served, no line-jumping 💅. Any option you specify on the command line stays in effect until it is explicitly changed by specifying the option again with a different effect. No cap.

Import options can appear on the command line or in your X resources file. See X(1) for that whole situation. Options on the command line supersede values specified in your X resources file, periodt.

For a more detailed description of each option, see Options, above. (Yes, the page is being extra and pointing you upward. Iconic.)

• -bordercolor <color> — the border color 🎀
• -colors <value> — preferred number of colors in the image 🌈
• -colorspace <value> — the type of colorspace 🎨
• -comment <string> — annotate an image with a comment 💬
• -crop <width>x<height>{+-}<x>{+-}<y>{%} — preferred size and location of the cropped image ✂️
• -debug <events> — enable debug printout 🐛
• -define <key>{=<value>},... — add coder/decoder specific options 🧚
• -delay <1/100ths of a second> — display the next image after pausing ⏳
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image 📏
• -depth <value> — depth of the image 🧬
• -descend — obtain image by descending window hierarchy 🪟
• -display <host:display[.screen] — specifies the X server to contact 🖥️
• -dispose <method> — GIF disposal method 🗑️
• -dither — apply Floyd/Steinberg error diffusion to the image ✨
• -encoding <type> — specify the text encoding 🔤
• -endian <type> — specify endianness (MSB, LSB, or Native) of image 💾
• -frame — include the X window frame in the imported image 🖼️
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — Specify dimension, offset, and resize options. 📐
• -help — print usage instructions 🆘
• -interlace <type> — the type of interlacing scheme 🧵
• -label <name> — assign a label to an image 🏷️
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit 🚦
• -log <string> — Specify format for debug log 📝
• -monitor — show progress indication 👀
• -monochrome — transform the image to black and white 🖤
• -negate — replace every pixel with its complementary color 🔄
• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas 📄
• -pause <seconds> — pause between snapshots [import] ⏸️
• -ping — efficiently determine image characteristics ⚡
• -pointsize <value> — pointsize of the PostScript, X11, or TrueType font 🔠
• -quality <value> — JPEG/MIFF/PNG/TIFF compression level 🗜️
• -resize <width>x<height>{%}{@}{!}{<}{>} — resize an image 📏
• -rotate <degrees>{<}{>} — rotate the image 🔄
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors 🎛️
• -scene <value> — set scene number 🎬
• -screen — specify the screen to capture 📸
• -set <attribute> <value> — set an image attribute 🛠️
• +set <attribute> — unset an image attribute 🧹
• -silent — operate silently 🤫
• -snaps <value> — number of screen snapshots 📸
• -thumbnail <width>x<height>{%}{@}{!}{<}{>} — resize an image (quickly) ⚡
• -transparent <color> — make this color transparent within the image 🫥
• -trim — trim an image ✂️
• -verbose — print detailed information about the image 🗣️
• -version — print detailed GraphicsMagick version/build information 🏷️
• -version — print compact GNU-style GraphicsMagick version information 🏷️

For a more detailed description of each option, see Options, above. 💖

---

🦋 GM MOGRIFY 🦋

Mogrify transforms an image or a sequence of images. These transforms include image scaling, image rotation, color reduction, and others. Each transmogrified image overwrites the corresponding original image, unless an option such as -format causes the output filename to be different from the input filename. 💅

The graphics formats supported by mogrify are listed in gm(1). 📚

---

💖 EXAMPLES 💖

To convert all the TIFF files in a particular directory to JPEG, use:

gm mogrify -format jpeg *.tiff

To convert a directory full of JPEG images to thumbnails, use:

gm mogrify -size 120x120 *.jpg -resize 120x120 +profile "*"

In this example, -size 120x120 gives a hint to the JPEG decoder that the images are going to be downscaled to 120x120, allowing it to run faster by avoiding returning full-resolution images to GraphicsMagick for the subsequent resizing operation. The -resize 120x120 specifies the desired dimensions of the output images. It will be scaled so its largest dimension is 120 pixels. The +profile "*" removes any ICM, EXIF, IPTC, or other profiles that might be present in the input and aren't needed in the thumbnails. ✨

To scale an image of a cockatoo to exactly 640 pixels in width and 480 pixels in height, use:

gm mogrify -resize 640x480! cockatoo.miff

Okay so before we get into the good stuff — a little rules-of-the-road moment, bestie, because this actually matters:

Options are processed in command line order. No taking numbers and sorting them later, no queue-jumping — they go down the line exactly like you typed them. And any option you drop on the command line stays in effect for the set of images that follows it, right up until that set gets terminated by the appearance of any option or -noop. So yeah — order is the whole vibe. ✨

For a more detailed description of each option, see Options, above. (Yes, above-above. She said what she said. 👑

---

• -affine <matrix> — the drawing transform matrix. straight up matrix math for your draw ops 💅
• -antialias — removes pixel aliasing, so your edges stop looking crunchy
• -asc-cdl <spec> — applies an ASC CDL color transform
• -authenticate <string> — decrypts the image with this password 🔐
• -auto-orient — orients (rotates) the image so it's upright, no more sideways pics
• -background <color> — sets the background color
• -black-threshold red[,green][,blue][,opacity] — everything below the threshold turns black
• -blue-primary <x>,<y> — the blue chromaticity primary point
• -blur <radius>{x<sigma>} — blurs the image with a Gaussian operator, soft-girl filter fr
• -border <width>x<height> — surrounds the image with a border of color
• -bordercolor <color> — the border color, obviously 🎀
• -channel <type> — the type of channel
• -charcoal <factor> — simulates a charcoal drawing, artsy era
• -colorize <value> — colorizes the image with the pen color
• -colors <value> — your preferred number of colors in the image
• -colorspace <value> — the type of colorspace
• -comment <string> — annotates an image with a comment
• -compose <operator> — the type of image composition
• -compress <type> — the type of image compression
• -contrast — enhance or reduce the image contrast
• -convolve <kernel> — convolves the image with the specified convolution kernel
• -create-directories — creates the output directory if it's required 👑
• -crop <width>x<height>{+-}<x>{+-}<y>{%} — preferred size and location of the cropped image ✂️
• -cycle <amount> — displaces the image colormap by amount
• -debug <events> — enables debug printout, for when you need to see everything
• -define <key>{=<value>},... — adds coder/decoder specific options
• -delay <1/100ths of a second> — displays the next image after pausing, frame timing bestie
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image
• -depth <value> — depth of the image
• -despeckle — reduces the speckles within an image, clearing the noise ✨
• -display <host:display[.screen] — specifies the X server to contact
• -dispose <method> — the GIF disposal method
• -dither — applies Floyd/Steinberg error diffusion to the image
• -draw <string> — annotates an image with one or more graphic primitives
• -edge <radius> — detects edges within an image
• -emboss <radius> — embosses an image
• -encoding <type> — specifies the text encoding
• -endian <type> — specifies endianness (MSB, LSB, or Native) of the image
• -enhance — applies a digital filter to enhance a noisy image
• -equalize — performs histogram equalization to the image
• -extent <width>x<height>{+-}<x>{+-}<y> — composites the image on a background color canvas image
• -fill <color> — the color to use when filling a graphic primitive
• -filter <type> — use this type of filter when resizing an image
• -flip — creates a "mirror image" 🪞
• -flop — creates a "mirror image" 🪞 (yes, also a mirror image — different axis energy)
• -font <name> — use this font when annotating the image with text
• -format <type> — the image format type
• -frame <width>x<height>+<outer bevel width>+<inner bevel width> — surrounds the image with an ornamental border 🖼️
• -fuzz <distance>{%} — colors within this Euclidean distance are considered equal
• -gamma <value> — level of gamma correction
• -gaussian <radius>{x<sigma>} — blurs the image with a Gaussian operator
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — specifies dimension, offset, and resize options
• -gravity <type> — the direction a primitive gravitates to when annotating the image
• -green-primary <x>,<y> — the green chromaticity primary point
• -hald-clut <clut> — applies a Hald CLUT to the image
• -help — prints usage instructions, the classic cry for help 🫶
• -implode <factor> — implodes image pixels about the center
• -interlace <type> — the type of interlacing scheme
• -label <name> — assigns a label to an image
• -lat <width>x<height>{+-}<offset>{%} — performs local adaptive thresholding
• -level <black_point>{,<gamma>}{,<white_point>}{%} — adjusts the level of image contrast
• -limit <type> <value> — a Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit
• -linewidth — the line width for subsequent draw operations
• -list <type> — the type of list
• -log <string> — specify the format for the debug log
• -loop <iterations> — adds the Netscape loop extension to your GIF animation 🔁
• -magnify — magnifies the image
• -map <filename> — chooses a particular set of colors from this image
• -mask <filename> — specifies a clipping mask
• -matte — stores the matte channel if the image has one
• -mattecolor <color> — specifies the color to be used with the -frame option
• -median <radius> — applies a median filter to the image
• -minify <factor> — minifies the image (the opposite of her magnify era)
• -modulate brightness[,saturation[,hue]] — varies the brightness, saturation, and hue of an image 🌈
• -monitor — shows progress indication, so you're not just staring at a blank terminal
• -monochrome — transforms the image to black and white
• -motion-blur <radius>{x<sigma>}{+angle} — simulates motion blur, zoom-zoom 🏃‍♀️
• -negate — replaces every pixel with its complementary color
• -noise <radius|type> — adds or reduces noise in an image
• -noop — NOOP (no option). she does nothing and she serves 💀
• -normalize — transforms the image to span the full range of color values
• -opaque <color> — changes this color to the pen color within the image
• -operator channel operator rvalue[%] — applies a mathematical, bitwise, or value operator to an image channel
• -ordered-dither <channeltype> <NxN> — ordered dithers the image
• -output-directory <directory> — outputs files to the directory you name 📁
• -orient <orientation> — sets the image orientation attribute
• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas
• -paint <radius> — simulates an oil painting 🎨
• -pen <color> — (this option has been replaced by the -fill option — she's retired, show some respect 🙈)
• -pointsize <value> — pointsize of the PostScript, X11, or TrueType font
• -profile <filename> — adds an ICM, IPTC, or generic profile to the image
• -preserve-timestamp — preserves the original timestamps of the file ⏰
• -quality <value> — JPEG/MIFF/PNG/TIFF compression level
• -raise <width>x<height> — lightens or darkens image edges
• -random-threshold <channeltype> <LOWxHIGH> — random thresholds the image
• -recolor <matrix> — applies a color translation matrix to image channels
• -map <filename> — chooses a particular set of colors from this image (yes, she's listed a second time — not me, not my fault, i just work here 💀)
• -red-primary <x>,<y> — the red chromaticity primary point
• -region <width>x<height>{+-}<x>{+-}<y> — applies options to a portion of the image
• -render — renders vector operations
• -repage <width>x<height>+xoff+yoff[!] — adjusts image page offsets
• -resample <horizontal>x<vertical> — resamples the image to the specified horizontal and vertical resolution
• -resize <width>x<height>{%}{@}{!}{<}{>} — resizes an image
• -roll {+-}<x>{+-}<y> — rolls an image vertically or horizontally 🌀
• -rotate <degrees>{<}{>} — rotates the image
• -sample <geometry> — scales the image using pixel sampling
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors
• -scale <geometry> — scales the image
• -scene <value> — sets the scene number
• -set <attribute> <value> — sets an image attribute
• +set <attribute> — unsets an image attribute (plus sign girlies, this is your moment ✨)
• -segment <cluster threshold>x<smoothing threshold> — segments an image
• -shade <azimuth>x<elevation> — shades the image using a distant light source 💡
• -sharpen <radius>{x<sigma>} — sharpens the image
• -shave <width>x<height>{%} — shaves pixels from the image edges
• -shear <x degrees>x<y degrees> — shears the image along the X or Y axis
• -size <width>x<height>{+offset} — width and height of the image
• -solarize <factor> — negates all pixels above the threshold level
• -spread <amount> — displaces image pixels by a random amount
• -strip — removes all profiles and text attributes from the image
• -stroke <color> — the color to use when stroking a graphic primitive
• -strokewidth <value> — sets the stroke width
• -swirl <degrees> — swirls image pixels about the center 🍥
• -texture <filename> — name of the texture to tile onto the image background
• -threshold <value>{%} — thresholds the image
• -thumbnail <width>x<height>{%}{@}{!}{<}{>} — resizes an image (quickly) ⚡
• -tile <filename> — tiles the image when filling a graphic primitive
• -transform — transforms the image
• -transparent <color> — makes this color transparent within the image 🫥
• -treedepth <value> — tree depth for the color reduction algorithm
• -trim — trims an image, bye-bye dead space
• -type <type> — the image type
• -units <type> — the units of image resolution
• -unsharp <radius>{x<sigma>}{+<amount>}{+<threshold>} — sharpens the image with an unsharp mask operator
• -verbose — prints detailed information about the image 🗣️
• -version — prints detailed GraphicsMagick version/build information
• -version — prints compact GNU-style GraphicsMagick version information (she's here twice too — both real, both iconic 👑)
• -view <string> — FlashPix viewing parameters
• -virtual-pixel <method> — specifies the contents of "virtual pixels"
• -wave <amplitude>x<wavelength> — alters an image along a sine wave 🌊
• -white-point <x>,<y> — the chromaticity white point
• -white-threshold red[,green][,blue][,opacity] — pixels above the threshold become white

For a more detailed description of each option, see Options, above. Still true, still up there, still slaying. 💖

🌈 GM MONTAGE 🌈

okay bestie, montage creates a composite image by combining several separate images. The images are tiled on the composite image with the name of the image optionally appearing just below the individual tile. It's giving collage era ✨

The composite image is constructed in the following manner. First, each image specified on the command line, except for the last, is scaled to fit the maximum tile size. The maximum tile size by default is 120x120. It can be modified with the -geometry command line argument or X resource. See Options for more information on command line arguments. See X(1) for more information on X resources. Note that the maximum tile size need not be a square. 🫶

Next the composite image is initialized with the color specified by the -background command line argument or X resource. The width and height of the composite image is determined by the title specified, the maximum tile size, the number of tiles per row, the tile border width and height, the image border width, and the label height. The number of tiles per row specifies how many images are to appear in each row of the composite image. The default is to have 5 tiles in each row and 4 tiles in each column of the composite. A specific value is specified with -tile. The tile border width and height, and the image border width defaults to the value of the X resource -borderwidth. It can be changed with the -borderwidth or -geometry command line argument or X resource. The label height is determined by the font you specify with the -font command line argument or X resource. If you do not specify a font, a font is chosen that allows the name of the image to fit the maximum width of a tiled area. The label colors is determined by the -background and -fill command line argument or X resource. Note, that if the background and pen colors are the same, labels will not appear. 💅

Initially, the composite image title is placed at the top if one is specified (refer to -fill). Next, each image is set onto the composite image, surrounded by its border color, with its name centered just below it. The individual images are left-justified within the width of the tiled area. The order of the images is the same as they appear on the command line unless the images have a scene keyword. If a scene number is specified in each image, then the images are tiled onto the composite in the order of their scene number. Finally, the last argument on the command line is the name assigned to the composite image. By default, the image is written in the MIFF format and can be viewed or printed with gm display. 🎀

bestie, note that if the number of tiles exceeds the default number of 20 (5 per row, 4 per column), more than one composite image is created. To ensure a single image is produced, use -tile to increase the number of tiles to meet or exceed the number of input images. ⚠️

Finally, to create one or more empty spaces in the sequence of tiles, use the "NULL:" image format. 🕳️

Note, a composite MIFF image displayed to an X server with gm display behaves differently than other images. You can think of the composite as a visual image directory. Choose a particular tile of the composite and press a button to display it. See gm and miff(5). 👑

---

✨ EXAMPLES ✨

To create a montage of a cockatoo, a parrot, and a hummingbird and write it to a file called birds, use:

gm montage cockatoo.miff parrot.miff hummingbird.miff birds.miff

To tile several bird images so that they are at most 256 pixels in width and 192 pixels in height, surrounded by a red border, and separated by 10 pixels of background color, use:

gm montage -geometry 256x192+10+10 -bordercolor red birds.* montage.miff

To create an unlabeled parrot image, 640 by 480 pixels, and surrounded by a border of black, use:

gm montage -geometry 640x480 -bordercolor black -label "" parrot.miff bird.miff

To create an image of an eagle with a textured background, use:

gm montage -texture bumps.jpg eagle.jpg eagle.png

To join several GIF images together without any extraneous graphics (e.g. no label, no shadowing, no surrounding tile frame), use:

gm montage +frame +shadow +label -tile 5x1 -geometry 50x50+0+0 *.png joined.png

---

🎀 OPTIONS 🎀

Any option you specify on the command line remains in effect for the group of images following it, until the group is terminated by the appearance of any option or -noop. For example, to make a montage of three images, the first with 32 colors, the second with an unlimited number of colors, and the third with only 16 colors, use:

gm montage -colors 32 cockatoo.1 -noop cockatoo.2 -colors 16 cockatoo.3 cockatoos.miff

For a more detailed description of each option, see Options, above.

• -adjoin — join images into a single multi-image file. one-file energy, slay 💅
• -affine <matrix> — drawing transform matrix.
• -authenticate <string> — decrypt image with this password.
• -background <color> — the background color.
• -blue-primary <x>,<y> — blue chromaticity primary point.
• -blur <radius>{x<sigma>} — blur the image with a Gaussian operator.
• -bordercolor <color> — the border color.
• -borderwidth <geometry> — the border width.
• -chop <width>x<height>{+-}<x>{+-}<y>{%} — remove pixels from the interior of an image.
• -colors <value> — preferred number of colors in the image.
• -colorspace <value> — the type of colorspace.
• -comment <string> — annotate an image with a comment.
• -compose <operator> — the type of image composition.
• -compress <type> — the type of image compression.
• -crop <width>x<height>{+-}<x>{+-}<y>{%} — preferred size and location of the cropped image.
• -debug <events> — enable debug printout.
• -define <key>{=<value>},... — add coder/decoder specific options.
• -density <width>x<height> — horizontal and vertical resolution in pixels of the image.
• -depth <value> — depth of the image.
• -display <host:display[.screen]> — specifies the X server to contact.
• -dispose <method> — GIF disposal method.
• -dither — apply Floyd/Steinberg error diffusion to the image.
• -draw <string> — annotate an image with one or more graphic primitives.
• -encoding <type> — specify the text encoding.
• -endian <type> — specify endianness (MSB, LSB, or Native) of image.
• -fill <color> — color to use when filling a graphic primitive.
• -filter <type> — use this type of filter when resizing an image.
• -font <name> — use this font when annotating the image with text.
• -frame <width>x<height>+<outer bevel width>+<inner bevel width> — surround the image with an ornamental border.
• -gamma <value> — level of gamma correction.
• -geometry <width>x<height>{+-}<x>{+-}<y>{%}{@}{!}{^}{<}{>} — specify dimension, offset, and resize options.
• -gravity <type> — direction primitive gravitates to when annotating the image.
• -green-primary <x>,<y> — green chromaticity primary point.
• -help — print usage instructions.
• -interlace <type> — the type of interlacing scheme.
• -label <name> — assign a label to an image.
• -limit <type> <value> — Disk, File, Images, Map, Memory, Pixels, Width, Height, Read, Threads, or Write resource limit.
• -log <string> — specify format for debug log.
• -matte — store matte channel if the image has one.
• -mattecolor <color> — specify the color to be used with the -frame option.
• -mode <value> — mode of operation.
• -monitor — show progress indication.
• -monochrome — transform the image to black and white.
• -noop — NOOP (no option). literally the no-op, bestie 😌
• -page <width>x<height>{+-}<x>{+-}<y>{%}{!}{<}{>} — size and location of an image canvas.
• -pen <color> — (This option has been replaced by the -fill option) heads up, it's replaced 💀
• -pointsize <value> — pointsize of the PostScript, X11, or TrueType font.
• -quality <value> — JPEG/MIFF/PNG/TIFF compression level.
• -red-primary <x>,<y> — red chromaticity primary point.
• -render — render vector operations.
• -repage <width>x<height>+xoff+yoff[!] — adjust image page offsets.
• -resize <width>x<height>{%}{@}{!}{<}{>} — resize an image.
• -rotate <degrees>{<}{>} — rotate the image.
• -sampling-factor <horizontal_factor>x<vertical_factor> — chroma subsampling factors.
• -scenes <value-value> — range of image scene numbers to read.
• -shadow <radius>{x<sigma>} — shadow the montage.
• -sharpen <radius>{x<sigma>} — sharpen the image.
• -size <width>x<height>{+offset} — width and height of the image.
• -strip — remove all profiles and text attributes from the image.
• -stroke <color> — color to use when stroking a graphic primitive.
• -strokewidth <value> — set the stroke width.
• -texture <filename> — name of texture to tile onto the image background.
• -thumbnail <width>x<height>{%}{@}{!}{<}{>} — resize an image (quickly).
• -tile <geometry> — layout of images [montage].
• -title <string> — assign title to displayed image [animate, display, montage].
• -transform — transform the image.
• -transparent <color> — make this color transparent within the image.
• -treedepth <value> — tree depth for the color reduction algorithm.
• -trim — trim an image.
• -type <type> — the image type.
• -verbose — print detailed information about the image.
• -version — print detailed GraphicsMagick version/build information.
• -version — print compact GNU-style GraphicsMagick version information. (yes, twin, it's listed again 💖)
• -white-point <x>,<y> — chromaticity white point.

For a more detailed description of each option, see Options, above. 🫶

🎀 X RESOURCES 🎀

okay so bestie, montage options can show up on the command line or in your X resource file. command-line options supersede values specified in your X resource file, no cap. see X(1) for more X resources lore ✨

all montage options have a corresponding X resource. on top of that, montage uses these:

• background (class Background) — background color. specifies the preferred color to use for the composite image background. the default is #ccc.
• borderColor (class BorderColor) — border color. specifies the preferred color to use for the composite image border. the default is #ccc.
• borderWidth (class BorderWidth) — border width. specifies the width in pixels of the composite image border. the default is 2.
• font (class Font) — font to use. specifies the name of the preferred font to use when displaying text within the composite image. the default is 9x15, fixed, or 5x8 determined by the composite image size.
• matteColor (class MatteColor) — color of the frame. specify the color of an image frame. a 3D effect is achieved by using highlight and shadow colors derived from this color. the default value is #697B8F.
• pen (class Pen) — text color. specifies the preferred color to use for text within the composite image. the default is black.
• title (class Title) — composite image title. this resource specifies the title to be placed at the top of the composite image. the default is not to place a title at the top of the composite image.

---

⏱️ GM TIME ⏱️

DESCRIPTION

time executes an arbitrary gm utility command (e.g. convert) and reports the user and elapsed time. this gives you a portable and consistent way to measure command execution times, kinda like the Unix time command but with main-character consistency 💅

EXAMPLES

to obtain time information for the execution of a command:

% gm time convert input.ppm -gaussian 0x2 output.ppm
convert input.ppm -gaussian 0x2 output.ppm    22.60s user 0.00s system 2354% cpu 0.960 total

here is the interpretation of the above output:

• user — the total user time consumed.
• system — the total system time consumed.
• total — the total elapsed time consumed.

OPTIONS

the time command requires no options other than the gm command to execute. iconic, honestly 🫶

---

🦋 GM VERSION 🦋

DESCRIPTION

version displays the software release version, build quantum (pixel sample) depth, web site URL, copyright notice, enabled features support, configuration parameters, and final build options used to build the software. the available information depends on how the software was configured and the host system. iykyk 🌈

EXAMPLES

to display the version information:

  GraphicsMagick 1.3.48 2026-07-23 Q16 http://www.GraphicsMagick.org/
  Copyright (C) 2002-2026 GraphicsMagick Group.
  Additional copyrights and licenses apply to this software.
  See http://www.GraphicsMagick.org/www/Copyright.html for details.
  Feature Support:
    Native Thread Safe         yes
    Large Files (> 32 bit)     yes
    Large Memory (> 32 bit)    yes
    BZIP                       yes
    FlashPix                   yes
    FreeType                   yes
    Ghostscript (Library)      no
    HEIF/HVEC ("HEIC")         yes
    JBIG                       yes
    JPEG-2000                  yes
    JPEG                       yes
    JPEG XL                    yes
    Little CMS                 yes
    Loadable Modules           yes
    Solaris mtmalloc           no
    Google perftools tcmalloc  no
    OpenMP                     yes (201511 "4.5")
    PNG                        yes
    TIFF                       yes
    TRIO                       no
    Solaris umem               no
    WebP                       yes
    WMF                        yes
    X11                        yes
    XML                        yes
    ZLIB                       yes
  Host type: x86_64-pc-linux-gnu
  Configured using the command:
    ./configure  ...
  Final Build Parameters:
    CC       = ...
    CFLAGS   = ...
    CPPFLAGS = ...
    CXX      = ...
    CXXFLAGS = ...
    LDFLAGS  = ...
    LIBS     = ...

OPTIONS

the version command does not currently support any options. slay 💖

---

okay bestie, that's the whole gm(1) page ✨ source/version: GraphicsMagick, date: 2026/10/05 — go be the main character of your terminal 🫶🎀