D-F(1)

📦 version 7.44.0, in nixpkgs unstable

💖 D-F(1) — the d-f page, bestie edition 💖

okay so welcome to the D-F manual page, from the BRL-CAD fam ✨ pull up a chair, grab a boba, and let's talk about squishing big floats into small floats 🫶

🌸 NAME 🌸

d-f — converts double-precision floats into single-precision floats. that's it, that's the whole job, and it does it beautifully 💅

✨ SYNOPSIS ✨

d-f [-n | scale] < doubles > floats

like… look at her. she's got options (literally two of them) but she's lowkey simple 🎀

🦋 DESCRIPTION 🦋

so here's the vibe: d-f reads a whole stream of machine-native double-precision (that's 8-byte) binary floating point values from standard input, converts every single one into a single-precision (4-byte) binary float, and writes the results out to standard output. no cap, that's the entire personality 💀

you can pass an optional argument to control scaling. if you give it a numeric scale, every value gets multiplied by it during the conversion. and if you use the -n flag instead, that selects normalized output, leaving the unit scale of 1.0 in effect. with no argument at all? the values just get converted unchanged, no drama, no notes ✨

now here's the part where you gotta listen, bestie 😭 because the input and output are raw binary, d-f straight up refuses to run when standard input or standard output is a terminal (fr fr, it will not do it), and it also rejects a scale of zero. those are the rules, we don't make them, we just respect them 🙈

the inverse conversion — going from single-precision floats back to doubles — is handled by f-d(1). so they're like a lil duo 🫶

⚠️ heads up bestie: this command is deprecated and is scheduled for removal. it prints a deprecation warning on startup, so don't say nobody told you 💀

🎀 OPTIONS 🎀

  • -n — normalizes the data. the unit scale factor of 1.0 is retained so that data already sitting in the range -1.0 to +1.0 is preserved. it's giving "leave it as it is but make it official" 💅
  • scale — a numeric multiplier applied to every value as it's converted. may not be zero 🚫 (that's the one hard no)

🌈 EXAMPLES 🌈

⭐ Example 1. Halve the amplitude of a double stream

d-f 0.5 < signal.d > signal.f

this reads double-precision samples out of signal.d, scales each one by a half, and writes single-precision floats into signal.f. the result? a more compact file at reduced amplitude — smaller file, chiller vibes, we love a glow-down that saves disk space 🍓

⭐ Example 2. Round-trip through single precision

d-f < values.d | f-d > rounded.d

converts doubles to floats and then back again with f-d(1). this is genuinely a convenient way to see the precision you lose when signal-processing data gets stored as single-precision floats by tools like dconv(1) or dmod(1). it's giving "let's go around the world and see what we lost" 🧚

💜 SEE ALSO 💜

brlcad(1), f-d(1), dconv(1), dmod(1) — go read them, they're iconic too 👑

👑 AUTHOR 👑

BRL-CAD Team — the main characters responsible for all of this 🫶

this software is Copyright (c) 1989-2026 by the United States Government as represented by U.S. Army Research Laboratory. periodt 💅

🦋 BUG REPORTS 🦋

reports of bugs or problems should be submitted via electronic mail to devs@brlcad.org. found something cursed? tell the devs, not me, i'm just the messenger 😭📬

and that's the whole page, bestie!! 💖 go forth and convert your doubles, remember she's deprecated (sad but true 😔), and if you see the BRL-CAD crew and the D-F(1) footer floating around, that's just the source, version and date doing their lil sign-off dance — mwah, love u, bye!! 💅✨🎀

📦 where this page comes from