SPHINX_PITCH(1)

💖 sphinx_pitch — the girl who finds the pitch 💖

okay bestie, so like. this is the manual for sphinx_pitch, a little program from the General Commands Manual section 1 (that's the "commands you type in a terminal" section, iykyk). she reads your audio files and tells you what pitch is in there and whether anybody's actually talking. it's giving vocal coach energy fr fr. 🎤✨

---

✨ NAME ✨

sphinx_pitch — extract pitch from audio files 🦋

---

🎀 SYNOPSIS 🎀

sphinx_pitch [ options ]...

that's it. that's the whole vibe. you throw some options at her and she goes. 💅

---

🌸 DESCRIPTION 🌸

okay so here's the tea: this program reads audio files and analyzes them for pitch and voicing. that's literally the whole job. she listens, she figures out the fundamental frequency, she tells you if there's a voice in there. main character energy honestly. 👑

now, all the little knobs and switches, in order, no skips, no skips periodt:

• -c — file for batch processing 🍓
• -di — directory, input file names are relative to this, if defined
• -do — directory, output files are relative to this
• -ei — extension to be applied to all input files
• -eo — extension to be applied to all output files
• -flen — Number of seconds in each analysis frame (needs to be greater than twice the longest period you wish to detect — to detect down to 80Hz you need a frame length of 2.0/80 = 0.025).

not me stressing this one: -flen genuinely has to be greater than twice the longest period you wanna detect, or the math just doesn't math. that's the rule, bestie. 🔥

• -fshift — Frame shift: number of seconds between each analysis frame.
• -i — audio input file
• -input_endian — of audio data (will be determined automatically if not given)
• -mswav — Defines input format as Microsoft Wav (RIFF)
• -nist — Defines input format as NIST sphere
• -nskip — If a control file was specified, the number of utterances to skip at the head of the file
• -o — text output file (standard output will be used if not given)
• -raw — Defines input format as raw binary data
• -runlen — If a control file was specified, the number of utterances to process (see -nskip too)
• -samprate — Sampling rate of audio data (will be determined automatically if 0)
• -search_range — Fraction of the best local estimate to use as a search range for smoothing.
• -smooth_window — Number of frames on either side of the current frame to use for smoothing.
• -voice_thresh — Threshold of normalized difference under which to search for the fundamental period.

see? every single one. we don't gatekeep here. 🫶

---

👑 AUTHOR 👑

David Huggins-Daines <dhdaines@gmail.com> — absolute icon, thank you for your service. 🙏

---

💜 COPYRIGHT 💜

Copyright © 2007-2008 Carnegie Mellon University. See the file COPYING included with this package for more information.

---

okay and that's the whole page!! it was written on 2007-05-12, it lives in the General Commands Manual as sphinx_pitch(1), and this version is 100% verbatim-facts, 100% vibe. go pitch something, bestie, you're gonna slay. 💖✨🎀🦋🌸💅🫶🔥👑🌈⭐💀🙈🍓🧚💜