Batch convert all .mp3 files in the current directory to 432Hz with ffmpeg

Convert2_432Hz.sh

#!/bin/bash
# Batch convert all .mp3 files in the current directory to 432Hz with ffmpeg

# Options
suffix="false" # Append the -432Hz suffix or not

oldIFS=$IFS
IFS=$'\n'
set -f # Deal with blanks and special characters in file names of the file command and in for loop

found=($(find . -name "*.mp3")) # Find the files in the current directory

IFS=$oldIFS
set +f

cpuCount=$(nproc --all)
started=0

# Wait for all remaining sub shells to complete, first argument is the counter of max subshells
wait4Completion() {
  if [ "$1" -gt 0 ]; then
    ((started++)) # Increase variable
  fi

  if [ $started -gt "$1" ]; then
    wait;
    started=0;
  fi
}

for file in "${found[@]}"; do # Iterate the found files
  (mv "$file" "$file.tmp"; ffmpeg -loglevel 8 -i "$file.tmp" -af "asetrate=48000*432/440,aresample=48000,atempo=440/432" "$file"; rm "$file.tmp"; echo "Pitched $file") &
  wait4Completion $cpuCount
done

wait4Completion 0

I know the -y flag of ffmpeg, but it did not work on .mp3 files.

Modifie la ligne 21 le dernier mot "$file" modifie le en "${file%.mp3}-432Hz.mp3"

I do not know French, but Google translate does fortunately. I added a "suffix" variable so the users can simply switch between the two file endings.

This is weird. The md says it's 441 -> 432 which is 9 Hz, which may mean you actually did calculate the percent change for 441 -> 432 instead of 440 -> 432 Hz which is 8 Hz.
That's a minor nitpick for anyone who doesn't use percent ratios or cents for retuning, though.

You are breaking the 8000, 16000, 22050, 44100, 48000 Hz which are the most common samplerates...
There are a few players out there that may dislike nonstandard rates.
(Edit: Lowering the sample rate may clamp high frequencies [tuned high frequencies/amplified {freq*2*2*2... or any multiplier that is divisible by 3, 6, 9, but not 13} are great] and make the music sound darker or slower which I dislike.)
I've no idea if 43200 will magically shift to 432 Hz, but here is a different command that does, and I've confirmed it works for the Solfeggios with another person who is more energy sensitive than me.

"asetrate=SAMPLE_RATE*NEW_FREQ/ORIG_FREQ,aresample=SAMPLE_RATE,atempo=ORIG_FREQ/NEW_FREQ"

For example:

ffmpeg -i file.ext -af "asetrate=48000*432/440,aresample=48000,atempo=440/432" out.ext

Found here https://www.mail-archive.com/[email protected]/msg30115.html

You are right, I got the math not entirely right and the ffmpeg params wrong too. I moved away from converting my music but with your suggestion I updated the script and it also sounds correct! Since, before the music seamed to slow down a bit.

Thanks a ton for pointing that out with the solution!

Additionally, I added parallel conversion limited by the CPU count. So whole music libraries can be converted much faster