Make a single MP3 audio file of all the voices from MacOS `say` command

helloer.sh

#!/bin/bash
# set -o xtrace

# generates an MP3 file of every voice in MacOS saying the appropriate 'Hello'
# Requires ffmpeg to be install, by homebrew is preferred.

OUTPUT="everyone_hello" # extension will be added below
OUTPUT_AIFF="$OUTPUT.aiff"
OUTPUT_MP3="$OUTPUT.mp3"

RATE=300 # speed of speech: higher number is faster
GAP_LENGTH=3 # length of silence between each voice

TEMP_FILE="hello_temp.aiff"

# It first outputs the work to lossless .aiff files, and then converts the final to .mp3

# Make a bunch of .aiff files, one for each voice in say(1)
say -v \? | awk 'BEGIN { FS = "，|、| +"} { print $1, $4 }' | sed "s/'//g" | while read LINE; do
  echo $LINE
  NAME=$(echo $LINE | awk 'BEGIN { FS = " " } { print $1 }')
  FN=hello_$NAME.aiff

  echo $FN

  HELLO=$(echo $LINE | awk 'BEGIN { FS = " " } { print $2 }' | sed 's/Most/Hello/' | sed "s/Isnt/Hello/g" | sed "s/I$/Hello/g")
  HELLO="$NAME says $HELLO"
  echo $HELLO

  say -o "$FN" -r $RATE -v $NAME "$HELLO"
done

# Generate a spacer of silence. Change `-t` increase/decrease time
ffmpeg -f lavfi -i anullsrc=r=44100:cl=mono -t $GAP_LENGTH -y silence.aiff

# create the initial output file
ffmpeg -i silence.aiff -filter_complex '[0:0]concat=n=1:v=0:a=1[out]' -map '[out]' -y $OUTPUT_AIFF

for FILENAME in hello_*.aiff; do
  echo $FILENAME
  ffmpeg -i $OUTPUT_AIFF -i $FILENAME -i silence.aiff -filter_complex '[0:0][1:0][2:0]concat=n=3:v=0:a=1[out]' -map '[out]' -y $TEMP_FILE
  mv $TEMP_FILE $OUTPUT_AIFF
  rm -v $FILENAME
done

# convert output to mp3
ffmpeg -i $OUTPUT_AIFF -y $OUTPUT_MP3

# cleanup
rm -v $OUTPUT_AIFF
rm -v silence.aiff

echo "wrote $OUTPUT_MP3. done."