Script for applying advanced crossfeed using Redline Monitor

apply_crossfeed.bash

#!/bin/bash

# Set up parameters
in=`realpath "${1}"`
echo $in
indir="${in%/*}"
infile="${in##*/}"
outdir="${indir}_headphone"
mkdir -p "$outdir"
tmpin="/tmp/in_${infile}"
tmpout="/tmp/out_${infile}"
tmpcover="/tmp/cover_${infile}"
out="${outdir}/${infile}"
# Get the original bit depth for later
original_bit_depth=`file -b "${in}" | cut -d " " -f 5`
if [ "${original_bit_depth}" == "16" ]; then
  sample_fmt="s16"
else
  sample_fmt="s32"
fi

function remove_temp_files() {
  rm "${tmpin}"
  rm "${tmpout}"
  rm "${tmpcover}"
}

function die() {
  echo $*
  [[ "${BASH_SOURCE[0]}" == "${0}" ]] && exit 1
  remove_temp_files
}

# Note - I'm using a bunch of temp files because I couldn't get streaming processing at high bit depths to work well with mrswatson64

# Reduce level to leave processing headroom for mrwatson
ffmpeg -y -i "${in}" -filter:a "volume=-3dB" -sample_fmt s32 "${tmpin}" || die "unable to reduce input level"

# Apply advanced crossfeed
mrswatson64 --bit-depth 24 -p /Library/Audio/Plug-Ins/VST/Redline\ Monitor.vst --parameter 2,0.333 --input "${tmpin}" --output "${tmpout}" || die "unable to apply crossfeed"

# Normalize to 0dB peak to reduce the amount of amplification needed at playback
vol=`ffmpeg -i "${tmpout}" -af "volumedetect" -vn -sn -dn -f null /dev/null 2>&1 | grep "max_volume" | cut -d ' ' -f 5 | sed 's/\-//'` || die "unable to determine max volume"

# Note, I'm using the highest compression level to save space, but this slows down encoding
ffmpeg -y -i "${tmpout}" -filter:a "volume=+${vol}dB" -sample_fmt "${sample_fmt}" "${out}" || die "unable to encode final output"

# Copy flac metadata
metaflac --no-utf8-convert --export-tags-to=- "${in}" | metaflac --import-tags-from=- "${out}" || die "unable to copy flac metadata"
metaflac --export-picture-to="${tmpcover}" "${in}" && metaflac --import-picture-from="${tmpcover}" "${out}"

remove_temp_files