Convert Multiple audio tracks into a Single Track with 12 channels using ffmpeg

Question

I have multiple MOV files and each file has different numbers of audio tracks. I want to convert those files to MOV and combine the audio into a single 12-channel audio track.

For example, I have a source file with 4 single-channel audio tracks, and that file has to be output to a file with a single 12-channel audio track. The 1st audio track from the source file must be output to the 1st channel; the 2nd track to the 2nd channel; the 3rd track to the 3rd channel; the 4th track to the 4th channel; and channels 5–12 in the output file will be left without audio.

Another example: If the source file has 6 single-channel audio tracks, then these 6 tracks will be mapped to channels 1-6 channels in the output file and channels 7-12 will be left without audio.

Below is my FFmpeg command. It completes the video transcode without issue and also outputs a single 12-channel audio track. The problem is that all the audio tracks from the source file are being combined in the 1st channel, and channels 2-12 have no audio. I'm using Switch Telestream Player to QC the files.

ffmpeg -i <input>.mp4 -c:v prores_ks -profile:v 0 -quant_mat 2 -c:a pcm_s16le -b:a 512K -filter_complex "[0:a]pan=mono|c0=c0[a0];[0:a]pan=mono|c0=c1[a1];[0:a]pan=mono|c0=c2[a2];[0:a]pan=mono|c0=c3[a3];[0:a]pan=mono|c0=c4[a4];[0:a]pan=mono|c0=c5[a5];[0:a]pan=mono|c0=c6[a6];[0:a]pan=mono|c0=c7[a7];[0:a]pan=mono|c0=c8[a8];[0:a]pan=mono|c0=c9[a9];[0:a]pan=mono|c0=c10[a10];[0:a]pan=mono|c0=c11[a11];[a0][a1][a2][a3][a4][a5][a6][a7][a8][a9][a10][a11]amerge=inputs=12[aout]" -map 0:v -map "[aout]" <out>.mov

I'm using FFAStrans and its included ffmpeg executable to automate the movement and transcode of every file that has been placed in its watchfolder. Can you help me build the correct ffmpeg command?

Answer 1

In the formula below I included only the code related to mapping the mono tracks into a 12-channel track.

ffmpeg -i test_file.mov -filter_complex "[0:a:0]pan=mono|c0=c0[a0];[0:a:1]pan=mono|c0=c0[a1];[0:a:2]pan=mono|c0=c0[a2];[0:a:3]pan=mono|c0=c0[a3];[0:a:0]pan=mono|c0=c1[a4];[0:a:0]pan=mono|c0=c1[a5];[0:a:0]pan=mono|c0=c1[a6];[0:a:0]pan=mono|c0=c1[a7];[0:a:0]pan=mono|c0=c1[a8];[0:a:0]pan=mono|c0=c1[a9];[0:a:0]pan=mono|c0=c1[a10];[0:a:0]pan=mono|c0=c1[a11];[a0][a1][a2][a3][a4][a5][a6][a7][a8][a9][a10][a11]amerge=inputs=12[aout]" -map 0:v -map "[aout]" -c:a pcm_s24le -c:v copy output_file.mov

In your original code, only [a0] contains sound after mixing. You weren't, as you thought, mixing/combining the four mono tracks/channels into the same c0 of the 12-channel track. [a1]-[a11] were mixed from mono channels that didn't exist (anything other than c0) and that's why the channels they were mapped to were silent. It was hard to pick up on this initially because your test file contains 4 identical mono tracks, so one can't distinguish by ear if mono tracks were being mixed together or simply lost.

The issue with your code is shown in the tables below that compare the original with the new. Note that using [0:a] for audio track 0 is interpreted by FFmpeg as equivalent to [0:a:0]

Original Code

Command	Input Track	Input Channel	Input Audio Note	Mapped Sound?
[0:a]pan=mono|c0=c0[a0]	[0:a:0]	c0	Mono track 0 - c0 has audio	c0: Yes
[0:a]pan=mono|c0=c1[a1]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c1:
[0:a]pan=mono|c0=c2[a2]	[0:a:0]	c2	Mono track 0 - c2 isn't valid	c2:
[0:a]pan=mono|c0=c3[a3]	[0:a:0]	c3	Mono track 0 - c3 isn't valid	c3:
[0:a]pan=mono|c0=c4[a4]	[0:a:0]	c4	Mono track 0 - c4 isn't valid	c4:
[0:a]pan=mono|c0=c5[a5]	[0:a:0]	c5	Mono track 0 - c5 isn't valid	c5:
[0:a]pan=mono|c0=c6[a6]	[0:a:0]	c6	Mono track 0 - c6 isn't valid	c6:
[0:a]pan=mono|c0=c7[a7]	[0:a:0]	c7	Mono track 0 - c7 isn't valid	c7:
[0:a]pan=mono|c0=c8[a8]	[0:a:0]	c8	Mono track 0 - c8 isn't valid	c8:
[0:a]pan=mono|c0=c9[a9]	[0:a:0]	c9	Mono track 0 - c9 isn't valid	c9:
[0:a]pan=mono|c0=c10[a10]	[0:a:0]	c10	Mono track 0 - c10 isn't valid	c10:
[0:a]pan=mono|c0=c11[a11]	[0:a:0]	c11	Mono track 0 - c11 isn't valid	c11:

New Code

Command	Input Track	Input Channel	Input Audio Note	Mapped Sound?
[0:a:0]pan=mono|c0=c0[a0]	[0:a:0]	c0	Mono track 0 - c0 has audio	c0: Yes
[0:a:1]pan=mono|c0=c0[a1]	[0:a:1]	c0	Mono track 1 - c0 has audio	c1: Yes
[0:a:2]pan=mono|c0=c0[a2]	[0:a:2]	c0	Mono track 2 - c0 has audio	c2: Yes
[0:a:3]pan=mono|c0=c0[a3]	[0:a:3]	c0	Mono track 3 - c0 has audio	c3: Yes
[0:a:0]pan=mono|c0=c1[a4]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c4:
[0:a:0]pan=mono|c0=c1[a5]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c5:
[0:a:0]pan=mono|c0=c1[a6]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c6:
[0:a:0]pan=mono|c0=c1[a7]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c7:
[0:a:0]pan=mono|c0=c1[a8]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c8:
[0:a:0]pan=mono|c0=c1[a9]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c9:
[0:a:0]pan=mono|c0=c1[a10]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c10:
[0:a:0]pan=mono|c0=c1[a11]	[0:a:0]	c1	Mono track 0 - c1 isn't valid	c11:

Answer 2

I did some bash script, but I dont know how it helps you

#!/bin/bash
f="input 4.mp4"
if [ ! -f "$f" ]; then
ffmpeg -lavfi "
testsrc[v];
aevalsrc=sin(440*2*PI*t):s=8000[a];
aevalsrc=sin(880*2*PI*t)|sin(1760*2*PI*t):s=16000[b];
" -map "[v]" -map "[a]" -map "[b]" -t 5 "$f" -hide_banner -y
fi

AUD=($(ffprobe -v 0 -select_streams a -show_entries stream=channels -of csv=p=0 "$f"))
MRG=""
CNT=0
INP=0
for a in ${AUD[@]}; do
  echo "channels=$a"
  MRG+="[0:a:${CNT}]"
  ((CNT+=$a))
  ((INP++))
done
SIL=""
while (( $CNT < 12 )); do
  SIL+="aevalsrc=0[a${CNT}];"
  MRG+="[a${CNT}]"
  ((CNT++))
  ((INP++))
done

ffmpeg -i "$f" -filter_complex "
${SIL}
${MRG}amerge=inputs=${INP}[a]
" -map 0:v -map "[a]" -c:v copy -c:a pcm_s16le /tmp/out.mov -hide_banner -y

ffplay -f lavfi "amovie='/tmp/out.mov',asplit[a][out0];[a]showspectrum=s=540x1080:mode=separate:legend=1[out1]" -hide_banner -autoexit