After further research and experimenting, I've figured out how to do it using ffmpeg from command line.

1. Extract the relevant video stream (in my case track No. 2, which has index 1) into a temporary MP4 file, using the same codec as the source. The resulting video will have no audio:

   ffmpeg -i original_video.mp4 -map 0:v:1 -codec copy  rear_no_audio.mp4
   

2. Extract the audio stream into a temporary M4A file:

   ffmpeg -i original_video.mp4 -map 0:a:0 -codec copy  audio.m4a
   

3. Merge the two streams into the final MP4 file.

   ffmpeg -i rear_no_audio.mp4 -i audio.m4a -codec copy  rear_with_audio.mp4
   

The argument of the -map option includes v for video and a for audio streams. The second digit is the stream ID, which can be found by running

ffprobe original_video.mp4

After further research and experimenting, I've figured out how to do it using ffmpeg from command line.

1. Extract the relevant video stream (in my case track No. 2, which has index 1) into a temporary MP4 file, using the same codec as the source. The resulting video will have no audio:

   ffmpeg -i original_video.mp4 -map v:1 -codec copy  rear_no_audio.mp4
   

2. Extract the audio stream into a temporary M4A file:

   ffmpeg -i original_video.mp4 -map a:0 -codec copy  audio.m4a
   

3. Merge the two streams into the final MP4 file.

   ffmpeg -i rear_no_audio.mp4 -i audio.m4a -codec copy  rear_with_audio.mp4
   

The argument of the -map option includes v for video and a for audio streams. The second digit is the stream ID, which can be found by running

ffprobe original_video.mp4