There is more than one way to achieve what you have planned.
Method one would be to encode the same Video Stream with several Audio Streams, I believe that mkv files support this. You could try and concatenate multiple audio tracks using ffmpeg. The tricky part would be selecting the correct audio track to play, but that depends on your embedded video-...
If you aren't adverse to using a terminal to extract it, FFmpeg is a very good tool that you can use. The command that you'd want to run in ffmpeg -i [input file name here] -c:a copy -vn -sn [output file name here].
-c:a copy: This argument copies the audio stream of the file.
-vn: This argument informs FFmpeg to not account for any video streams.
-sn: This ...
When using sound for speech-to-text, the quality is actually not very important. You can extract audio using ffmpeg, After Effects, Premiere, Nuke, Resolve or any editing software you like. Load in the clip, export again but as a .wav. Boom, you have only the sound.