Separate audo & video tracks

# Background

Currently, videos from YouTube will in many cases have distinct audio tracks, we are working around this in YouTube-Synch by doing on-the-fly merging of audio and video in such cases, but it is wasteful and complex. Distinct audio tracks may also support things like having distinct tracks in different languages, although that may require distinct metadata treatment.

There is a similar issue here https://github.com/Joystream/joystream/issues/3806, but it got too complex to really be a good basis for continued work on this issue.

# Proposal

Introduce changes in metadata standard for videos to allow this, however, in the future we would like to have adaptive streaming (e.g. DASH), hence it should be resolved whether these two initiatives really are distinct, or can be combined.