When there are two audio renderers (e.g. because the FFmpeg extension is being used), track groups are mapped to both of them, and the application specifies a constraint on track selection, DefaultTrackSelector will always prefer a track mapped to the first renderer, even if the constraint is better satisfied by a track in the second selection. We should instead select a track for the second audio renderer in this case.
The same argument may apply for video as well, although this might be less clear cut.
When there are two audio renderers (e.g. because the FFmpeg extension is being used), track groups are mapped to both of them, and the application specifies a constraint on track selection,
DefaultTrackSelectorwill always prefer a track mapped to the first renderer, even if the constraint is better satisfied by a track in the second selection. We should instead select a track for the second audio renderer in this case.The same argument may apply for video as well, although this might be less clear cut.