Commentary tracks complicate things.
Usually when a TV sees that one audio track is in an unsupported format it will automatically skip it to play the next supported one.
When UMS checks if a video is supported it used to allow it to stream as long as the video and at least 1 audio stream is supported.
Now if the first track it unsupported it will transcode.
There is currently no good way for UMS to know if the other audio tracks are commentary as usually they just have the same language tag, or it might just be the same audio but in AC3 instead of DTS.
I think attempts have been made to improve this but has been unreliable with too many false positives.
One problem with transcoding video is that it does not preserve the HDR data.
I think this is because FFmpeg does not report something in the metadata so UMS cannot use in the command to save it, but I know very little of the process.
And it is better to be able to play a video even if something is lost than not being able to play at all.
I think this is a common problem with other media servers.
The way around to preserve HDR is to remux (Passthrough) instead of transcode so it copies the video stream while converting the unsupported audio.
However as the output only has one audio stream and no subtitles, if there are subtitles that are required these have to be burned in which means a full transcode of the video which will lose HDR.
I think this is a limitation on the MPEGTS container that does not support multiple streams like a MKV file has.
Some devices will support streaming subs separately and this might be a workaround if you need them. I have altered this conf to do that.
This will only work with external sub files though. It would be too slow to extract internal subs first to stream them as it has been tried before.
If you do not want subs then make sure there are no sub files in the same folder as the video (.srt) and turn off subtitles in the settings and it should just remux automatically if the audio uses DTS.
I think the best solution is to avoid videos that use DTS altogether and find another source.
Or you could manually remux/transcode DTS audio tracks separately so UMS can just stream it to the TV unaltered. You do not need to remove the subs either.
There are several ways to do this like doing it by hand with FFmpeg, but something like MKVtools would probably be easier if you had to do it regularly.
I think you can set up some automation to do this and should not take long to only change the audio.
When using the hls transcoding method it is basically a playlist of tiny segments that can be switched easily to another audio track/subtitle which is more flexible. Its the same way Youtube works.
Because of this it will show as a *-mpegURL file instead of the original .mkv
The same thing happens with normal transcoding and would show as .mpeg for example.
When it works HLS should be like if it was streaming instead of transcoding and allow switching of audio and subs during playback.
I checked the packing logs box on v15.8.0 and it does not give the option to attach the renderer conf like it should.

- Screenshot 2026-08-21 113750.png (20.93 KiB) Viewed 14918 times
I'll report this as it should be including them.