You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Have you read the FAQ and checked for duplicate open issues?
Yes
Is your feature request related to a problem? Please describe.
I have been testing with some production content (which I cannot share the source of) but will describe and include some data below to set the scene.
There are several audio AdaptationSets representing two languages with stereo but two different audio codecs renditions, e.g.
language 1, stereo, mp4a.40.2, audio channel config 2
language 1, stereo, ac-3, audio channel config A000
language 2, stereo, mp4a.40.2, audio channel config 2
language 2, stereo, ac-3, audio channel config A000
Whilst language track selection is possible, and for stereo vs surround channels we can select and isolate tracks based on the channel count, it is not currently possible to isolate the stereo channel counts from each other. Shaka-Player treats all the ac-3 and mp4a.40.2 renditions as one adaptation set. I have concerns that audio gaps could happen during adaptation when you should only expect a small gap on track change.
The desire here would be to separate them based on their audio channel config and permit the user to select their choice of audio quality/codec.
Request content utilises the role attribute to differentiate alternative tracks.
Additional context
Here's the result of calling getVariantTracks() on an example stream.
Note I have change the language code to protect the obvious country of origin.
I have also preliminarily added support for audioChannelConfig into variant tracks, as a stepping stone to a potential solution and to demonstrate the different values that can hold. e.g.
2
A000
As you can guess I have been already experimenting here and can continue to a full merge request, just wanting to gain buy in and also peer review for any gotchas I might have missed in this direction.
Closing due to inactivity. If this is still an issue for you or if you have further questions, the OP can ask shaka-bot to reopen it by including @shaka-bot reopen in a comment.
Have you read the FAQ and checked for duplicate open issues?
Yes
Is your feature request related to a problem? Please describe.
I have been testing with some production content (which I cannot share the source of) but will describe and include some data below to set the scene.
There are several audio
AdaptationSet
s representing two languages with stereo but two different audio codecs renditions, e.g.mp4a.40.2
, audio channel config2
ac-3
, audio channel configA000
mp4a.40.2
, audio channel config2
ac-3
, audio channel configA000
Whilst language track selection is possible, and for stereo vs surround channels we can select and isolate tracks based on the channel count, it is not currently possible to isolate the stereo channel counts from each other. Shaka-Player treats all the
ac-3
andmp4a.40.2
renditions as one adaptation set. I have concerns that audio gaps could happen during adaptation when you should only expect a small gap on track change.The desire here would be to separate them based on their audio channel config and permit the user to select their choice of audio quality/codec.
Describe the solution you'd like
PreferenceBasedCriteria
addition:static filterVariantsByAudioChannelConfig_(variants, audioChannelConfig) {}
Potential addition of a config attribute alternative called
preferredAudioChannelConfig
default to 2.Extend the
selectAudioLanguage()
API to take an option audio channel config attribute as an alternative tochannelsCount
selectAudioLanguage(language, role, channelsCount = 0, safeMargin = 0, audioChannelConfig = '') {}
Describe alternatives you've considered
Request content utilises the
role
attribute to differentiate alternative tracks.Additional context
Here's the result of calling
getVariantTracks()
on an example stream.getVariantTracks() output
The text was updated successfully, but these errors were encountered: