Should we document in the spec that there is no consensus on whether supporting MediaStreamTrackProcessor for audio?

[youennf]

Should we document in the spec that there is no consensus on whether supporting MediaStreamTrackProcessor for audio?

[youennf]

The spec says:

Note: There is no WG consensus on whether or not audio use cases should be supported.

It is precise about audio track generator.
But it is not precise about MediaStreamTrackProcessor.
It might be good to add a note similar to the one done for VideoTrackGenerator.

[youennf]

I uploaded #104 which partially fixes the issue.
What remains is what to do when MediaStreamTrackProcessor constructor is given an audio track.
The spec should probably say something, safari current implementation throws.

@jan-ivar, @alvestrand, thoughts?

[youennf]

I guess audio track throwing may be already allowed since it might not be considered as an invalid track.
This could be tackled in #94.

[guidou]

We should reopen this debate. In Chrome, MSTP for audio is used 3X more than MSTP for video.
I don't see any argument not to support it. The real-time one is not valid, because a lot of the applications (presumably all of the ones currently using it) do not need to run on a real-time thread (in fact, for many such applications a real-time thread would be worse than a worker thread).

[youennf]

We should reopen this debate

That is a different issue, that we can tackle on with WG discussions.

The issue here is specifically about representing in the spec the current state of discussions, nothing less, nothing more. This is more editorial than anything else (except the part to represent lack of audio support).

[guidou]

That is a different issue, that we can tackle on with WG discussions.

The issue here is specifically about representing in the spec the current state of discussions, nothing less, nothing more. This is more editorial than anything else (except the part to represent lack of audio support).

Absolutely. I commented on the wrong issue. I agree that the spec should reflect that and your PR LGTM.