- From: guidou via GitHub <sysbot+gh@w3.org>
- Date: Tue, 01 Jun 2021 15:52:18 +0000
- To: public-webrtc-logs@w3.org
> > It seems that we have agreement that processing on main thread should not be impossible. > > I would phrase it this way: > > * There is no consensus on a goal to provide an API that allows processing on main thread. > * There is consensus on not having a goal to forbid main thread processing. I agree with this phrasing. > > The current spec would be really hard to implement in an interoperable manner. Are you referring to the streams algorithms or something else specifically? > Some JS exposed behaviours are actually not specified for instance. Can you file specific bugs about those behaviors? > Describing these algorithms will allow reaching interoperability. I'll prepare a PR to include those algorithms. Tracked on #36. > * We have MediaStreamTrack data that will often be backed by GPU data > * We have WebGPU API that will allow processing GPU data > * Is this API able to do WebGPU processing on MediaStreamTrack data in an optimal manner, ideally no conversion/no memory copies? If not, do we have a good idea on how to add that support at a later stage. I think we should continue this discussion on a separate bug, perhaps on the WebCodecs spec if we agree that WebCodec's VideoFrame is the right API to expose video frames coming from a MediaStreamTrack. > > > > * Can one JS document hose a browser's camera capturing ability? #37 > > > > > > Not by default, but yes, if the document reads enough frames and does not close them a camera capturer can suspend producing frames. > > The general question is how much we can reduce this risk and whether APIs could be designed with this in mind. > For instance, the current API is exposing a pattern that seems nice: when enqueueing a frame in a MediaStreamTrack generator, the frame gets consumed. > I would hope we can continue improving the design to further reduce these risks. I agree that the goal should be risk reduction since completely eliminating this possibility would negate the goal of optimal WebGPU processing. -- GitHub Notification of comment by guidou Please view or discuss this issue at https://github.com/w3c/mediacapture-transform/issues/38#issuecomment-852236943 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Tuesday, 1 June 2021 15:52:46 UTC