Re: [mediacapture-screen-share] Cropping Video Tracks (#195)

I am restricting the discussion to getDisplayMedia tracks, getViewPortMedia model is potentially different enough that I am unclear whether such a mechanism would be useful for getViewPortMedia. I am also assuming there will be some sort of CaptureHandle API to allow some form of communication between getDisplayMedia capturer and capturee.

The main drawback I see with the above cropping approach is that it requires tight coupling between capturee and capturer: they need to create a communication channel and define a custom protocol/API to exchange the region information.

We should aim at decoupling as much as possible capturer and capturee through a standard API.
Capturee can declare regions of interest through an API a la CaptureHandle, with the same origin protection mechanism.
Capturer would get and use that information from the MediaStreamTrack object itself.

Some additional thoughts:
- On capturer side, having a single object to handle (MediaStreamTrack) is nice, for instance in case the track is transferred to other contexts, say workers.
- I wonder whether it makes sense to allow User Agents to use that mechanism to expose subregions of window surfaces (say for instance a browser window MediaStreamTrack would expose a region that only includes the web page content, not the User Agent UI)? Might be somehow redundant with constraints though.

-- 
GitHub Notification of comment by youennf
Please view or discuss this issue at https://github.com/w3c/mediacapture-screen-share/issues/195#issuecomment-993566799 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Tuesday, 14 December 2021 13:55:57 UTC