- From: Elad Alon via GitHub <sysbot+gh@w3.org>
- Date: Fri, 31 Jan 2025 09:43:19 +0000
- To: public-webrtc-logs@w3.org
This makes both reasonable and unreasonable assumptions. **Reasonable** assumptions: 1. The existence of a preview tile. **Unreasonable** assumptions: 1. The preview tile is implemented using a video tile. * Alternative not considered here - canvas. (Note: Overlay can also be a canvas; does not change anything, but has been the source of confusion in the past. Let's call them preview-canvas and overlay-canvas going forward.) 2. The video frames are an exact match to the preview tile's bounding box. * What if the aspect ratio is wrong and letterboxing occurs? * What if the frames are intentionally scaled down and only take up a portion of the preview-canvas? (Or the same thing for videos using Media Capture Transform, allowing injection of content in the space left behind after scaling.) 3. The video frames are trivially associable with the captured surface without any indication from the app. * Inherently challenging with canvas-previews. * Even for video-based previews, this is not a reasonable assumption. What if the Web app uses Media Capture Transform to combine the shared content and the camera content and produce a [weatherman effect](https://www.youtube.com/watch?v=fHxO0UdpoxM)? We should carefully weigh the undesirable limitations this proposal would produce on Web developers, against the likelihood, frequency and magnitude of bugs this would prevent. IMHO, the issues this proposal seeks to avert are unlikely, minor and easy to fix by Web applications. -- GitHub Notification of comment by eladalon1983 Please view or discuss this issue at https://github.com/w3c/mediacapture-surface-control/issues/69#issuecomment-2626774199 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Friday, 31 January 2025 09:43:19 UTC