On 13-01-22 6:34 PM, Adam Roach wrote: > I have to admit that [WebVTT's] relationship to WebRTC isn't > immediately obvious (at least to me). Could you give a short > executive summary of how you see them interacting?
AFAIK they don't interact at all, currently. The 'webvtt' project we're talking about in our implementation of the HTML <track> element. That's several pieces: a javascript API for manipulating text tracks, i.e. textual data associated with a playback timeline; a file format for expressing those tracks; and native controls for selecting among available text tracks and displaying them in sync with an HTML <video> element. This all assumes traditional delivery of static media files over HTTP. <video src=somefile.webm controls> <track src=somefile.vtt> </video> WebRTC is a framework for sending audio, video, and data streams directly between web user agents. It focusses on low latency and does not use HTTP for stream transmission, although the audio and video streams can be played through HTML <video> and <audio> elements. While there has been some discussion of supporting text streams, tty, caption data, etc. in WebRTC, my understanding is that there's no concensus to standardize any of them at this time, so there's no direct connection between the two features. What can be done is to send WebVTT 'cues', as the individual timeline elements are called, over an WebRTC data connection, and then use the TextTrack api to insert them into the playback context of a <video> element. This is only a few lines of code, and can simplify implementation of any particular subtitle or captioning delivery in the context of WebRTC. Does that help with the relative context? -r _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media
