Hi all, Use Case: 1. MediaStream will be used to represent TV broadcasting stream. 2. Then there are multiple MediaStreamTrack to represent multiple audio/video tracks in TV stream. 3. Web content will put this MediaStream into video tag to play the TV content. 4. Web content can check or select different audio languages or visual angles from video tag's multitracks API. 5. This introduced that MediaStreamTrack needs to relay enough information to construct Audio/VideoTrack.
After discussing with Shelly offline, we think that both of Constraints and MediaStreamTrack::lable are not suitable to append extra information from AudioTrack::kind/languange. And according to these extra info doesn't expose to Web Content, we can just create new member variables in class MediaStreamTrack to record them. Then finally mapping these new variables from MediaStreamTrack to AudioTrack. Appreciate for any suggestion. Thanks, Sincerely yours. ----- 原始郵件 ----- 寄件者: "Shelly Lin" <[email protected]> 收件者: [email protected] 寄件備份: 2014 4 月 16 星期三 下午 4:11:58 主旨: Mapping AudioStreamTrack to AudioTrack Hi everyone, I've picked up the implementation of bug 744896 to support multi-tracks media sources from DTV. https://bugzilla.mozilla.org/show_bug.cgi?id=744896 Although the Tuner API is till in draft, it is pretty solid that the source of a tuner would be MediaStream. http://airpingu.github.io/tv-tuner-api/index.html#idl-def-TVTuner The problem is that the mapping of AudioStreamTrack to AudioTrack is not one-on-one, for example: An audio track has: interface AudioTrack { readonly attribute DOMString id; readonly attribute DOMString kind; --> "alternative", "descriptions", "main" readonly attribute DOMString label; --> more like free string for users(resource) to define readonly attribute DOMString language; --> "eng", "fra" attribute boolean enabled; }; For AudioTrack.kind() http://www.w3.org/TR/html5/embedded-content-0.html#dom-audiotrack-kind For AudioTrack.language() http://www.w3.org/TR/html5/embedded-content-0.html#dom-audiotrack-language An audio stream track has: interface MediaStreamTrack : EventTarget { readonly attribute DOMString kind; --> "audio" or "video" readonly attribute DOMString id; readonly attribute DOMString label; --> Same as the label in audio track attribute boolean enabled; readonly attribute boolean muted; Should I use Constraints to setup the information when constructing a stream track? http://dev.w3.org/2011/webrtc/editor/getusermedia.html#tracks-and-constraints Or, if this mapping will not reveal to js anyway, is it okay to just implement the mapping internally? Say the label attribute of MediaStreamTrack contains a field of "kind" and a field of "language", so that we can parse them out at constructing an audio track from media stream. Cheers, Shelly _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media
