1066102 - send information about the microphone level to the UI

Status: NEW

(Core :: WebRTC: Audio/Video, defect, P5)

Description

[Florian Quèze [:florian]]

The redesigned WebRTC UI in attachment 8485640 [details] has microphone meters both when permission to use a microphone is requested (image 2) and when a microphone is already shared (image 4).

There's currently no API providing this information to the UI.

Information from an IRC discussion on #media Monday:
21:20:28 - jesup: flo-retina: microphone: two options:  
21:20:30 - jesup: 1) provide preliminary selection info from the gui ("I'm open and need mic level for XXXX") back to media manager.  MediaManager would temporarily open the device, and instead of inserting into a stream would feed into level-sensing code.  That would throw notifications of level N times per second.  Lot of careful state management and error-case management on platform side
21:20:32 - jesup: 2) have the gui spawn off a privileged getUserMedia request for the mic, and feed the stream into a WebAudio level analyzer.  Downside: bunch of stuff called from (I presume) chrome.  Some management complexity though most would be handed by normal mechanisms.    Upside: reuse lots of existing stuff, fairly automatic error handling

I think 2) would cause us to start media manager on the chrome process, which if I understood correctly is something we would like to avoid for security reasons.

Comment 1

[Florian Quèze [:florian]]

23:54:53 - jesup: we don't want to be starting MediaManager/webrtc in the parent/UI process; huge sec/reliability issue.  I was presuming "preview" would fire off (periodic) small dumps of data (audio level, images) to the UI (cross-process as needed)
23:56:11 - flo-retina: so you would send the audio level or a small png several times per second?
23:58:03 - jesup: flo-retina: yes, that was the idea (audio maybe 5-10/s, video lower rate (and small)

Comment 2

[Bulk Bug Changes for mreavy's org]

Mass change P4->P5 to align with new Mozilla triage process.