Provide command line arg to pass an audio file to be used as a mic input in combination with media.navigator.streams.fake
Categories
(Core :: WebRTC: Audio/Video, task, P5)
Tracking
()
People
(Reporter: damencho, Unassigned)
Details
Steps to reproduce:
We (Jitsi Team) have a suite of tests testing WebRTC functionality.
Many tests which depend on dominant speaker we need to disable those tests when Firefox is involved.
Actual results:
The reason is that we cannot specify an audio file to be used as mic input when using "media.navigator.streams.fake". For another browser, we use a file with a natural speech and pass it as a command line argument.
Expected results:
Passing a file on the local filesystem as command line arg to be used as a mic input when "media.navigator.streams.fake" flag is true.
|
||
Comment 1•1 month ago
|
||
The Bugbug bot thinks this bug should belong to the 'Core::WebRTC' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.
|
||
Comment 2•1 month ago
|
||
We have lots to do but this isn't a priority, sorry. I'm marking P5 because I could be open to accepting a patch.
As an alternative solution I would suggest using virtual mics/loopback devices. We do this in our CI where the test itself can then decide what a virtual mic should receive by playing the audio to the corresponding speaker device. Passing data to the loopback device could also be handled in the environment/test runner, of course.
Let me know if you need help getting Firefox to play along with such a setup. There are some prefs that are handy for bypassing permissions.
A second alternative would be to override getUserMedia with a WebAudio-based polyfill. Then you could get audio tracks that contain any content you choose.
Description
•