- From: Harald Alvestrand <harald@alvestrand.no>
- Date: Fri, 08 Feb 2013 22:28:25 +0100
- To: stephane.cazeaux@orange.com
- CC: public-webrtc@w3.org
- Message-ID: <51156DF9.5020200@alvestrand.no>
On 02/08/2013 08:53 PM, stephane.cazeaux@orange.com wrote: > > Hi all > > I am wondering if something is not missing in audio device selection > for some communication case. > Well, first off, this is a media capture matter, so should go to public-media-capture@w3.org.... > Regarding to audio device selection, the media capture scenario > document > (https://dvcs.w3.org/hg/dap/raw-file/tip/media-stream-capture/scenarios.html) > talks of the case where the user selects between a microphone device > plugged to the audio jack, or the microphone of a webcam. But there is > another common case: the user has a microphone in the audio jack and a > usb headset (speaker+microphone). To me, in this case: > > -When the user selects the microphone device in the audio jack, he > will most likely want to hear the remote party in the speakers of the > computer > > -When the user selects the USB headset device, he will most likely > want to hear the remote party in the headset > "Most likely" isn't the same as "always". I regularly use a headset with a broken microphone; in those cases, I need to select the camera microphone (because that's the one that gives the best sound). I think it is better that the API present those as independent choices. > But with some tests based on Chrome 24, the audio is always played > through the OS default audio device, even if I select a usb headset as > audio capture device (which is of course not the OS default device). > > AFAIK, this is explained by the fact that on one hand the media > capture API enables the selection of the capture device only. On the > other hand, audio HTML tag uses the default rendering device of the > system. I could not find any document talking of audio rendering > device selection. > Chrome 24 doesn't have input microphone selection yet. Please don't judge the specification by the state of the implementation. > Any thought about this? Did I miss something? > > Stéphane. > > ** >
Received on Friday, 8 February 2013 21:28:57 UTC