W3C home > Mailing lists > Public > public-webrtc@w3.org > November 2011

Re: Where to attach a DTMF API

From: Randell Jesup <randell-ietf@jesup.org>
Date: Tue, 29 Nov 2011 11:19:01 -0500
Message-ID: <4ED505F5.5040800@jesup.org>
To: public-webrtc@w3.org
On 11/29/2011 5:23 AM, Timothy B. Terriberry wrote:
>>> How should we handle these cases? Can we create a null audio track using
>>> the current API?
>> How would you do it for a legacy client? My understanding was that
>> RFC4733 required and existing (audio) RTP stream (identified by SSRC) to
>> insert its DTMF payloads into.
> I think the question is simply whether or not we can create such an RTP
> stream without acquiring a MediaStream via getUserMedia(). roc's
> MediaStream Processing API certainly provides a way to do this. Chris
> Rogers' Web Audio API doesn't seem to provide MediaStream objects that I
> can tell, but I assume that's something that could be fixed.

This is a case where we want to create a one-way-connection, or maybe 
one where the user doesn't want to give mic access (but does want the 
call to go through).

While microphone/file-source access needs to be use-approved, we could 
allow an application to create a 'null' audio (or video) mediatrack 
without permission (silence for audio, black (or black with "no video" 
or "camera off" or a camera with a red circle-slash)).  That could then 
be used in place of a GetUserMedia() stream.

We may have uses for this elsewhere, perhaps for muting (you don't want 
to wait for signaling on mute, and just stopping sending is a problem in 
some cases, especially if the signaling channel is slow/broken) or to 
open channels immediately while waiting for the user to respond about 
media access.  However, these may be spurious uses if the media tracks 
support a Mute operation.

Perhaps more to the point, being able to switch the source for a channel 
from the mic to self-generated data (silence, "the doctor will be with 
you shortly", black, a test pattern, camera-with-slash, etc) is more the 
point of what we need, and for this we'd just need to be able to build a 
stream using a source like that instead of switching the source of an 
existing stream/track.

Randell Jesup
Received on Tuesday, 29 November 2011 16:20:55 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 15:19:26 UTC