Create a private API for controlling WebRTC's AEC3

chrome/common/extensions/api/webrtc_audio_private.idl

 // Copyright 2013 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 // The <code>chrome.webrtcAudioPrivate</code> API allows enumeration
 // of audio output (sink) devices as well as getting and setting the
 // active device for a given requesting process.
 //
 // Note that device IDs as used in this API are opaque (i.e. they are
 // not the hardware identifier of the device) and while they are
@@ skipping 32 matching lines @@
 
   dictionary RequestInfo {
     // The tab identifier from the chrome.tabs API, if the request is from a
     // tab.
     long? tabId;
     // The guest process id for the requester, if the request is from a
     // webview.
     long? guestProcessId;
   };
 
+  dictionary AudioExperiments {
+    // Enables or disables the new acoustic echo canceller (AEC3) in WebRTC.
+    boolean? enableAec3;
+  };
+
   interface Functions {
     // Retrieves a list of available audio sink devices.
     static void getSinks(GetSinksCallback callback);
 
     // Retrieves the currently active audio sink for the given requesting
     // process.
     static void getActiveSink(RequestInfo request,
                               SinkIdCallback callback);
 
     // Sets the active audio sink device for the specified requesting process.
     static void setActiveSink(RequestInfo request,
                               DOMString sinkId,
                               optional CompletionCallback callback);
 
     // Given a security origin and an input device ID valid for that
     // security origin, retrieve an audio sink ID valid for the
     // extension, or the empty string if there is no associated audio
     // sink.
     //
     // The associated sink ID can be used as a sink ID for
     // setActiveSink. It is valid irrespective of which process you are
     // setting the active sink for.
     static void getAssociatedSink(DOMString securityOrigin,
                                   DOMString sourceIdInOrigin,
                                   SinkIdCallback cb);
+
+    // Provides settings to experiments related to WebRTC audio.
+    static void setAudioExperiments(RequestInfo request,

[Devlin, 2017/04/10 21:34:25]

Document the parameters this accepts. Also, probably add an optional callback
unless there's a concrete reason not to.

[Henrik Grunell, 2017/04/11 06:49:39]

I asked the author to remove the callback that was originally there since there
is no need to signal success/failure (and it was actually never run in the
original patch). There's also discussion in another comment whether the
application would need to know when the operation has finished. If that's not
necessary either, I see no reason for a callback.

[hlundin-chromium, 2017/04/11 09:19:37]

Added documentation.

[Devlin, 2017/04/11 20:49:37]

The main reason is consistency with other APIs and because it's difficult to
predict all use cases (it's entirely possible that there could be a reason to
wait for the operation to complete).  But, given it's a private API, we can hold
off if you prefer.

[hlundin-chromium, 2017/04/12 05:54:16]

Time is not on our side now, so I'll prefer to hold off.

+                                    DOMString securityOrigin,
+                                    AudioExperiments audioExperiments);
   };
 
   interface Events {
     // Fired when audio sink devices are added or removed.
     static void onSinksChanged();
   };
 };