Add method on AVFoundation capturer to adapt output format.

webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h

Index: webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h
diff --git a/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h b/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h
index 2a732b9686b877644541814da1d6a025611362d3..254196d9e1614616319516ee37ca008011324d4f 100644
--- a/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h
+++ b/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h
@@ -29,6 +29,16 @@ RTC_EXPORT
 
 - (instancetype)init NS_UNAVAILABLE;
 
+/**
+ * Calling this function will cause frames to be scaled down to the
+ * requested resolution. Also, frames will be cropped to match the
+ * requested aspect ratio, and frames will be dropped to match the
+ * requested fps. The requested aspect ratio is orientation agnostic and
+ * will be adjusted to maintain the input orientation, so it doesn't
+ * matter if e.g. 1280x720 or 720x1280 is requested.
+ */
+- (void)adaptOutputFormatToWidth:(int)width height:(int)height fps:(int)fps;

[tkchin_webrtc, 2016/11/30 22:08:52]

is there a way to do this without touching the capturer directly like this?
ie what is the "WebRTC" way of doing things here

[magjed_webrtc, 2016/12/01 09:31:44]

I think this is the "WebRTC" way of doing it. We have pushed all cropping and
scaling to the video source since it's most efficient to do it as early as
possible (this was more a concern when we did like 5 frame copies in the
pipeline before scaling). So the one scaler we have is a cricket::VideoAdapter
in the video source. CPU overuse detector and QualityScaler will signal back to
this cricket::VideoAdapter. I think it makes sense to also expose a direct way
in the video source to control the VideoAdapter. This is how it's done for
Android as well.

One alternative might be to use media constraints in
PeerConnectionFactory::CreateVideoSource to control this, but we are trying to
move away from media constraints for that function, see
https://cs.chromium.org/chromium/src/third_party/webrtc/api/peerconnectionfac....
We are trying to move towards a more simple and direct interface and leave the
decision logic to the app. So the app has to select one of the supported
resolutions from the capturer and can also control the VideoAdapter directly.

[tkchin_webrtc, 2016/12/02 22:02:20]

I'm totally with you that mobile is a different world, and should be treated
differently when needed, so this change looks good. I guess what I meant was,
how would you do this in JS? And I guess the answer is, mobile is different.

+
 /** Returns whether rear-facing camera is available for use. */
 @property(nonatomic, readonly) BOOL canUseBackCamera;