Add method cricket::VideoCapturer::NeedsDenoising, use in VideoCapturerTrackSource.

webrtc/media/base/videocapturer.h

 /*
  *  Copyright (c) 2010 The WebRTC project authors. All Rights Reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS.  All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
@@ skipping 157 matching lines @@
   CaptureState capture_state() const {
     return capture_state_;
   }
 
   virtual bool apply_rotation() { return apply_rotation_; }
 
   // Returns true if the capturer is screencasting. This can be used to
   // implement screencast specific behavior.
   virtual bool IsScreencast() const = 0;
 
+  // Returns denoising preference. Unset means use encoder-specific default.

[perkj_webrtc, 2016/09/14 13:08:27]

This should not have anything todo with an encoder. So I think this comment
should not mention an encoder.
 

[nisse-webrtc, 2016/09/14 13:20:23]

Behavior if unset depends on the encoder. I think that's relevant.   

[perkj_webrtc, 2016/09/14 13:28:22]

Can you elaborate on this. To me- the whole thing is very weird and hacky. Is
there a better comment in VideoTrackSource? 
Maybe also add a todo and a comment why this is added at all here.

+  virtual rtc::Optional<bool> NeedsDenoising() const {

[perkj_webrtc, 2016/09/14 13:08:27]

Also - I think this api should be identical to IsScreenCast. 
Its the same hack. 
virtual bool NeedsDenoising() { return false; }

[nisse-webrtc, 2016/09/14 13:20:23]

I'm afraid it has to be three-state, it's passed on to be returned by
VideoCapturerTrackSource::needs_denoising, then copied into VideoOptions for
SetVideoSend by VideoRtpSender.

For consistency, we could change IsScreencast to the three-state too, but I
don't think that is a real improvement.

[perkj_webrtc, 2016/09/14 13:28:22]

Acknowledged.

+    return rtc::Optional<bool>();
+  }
+
   // Caps the VideoCapturer's format according to max_format. It can e.g. be
   // used to prevent cameras from capturing at a resolution or framerate that
   // the capturer is capable of but not performing satisfactorily at.
   // The capping is an upper bound for each component of the capturing format.
   // The fourcc component is ignored.
   void ConstrainSupportedFormats(const VideoFormat& max_format);
 
   void set_enable_camera_list(bool enable_camera_list) {
     enable_camera_list_ = enable_camera_list;
   }
@@ skipping 139 matching lines @@
   bool apply_rotation_;
 
   // State for the timestamp translation.
   rtc::TimestampAligner timestamp_aligner_;
   RTC_DISALLOW_COPY_AND_ASSIGN(VideoCapturer);
 };
 
 }  // namespace cricket
 
 #endif  // WEBRTC_MEDIA_BASE_VIDEOCAPTURER_H_