Adding support for simulcast and spatial layers into VideoQualityTest

webrtc/modules/video_coding/codecs/vp9/vp9_impl.h

Index: webrtc/modules/video_coding/codecs/vp9/vp9_impl.h
diff --git a/webrtc/modules/video_coding/codecs/vp9/vp9_impl.h b/webrtc/modules/video_coding/codecs/vp9/vp9_impl.h
index c164a63980f2dedf54149e4f6e30477682f947e8..646b398729e5c51a6223b19f59189d6316e95ca2 100644
--- a/webrtc/modules/video_coding/codecs/vp9/vp9_impl.h
+++ b/webrtc/modules/video_coding/codecs/vp9/vp9_impl.h
@@ -56,6 +56,7 @@ class VP9EncoderImpl : public VP9Encoder {
                              const vpx_codec_cx_pkt& pkt,
                              uint32_t timestamp);
 
+  bool ExplicitlyConfiguredSpatialLayers() const;

[mflodman, 2015/10/06 11:27:27]

Are there cases we configure spatial layers implicitly and want to do things
differently? If not, can we name it something shorter like
'SpatialLayersConfigured()', 'UsingSpatialLayers()' or similar?

[mflodman, 2015/10/07 11:23:43]

Ping.

[ivica, 2015/10/07 11:59:50]

The manual configuration is currently used only from loopback tests currently,
where you specify bitrates directly via command line args. The other way is just
to let VP9Impl calculate them for you (it takes into account layer resolutions
and similar).
Because spatialLayers is not a std::vector, but a C-style array, I have to do
magic to detect if the user wants to manually specify bitrates. That's why I
created this helper method.
'UsingSpatialLayers()' wouldn't fit the usage, but 'SpatialLayersConfigued()'
might in some sense, but still not perfectly.

   bool SetSvcRates();
 
   virtual int GetEncodedLayerFrame(const vpx_codec_cx_pkt* pkt);