rtc-svc: Fix to make SVC work for Profile 1

Added datarate unittest for 4:4:4 and 4:2:2 input,
for spatial and temporal layers.

Fix is needed in vp9_set_literal_size():
the sampling_x/y should be passed into update_inital_width(),
othewise sampling_x/y = 1/1 (4:2:0) was forced.
vp9_set_literal_size() is only called by the svc and
on dynamic resize.

Fix issue with the normative optimized scaler:
UV width/height was assumed to be 1/2 of Y, for
the ssse and neon code.

Also fix to assert for the scaled width/height:
in case scaled width/height is odd it should be
incremented by 1 (make it even).

Change-Id: I3a2e40effa53c505f44ef05aaa3132e1b7f57dd5

diff --git a/test/svc_datarate_test.cc b/test/svc_datarate_test.cc
index 291cb01..f3f76a0 100644 (file)
--- a/test/svc_datarate_test.cc
+++ b/test/svc_datarate_test.cc
@@ -548,13 +548,16 @@ class DatarateOnePassCbrSvc : public OnePassCbrSvc {
       }
 
       if (!single_layer_resize_) {
-        ASSERT_EQ(pkt->data.frame.width[sl],
-                  top_sl_width_ * svc_params_.scaling_factor_num[sl] /
-                      svc_params_.scaling_factor_den[sl]);
-
-        ASSERT_EQ(pkt->data.frame.height[sl],
-                  top_sl_height_ * svc_params_.scaling_factor_num[sl] /
-                      svc_params_.scaling_factor_den[sl]);
+        unsigned int scaled_width = top_sl_width_ *
+                                    svc_params_.scaling_factor_num[sl] /
+                                    svc_params_.scaling_factor_den[sl];
+        if (scaled_width % 2 != 0) scaled_width += 1;
+        ASSERT_EQ(pkt->data.frame.width[sl], scaled_width);
+        unsigned int scaled_height = top_sl_height_ *
+                                     svc_params_.scaling_factor_num[sl] /
+                                     svc_params_.scaling_factor_den[sl];
+        if (scaled_height % 2 != 0) scaled_height += 1;
+        ASSERT_EQ(pkt->data.frame.height[sl], scaled_height);
       } else if (superframe_count_ > 0) {
         if (pkt->data.frame.width[sl] < prev_frame_width[sl] &&
             pkt->data.frame.height[sl] < prev_frame_height[sl])
@@ -678,6 +681,70 @@ class DatarateOnePassCbrSvcSingleBR
   }
 };
 
+// Check basic rate targeting for 1 pass CBR SVC: 3 spatial layers and 3
+// temporal layers, for 4:4:4 Profile 1.
+TEST_P(DatarateOnePassCbrSvcSingleBR, OnePassCbrSvc3SL3TL444Profile1) {
+  SetSvcConfig(3, 3);
+  ::libvpx_test::Y4mVideoSource video("rush_hour_444.y4m", 0, 140);
+  cfg_.g_profile = 1;
+  cfg_.g_bit_depth = VPX_BITS_8;
+  cfg_.rc_buf_initial_sz = 500;
+  cfg_.rc_buf_optimal_sz = 500;
+  cfg_.rc_buf_sz = 1000;
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.g_threads = 1;
+  cfg_.rc_dropframe_thresh = 0;
+  cfg_.kf_max_dist = 9999;
+
+  top_sl_width_ = 352;
+  top_sl_height_ = 288;
+  cfg_.rc_target_bitrate = 500;
+  ResetModel();
+  AssignLayerBitrates();
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  CheckLayerRateTargeting(number_spatial_layers_, number_temporal_layers_, 0.78,
+                          1.15);
+#if CONFIG_VP9_DECODER
+  // The non-reference frames are expected to be mismatched frames as the
+  // encoder will avoid loopfilter on these frames.
+  EXPECT_EQ(GetNonRefFrames(), GetMismatchFrames());
+#endif
+}
+
+// Check basic rate targeting for 1 pass CBR SVC: 2 spatial layers and 3
+// temporal layers, for 4:2:2 Profile 1.
+TEST_P(DatarateOnePassCbrSvcSingleBR, OnePassCbrSvc2SL3TL422Profile1) {
+  SetSvcConfig(2, 3);
+  ::libvpx_test::Y4mVideoSource video("park_joy_90p_8_422.y4m", 0, 20);
+  cfg_.g_profile = 1;
+  cfg_.g_bit_depth = VPX_BITS_8;
+  cfg_.rc_buf_initial_sz = 500;
+  cfg_.rc_buf_optimal_sz = 500;
+  cfg_.rc_buf_sz = 1000;
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.g_threads = 1;
+  cfg_.rc_dropframe_thresh = 0;
+  cfg_.kf_max_dist = 9999;
+
+  top_sl_width_ = 160;
+  top_sl_height_ = 90;
+  cfg_.rc_target_bitrate = 500;
+  ResetModel();
+  AssignLayerBitrates();
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  // Use large under/over shoot thresholds as this is a very short clip,
+  // so not good for testing rate-targeting.
+  CheckLayerRateTargeting(number_spatial_layers_, number_temporal_layers_, 0.5,
+                          1.7);
+#if CONFIG_VP9_DECODER
+  // The non-reference frames are expected to be mismatched frames as the
+  // encoder will avoid loopfilter on these frames.
+  EXPECT_EQ(GetNonRefFrames(), GetMismatchFrames());
+#endif
+}
+
 // Check basic rate targeting for 1 pass CBR SVC: 2 spatial layers and 1
 // temporal layer, with screen content mode on and same speed setting for all
 // layers.

diff --git a/vp9/encoder/arm/neon/vp9_frame_scale_neon.c b/vp9/encoder/arm/neon/vp9_frame_scale_neon.c
index e46f789..69b8cff 100644 (file)
--- a/vp9/encoder/arm/neon/vp9_frame_scale_neon.c
+++ b/vp9/encoder/arm/neon/vp9_frame_scale_neon.c
@@ -710,8 +710,8 @@ void vp9_scale_and_extend_frame_neon(const YV12_BUFFER_CONFIG *src,
   const int src_h = src->y_crop_height;
   const int dst_w = dst->y_crop_width;
   const int dst_h = dst->y_crop_height;
-  const int dst_uv_w = dst_w / 2;
-  const int dst_uv_h = dst_h / 2;
+  const int dst_uv_w = dst->uv_crop_width;
+  const int dst_uv_h = dst->uv_crop_height;
   int scaled = 0;
 
   // phase_scaler is usually 0 or 8.

diff --git a/vp9/encoder/vp9_encoder.c b/vp9/encoder/vp9_encoder.c
index 85bd706..371779e 100644 (file)
--- a/vp9/encoder/vp9_encoder.c
+++ b/vp9/encoder/vp9_encoder.c
@@ -8155,9 +8155,11 @@ int vp9_set_size_literal(VP9_COMP *cpi, unsigned int width,
                          unsigned int height) {
   VP9_COMMON *cm = &cpi->common;
 #if CONFIG_VP9_HIGHBITDEPTH
-  update_initial_width(cpi, cm->use_highbitdepth, 1, 1);
+  update_initial_width(cpi, cm->use_highbitdepth, cpi->common.subsampling_x,
+                       cpi->common.subsampling_y);
 #else
-  update_initial_width(cpi, 0, 1, 1);
+  update_initial_width(cpi, 0, cpi->common.subsampling_x,
+                       cpi->common.subsampling_y);
 #endif  // CONFIG_VP9_HIGHBITDEPTH
 
 #if CONFIG_VP9_TEMPORAL_DENOISING

diff --git a/vp9/encoder/x86/vp9_frame_scale_ssse3.c b/vp9/encoder/x86/vp9_frame_scale_ssse3.c
index 7685e7b..bf0e8b1 100644 (file)
--- a/vp9/encoder/x86/vp9_frame_scale_ssse3.c
+++ b/vp9/encoder/x86/vp9_frame_scale_ssse3.c
@@ -754,8 +754,8 @@ void vp9_scale_and_extend_frame_ssse3(const YV12_BUFFER_CONFIG *src,
   const int src_h = src->y_crop_height;
   const int dst_w = dst->y_crop_width;
   const int dst_h = dst->y_crop_height;
-  const int dst_uv_w = dst_w / 2;
-  const int dst_uv_h = dst_h / 2;
+  const int dst_uv_w = dst->uv_crop_width;
+  const int dst_uv_h = dst->uv_crop_height;
   int scaled = 0;
 
   // phase_scaler is usually 0 or 8.