commit e953224613c3773de8398de05a6f28b7bed9e334
author Jingning Han <jingning@google.com> Fri Nov 20 02:06:04 2020 -0800
committer Jingning Han <jingning@google.com> Fri Nov 20 19:52:02 2020 +0000
tree 0cbae0448b4fece67a38208ec083fcec6b453a32
parent de22cec5e6a793dea02eb5e7e0b5b64351da3ca4

Allow all modes to reuse wedge index search results

Apply the prior wedge index search results to all inter modes. This
and its parent CL together improve the encoding speed by more than
10%. The coding performance loss in speed 1:

        overall PSNR     SSIM
lowres   0.056%          0.032%
midres2  0.033%          0.015%

STATS_CHANGED

Change-Id: Ib4e086b924cf4321d01ad73c97db5b5d0f22e849

av1/encoder/compound_type.c

diff --git a/av1/encoder/compound_type.c b/av1/encoder/compound_type.c
index fd39bde..4256289 100644
--- a/av1/encoder/compound_type.c
+++ b/av1/encoder/compound_type.c
@@ -1355,8 +1355,7 @@
       int best_rs2 = 0;
       int best_rate_mv = *rate_mv;
       const int wedge_mask_size = get_wedge_types_lookup(bsize);
-      int need_mask_search =
-          args->wedge_index == -1 || !have_newmv_in_inter_mode(this_mode);
+      int need_mask_search = args->wedge_index == -1;
 
       for (int wedge_mask = 0; wedge_mask < wedge_mask_size && need_mask_search;
            ++wedge_mask) {
@@ -1398,7 +1397,7 @@
       }
 
       if (need_mask_search) {
-        if (have_newmv_in_inter_mode(this_mode)) {
+        if (this_mode == NEW_NEWMV) {
           args->wedge_index = best_mask_index;
           args->wedge_sign = best_wedge_sign;
         }
@@ -1407,8 +1406,11 @@
         mbmi->interinter_comp.wedge_sign = args->wedge_sign;
         rs2 = masked_type_cost[cur_type];
         rs2 += get_interinter_compound_mask_rate(&x->mode_costs, mbmi);
-        tmp_rate_mv = av1_interinter_compound_motion_search(cpi, x, cur_mv,
-                                                            bsize, this_mode);
+
+        if (have_newmv_in_inter_mode(this_mode)) {
+          tmp_rate_mv = av1_interinter_compound_motion_search(cpi, x, cur_mv,
+                                                              bsize, this_mode);
+        }
 
         best_mask_index = args->wedge_index;
         best_wedge_sign = args->wedge_sign;
@@ -1437,8 +1439,7 @@
       int best_mask_index = 0;
       rs2 += get_interinter_compound_mask_rate(&x->mode_costs, mbmi);
 
-      int need_mask_search =
-          args->diffwtd_index == -1 || !have_newmv_in_inter_mode(this_mode);
+      int need_mask_search = args->diffwtd_index == -1;
 
       for (int mask_index = 0; mask_index < 2 && need_mask_search;
            ++mask_index) {
@@ -1470,8 +1471,7 @@
       }
 
       if (need_mask_search) {
-        if (have_newmv_in_inter_mode(this_mode))
-          args->diffwtd_index = best_mask_index;
+        if (this_mode == NEW_NEWMV) args->diffwtd_index = best_mask_index;
       } else {
         mbmi->interinter_comp.mask_type = args->diffwtd_index;
         rs2 = masked_type_cost[cur_type];
@@ -1480,8 +1480,10 @@
         int mask_value = mbmi->interinter_comp.mask_type == 0 ? 38 : 26;
         memset(xd->seg_mask, mask_value, sizeof(xd->seg_mask));
 
-        tmp_rate_mv = av1_interinter_compound_motion_search(cpi, x, cur_mv,
-                                                            bsize, this_mode);
+        if (have_newmv_in_inter_mode(this_mode)) {
+          tmp_rate_mv = av1_interinter_compound_motion_search(cpi, x, cur_mv,
+                                                              bsize, this_mode);
+        }
         best_mask_index = mbmi->interinter_comp.mask_type;
         tmp_mv[0] = mbmi->mv[0];
         tmp_mv[1] = mbmi->mv[1];