Move the SATD model pruning in the intra mode decision
Allow directional intra prediction modes to use SATD model to
test angle changes. The compression efficiency is improved by
0.088% in speed 6. The encoding speed is slightly up by 2%, as
the larger coding blocks' prediction quality is improved.
The compression efficieny in speed 1 is improved by 0.2% for
all intra mode.
BUG=aomedia:2959
STATS_CHANGED
Change-Id: I7183e77a7badbf1c82b3e1784f555c88f46d58d6
diff --git a/av1/encoder/intra_mode_search.c b/av1/encoder/intra_mode_search.c
index da47610..90d5de9 100644
--- a/av1/encoder/intra_mode_search.c
+++ b/av1/encoder/intra_mode_search.c
@@ -1098,11 +1098,6 @@
continue;
mbmi->angle_delta[PLANE_TYPE_Y] = 0;
- if (model_intra_yrd_and_prune(cpi, x, bsize, bmode_costs[mbmi->mode],
- &best_model_rd)) {
- continue;
- }
-
is_directional_mode = av1_is_directional_mode(mbmi->mode);
if (is_directional_mode && directional_mode_skip_mask[mbmi->mode]) continue;
if (is_directional_mode && av1_use_angle_delta(bsize) &&
@@ -1113,6 +1108,11 @@
bmode_costs[mbmi->mode], best_rd, &best_model_rd,
1);
} else {
+ if (model_intra_yrd_and_prune(cpi, x, bsize, bmode_costs[mbmi->mode],
+ &best_model_rd)) {
+ continue;
+ }
+
// Builds the actual prediction. The prediction from
// model_intra_yrd_and_prune was just an estimation that did not take into
// account the effect of txfm pipeline, so we need to redo it for real