Refine interpolation_filter_search
This patch will restructure the match found logic in interpolation_filter_search().
In interpolation stats, rd and sse of interpolation search result is included.
Same values are restored, if current mode is matching with interpolation stats.
In this way, calculation of model_rd is avoided. It also reduces the number of calls
to build_inter_predictor() function for early exit case.
This change is bit exact. For speed = 1, 2, 3 and 4 presets, encode time reduction
is seen as 0.40%, 0.85%, 1.40% and 1.20% (averaged across multiple test cases).
Change-Id: Ie8ace3f758c9da9143cc602ce14b876a7b78f299
diff --git a/av1/encoder/block.h b/av1/encoder/block.h
index b87411f..46947ec 100644
--- a/av1/encoder/block.h
+++ b/av1/encoder/block.h
@@ -184,6 +184,10 @@
int_mv mv[2];
int8_t ref_frames[2];
COMPOUND_TYPE comp_type;
+ int64_t rd;
+ int skip_txfm_sb;
+ int64_t skip_sse_sb;
+ unsigned int pred_sse;
} INTERPOLATION_FILTER_STATS;
#define MAX_COMP_RD_STATS 64
diff --git a/av1/encoder/rdopt.c b/av1/encoder/rdopt.c
index dcd3465..ef079bf 100644
--- a/av1/encoder/rdopt.c
+++ b/av1/encoder/rdopt.c
@@ -8481,7 +8481,10 @@
}
static INLINE void save_interp_filter_search_stat(MACROBLOCK *x,
- MB_MODE_INFO *const mbmi) {
+ MB_MODE_INFO *const mbmi,
+ int64_t rd, int skip_txfm_sb,
+ int64_t skip_sse_sb,
+ unsigned int pred_sse) {
const int comp_idx = mbmi->compound_idx;
const int offset = x->interp_filter_stats_idx[comp_idx];
if (offset < MAX_INTERP_FILTER_STATS) {
@@ -8489,7 +8492,11 @@
{ mbmi->mv[0], mbmi->mv[1] },
{ mbmi->ref_frame[0],
mbmi->ref_frame[1] },
- mbmi->interinter_comp.type };
+ mbmi->interinter_comp.type,
+ rd,
+ skip_txfm_sb,
+ skip_sse_sb,
+ pred_sse };
x->interp_filter_stats[comp_idx][offset] = stat;
x->interp_filter_stats_idx[comp_idx]++;
}
@@ -8526,8 +8533,8 @@
const BUFFER_SET *const tmp_dst, const BUFFER_SET *const orig_dst,
InterpFilter (*const single_filter)[REF_FRAMES], int64_t *const rd,
int *const switchable_rate, int *const skip_txfm_sb,
- int64_t *const skip_sse_sb, const int skip_build_pred,
- HandleInterModeArgs *args, int64_t ref_best_rd) {
+ int64_t *const skip_sse_sb, int *skip_build_pred, HandleInterModeArgs *args,
+ int64_t ref_best_rd) {
const AV1_COMMON *cm = &cpi->common;
const int num_planes = av1_num_planes(cm);
MACROBLOCKD *const xd = &x->e_mbd;
@@ -8544,12 +8551,23 @@
const int ref_frame = xd->mi[0]->ref_frame[0];
(void)single_filter;
- int match_found = -1;
+ int match_found_idx = -1;
const InterpFilter assign_filter = cm->interp_filter;
if (cpi->sf.skip_repeat_interpolation_filter_search && need_search) {
- match_found = find_interp_filter_in_stats(x, mbmi);
+ match_found_idx = find_interp_filter_in_stats(x, mbmi);
}
- if (!need_search || match_found == -1) {
+ if (match_found_idx != -1) {
+ const int comp_idx = mbmi->compound_idx;
+ *rd = x->interp_filter_stats[comp_idx][match_found_idx].rd;
+ *skip_txfm_sb =
+ x->interp_filter_stats[comp_idx][match_found_idx].skip_txfm_sb;
+ *skip_sse_sb =
+ x->interp_filter_stats[comp_idx][match_found_idx].skip_sse_sb;
+ x->pred_sse[ref_frame] =
+ x->interp_filter_stats[comp_idx][match_found_idx].pred_sse;
+ return 0;
+ }
+ if (!need_search || match_found_idx == -1) {
set_default_interp_filters(mbmi, assign_filter);
}
int switchable_ctx[2];
@@ -8557,9 +8575,10 @@
switchable_ctx[1] = av1_get_pred_context_switchable_interp(xd, 1);
*switchable_rate =
get_switchable_rate(x, mbmi->interp_filters, switchable_ctx);
- if (!skip_build_pred) {
+ if (!(*skip_build_pred)) {
av1_enc_build_inter_predictor(cm, xd, mi_row, mi_col, orig_dst, bsize, 0,
av1_num_planes(cm) - 1);
+ *skip_build_pred = 1;
}
#if CONFIG_COLLECT_RD_STATS == 3
@@ -8586,7 +8605,7 @@
*skip_sse_sb = best_skip_sse_sb[1];
x->pred_sse[ref_frame] = (unsigned int)(best_skip_sse_sb[0] >> 4);
- if (assign_filter != SWITCHABLE || match_found != -1) {
+ if (assign_filter != SWITCHABLE || match_found_idx != -1) {
return 0;
}
if (!need_search) {
@@ -8756,8 +8775,9 @@
// save search results
if (cpi->sf.skip_repeat_interpolation_filter_search) {
- assert(match_found == -1);
- save_interp_filter_search_stat(x, mbmi);
+ assert(match_found_idx == -1);
+ save_interp_filter_search_stat(x, mbmi, *rd, *skip_txfm_sb, *skip_sse_sb,
+ x->pred_sse[ref_frame]);
}
return 0;
}
@@ -10240,7 +10260,7 @@
ret_val = interpolation_filter_search(
x, cpi, tile_data, bsize, mi_row, mi_col, &tmp_dst, &orig_dst,
args->single_filter, &rd, &rs, &skip_txfm_sb, &skip_sse_sb,
- skip_build_pred, args, ref_best_rd);
+ &skip_build_pred, args, ref_best_rd);
if (args->modelled_rd != NULL && !is_comp_pred) {
args->modelled_rd[this_mode][ref_mv_idx][refs[0]] = rd;
}
@@ -10271,6 +10291,10 @@
}
}
rd_stats->rate += compmode_interinter_cost;
+ if (skip_build_pred != 1) {
+ av1_enc_build_inter_predictor(cm, xd, mi_row, mi_col, &orig_dst, bsize,
+ 0, av1_num_planes(cm) - 1);
+ }
if (cpi->sf.second_loop_comp_fast_tx_search && comp_idx == 0) {
// TODO(chengchen): this speed feature introduces big loss.
