No need to calculate baseline VMAF
Less than 0.01% RD-rate change.
Change-Id: I6955f21f63383ce31304f3460b0b9c76d8cb31be
diff --git a/av1/encoder/encoder.h b/av1/encoder/encoder.h
index 1d84500..86fd720 100644
--- a/av1/encoder/encoder.h
+++ b/av1/encoder/encoder.h
@@ -1138,7 +1138,6 @@
#if CONFIG_TUNE_VMAF
double *vmaf_rdmult_scaling_factors;
double last_frame_ysse;
- double last_frame_bvmaf;
double last_frame_vmaf;
#endif
diff --git a/av1/encoder/tune_vmaf.c b/av1/encoder/tune_vmaf.c
index c8ebe59..79220a6 100644
--- a/av1/encoder/tune_vmaf.c
+++ b/av1/encoder/tune_vmaf.c
@@ -17,6 +17,8 @@
#include "av1/encoder/extend.h"
#include "av1/encoder/rdopt.h"
+static const double kBaselineVmaf = 97.42773;
+
// TODO(sdeng): Add the SIMD implementation.
static AOM_INLINE void highbd_unsharp_rect(const uint16_t *source,
int source_stride,
@@ -173,7 +175,6 @@
cpi->oxcf.border_in_pixels, cm->byte_alignment);
const double baseline_variance = frame_average_variance(cpi, source);
- const double baseline_vmaf = 97.42773;
int loop_count = 0;
double approx_vmaf = 0.0;
double best_vmaf, new_vmaf, unsharp_amount = unsharp_amount_start;
@@ -185,7 +186,7 @@
&new_vmaf);
const double sharpened_var = frame_average_variance(cpi, &sharpened);
approx_vmaf =
- baseline_variance / sharpened_var * (new_vmaf - baseline_vmaf);
+ baseline_variance / sharpened_var * (new_vmaf - kBaselineVmaf);
loop_count++;
} while (approx_vmaf > best_vmaf && loop_count < max_loop_count);
@@ -419,11 +420,7 @@
}
}
}
- if (row < 0 && col < 0) {
- frames->row = 0;
- frames->col = 0;
- return 0;
- } else if (row < num_rows && col < num_cols) {
+ if (row < num_rows && col < num_cols) {
// Set current block
const int row_offset = row * block_h;
const int col_offset = col * block_w;
@@ -488,8 +485,8 @@
cpi->oxcf.border_in_pixels, cm->byte_alignment);
gaussian_blur(cpi, cpi->source, &blurred);
- double *scores = aom_malloc(sizeof(*scores) * (num_rows * num_cols + 1));
- memset(scores, 0, sizeof(*scores) * (num_rows * num_cols + 1));
+ double *scores = aom_malloc(sizeof(*scores) * (num_rows * num_cols));
+ memset(scores, 0, sizeof(*scores) * (num_rows * num_cols));
FrameData frame_data;
frame_data.source = cpi->source;
frame_data.blurred = &blurred;
@@ -497,12 +494,11 @@
frame_data.block_h = block_h;
frame_data.num_rows = num_rows;
frame_data.num_cols = num_cols;
- frame_data.row = -1;
- frame_data.col = -1;
+ frame_data.row = 0;
+ frame_data.col = 0;
frame_data.bit_depth = bit_depth;
aom_calc_vmaf_multi_frame(&frame_data, cpi->oxcf.vmaf_model_path,
update_frame, y_width, y_height, bit_depth, scores);
- const double baseline_vmaf = scores[0];
// Loop through each 'block_size' block.
for (int row = 0; row < num_rows; ++row) {
@@ -518,8 +514,8 @@
uint8_t *const blurred_buf =
blurred.y_buffer + row_offset_y * blurred.y_stride + col_offset_y;
- const double vmaf = scores[index + 1];
- const double dvmaf = baseline_vmaf - vmaf;
+ const double vmaf = scores[index];
+ const double dvmaf = kBaselineVmaf - vmaf;
unsigned int sse;
cpi->fn_ptr[block_size].vf(orig_buf, y_stride, blurred_buf,
blurred.y_stride, &sse);
@@ -689,7 +685,7 @@
const double approx_sse =
cpi->last_frame_ysse /
(double)((1 << (bit_depth - 8)) * (1 << (bit_depth - 8)));
- const double approx_dvmaf = cpi->last_frame_bvmaf - cpi->last_frame_vmaf;
+ const double approx_dvmaf = kBaselineVmaf - cpi->last_frame_vmaf;
const double sse_threshold =
0.01 * cpi->source->y_width * cpi->source->y_height;
const double vmaf_threshold = 0.01;
@@ -739,8 +735,6 @@
void av1_update_vmaf_curve(AV1_COMP *cpi, YV12_BUFFER_CONFIG *source,
YV12_BUFFER_CONFIG *recon) {
const int bit_depth = cpi->td.mb.e_mbd.bd;
- aom_calc_vmaf(cpi->oxcf.vmaf_model_path, source, source, bit_depth,
- &cpi->last_frame_bvmaf);
aom_calc_vmaf(cpi->oxcf.vmaf_model_path, source, recon, bit_depth,
&cpi->last_frame_vmaf);
if (bit_depth > 8) {
