-
Ronald S. Bultje authored
Makes first 50 frames of bus @ 1500kbps encode from 3min22.7 to 3min18.2, i.e. 2.3% faster. In addition, use the sub_pixel_avg functions to calc the variance of the averaging predictor. This is slightly suboptimal because the function is subpixel-position-aware, but it will (at least for the SSE2 version) not actually use a bilinear filter for a full-pixel position, thus leading to approximately the same performance compared to if we implemented an actual average-aware full-pixel variance function. That gains another 0.3 seconds (i.e. encode time goes to 3min17.4), thus leading to a total gain of 2.7%. Change-Id: I3f059d2b04243921868cfed2568d4fa65d7b5acd
c24d9223
/* * Copyright (c) 2010 The WebM project authors. All Rights Reserved. * * Use of this source code is governed by a BSD-style license * that can be found in the LICENSE file in the root of the source * tree. An additional intellectual property rights grant can be found * in the file PATENTS. All contributing project authors may * be found in the AUTHORS file in the root of the source tree. */#ifndef VP9_ENCODER_VP9_VARIANCE_H_#define VP9_ENCODER_VP9_VARIANCE_H_#include "vpx/vpx_integer.h"// #include "./vpx_config.h"typedef unsigned int(*vp9_sad_fn_t)(const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad);typedef unsigned int(*vp9_sad_avg_fn_t)(const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad);typedef void (*vp9_sad_multi_fn_t)(const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sad_array);typedef void (*vp9_sad_multi1_fn_t)(const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sad_array);typedef void (*vp9_sad_multi_d_fn_t)(const uint8_t *src_ptr, int source_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array);typedef unsigned int (*vp9_variance_fn_t)(const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse);typedef unsigned int (*vp9_subpixvariance_fn_t)(const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int Refstride, unsigned int *sse);typedef unsigned int (*vp9_subp_avg_variance_fn_t)(const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int Refstride, unsigned int *sse, const uint8_t *second_pred);typedef void (*vp9_ssimpf_fn_t)(uint8_t *s, int sp, uint8_t *r,7172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113
int rp, unsigned long *sum_s,
unsigned long *sum_r, unsigned long *sum_sq_s,
unsigned long *sum_sq_r,
unsigned long *sum_sxr);
typedef unsigned int (*vp9_getmbss_fn_t)(const short *);
typedef unsigned int (*vp9_get16x16prederror_fn_t)(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int ref_stride);
typedef struct vp9_variance_vtable {
vp9_sad_fn_t sdf;
vp9_sad_avg_fn_t sdaf;
vp9_variance_fn_t vf;
vp9_subpixvariance_fn_t svf;
vp9_subp_avg_variance_fn_t svaf;
vp9_variance_fn_t svf_halfpix_h;
vp9_variance_fn_t svf_halfpix_v;
vp9_variance_fn_t svf_halfpix_hv;
vp9_sad_multi_fn_t sdx3f;
vp9_sad_multi1_fn_t sdx8f;
vp9_sad_multi_d_fn_t sdx4df;
} vp9_variance_fn_ptr_t;
static void comp_avg_pred(uint8_t *comp_pred, const uint8_t *pred, int width,
int height, const uint8_t *ref, int ref_stride) {
int i, j;
for (i = 0; i < height; i++) {
for (j = 0; j < width; j++) {
int tmp;
tmp = pred[j] + ref[j];
comp_pred[j] = (tmp + 1) >> 1;
}
comp_pred += width;
pred += width;
ref += ref_stride;
}
}
#endif // VP9_ENCODER_VP9_VARIANCE_H_