-
James Zern authored
Change-Id: I481d9bb2fa3ec72b6a83d5f04d545ad8013f295c
c8ba8c51
/* * 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. */#include "vp9/encoder/vp9_variance.h"#include "vp9/common/vp9_filter.h"#include "vp9/common/vp9_subpelvar.h"#include "vpx/vpx_integer.h"#include "vpx_ports/mem.h"#include "./vp9_rtcd.h"unsigned int vp9_get_mb_ss_c(const int16_t *src_ptr) { unsigned int i, sum = 0; for (i = 0; i < 256; i++) { sum += (src_ptr[i] * src_ptr[i]); } return sum;}unsigned int vp9_variance64x32_c(const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int recon_stride, unsigned int *sse) { unsigned int var; int avg; variance(src_ptr, source_stride, ref_ptr, recon_stride, 64, 32, &var, &avg); *sse = var; return (var - (((int64_t)avg * avg) >> 11));}unsigned int vp9_sub_pixel_variance64x32_c(const uint8_t *src_ptr, int src_pixels_per_line, int xoffset, int yoffset, const uint8_t *dst_ptr, int dst_pixels_per_line, unsigned int *sse) { uint16_t fdata3[65 * 64]; // Temp data buffer used in filtering uint8_t temp2[68 * 64]; const int16_t *hfilter, *vfilter; hfilter = BILINEAR_FILTERS_2TAP(xoffset); vfilter = BILINEAR_FILTERS_2TAP(yoffset); var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line, 1, 33, 64, hfilter); var_filter_block2d_bil_second_pass(fdata3, temp2, 64, 64, 32, 64, vfilter); return vp9_variance64x32(temp2, 64, dst_ptr, dst_pixels_per_line, sse);}unsigned int vp9_sub_pixel_avg_variance64x32_c(const uint8_t *src_ptr, int src_pixels_per_line, int xoffset, int yoffset, const uint8_t *dst_ptr, int dst_pixels_per_line, unsigned int *sse, const uint8_t *second_pred) {7172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140
uint16_t fdata3[65 * 64]; // Temp data buffer used in filtering
uint8_t temp2[68 * 64];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 64 * 64); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 33, 64, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 64, 64, 32, 64, vfilter);
comp_avg_pred(temp3, second_pred, 64, 32, temp2, 64);
return vp9_variance64x32(temp3, 64, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_variance32x64_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 32, 64, &var, &avg);
*sse = var;
return (var - (((int64_t)avg * avg) >> 11));
}
unsigned int vp9_sub_pixel_variance32x64_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[65 * 64]; // Temp data buffer used in filtering
uint8_t temp2[68 * 64];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 65, 32, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 32, 32, 64, 32, vfilter);
return vp9_variance32x64(temp2, 32, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance32x64_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[65 * 64]; // Temp data buffer used in filtering
uint8_t temp2[68 * 64];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 32 * 64); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 65, 32, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 32, 32, 64, 32, vfilter);
comp_avg_pred(temp3, second_pred, 32, 64, temp2, 32);
return vp9_variance32x64(temp3, 32, dst_ptr, dst_pixels_per_line, sse);
141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210
}
unsigned int vp9_variance32x16_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 32, 16, &var, &avg);
*sse = var;
return (var - (((int64_t)avg * avg) >> 9));
}
unsigned int vp9_sub_pixel_variance32x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[33 * 32]; // Temp data buffer used in filtering
uint8_t temp2[36 * 32];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 17, 32, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 32, 32, 16, 32, vfilter);
return vp9_variance32x16(temp2, 32, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance32x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[33 * 32]; // Temp data buffer used in filtering
uint8_t temp2[36 * 32];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 32 * 16); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 17, 32, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 32, 32, 16, 32, vfilter);
comp_avg_pred(temp3, second_pred, 32, 16, temp2, 32);
return vp9_variance32x16(temp3, 32, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_variance16x32_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 32, &var, &avg);
*sse = var;
return (var - (((int64_t)avg * avg) >> 9));
211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280
}
unsigned int vp9_sub_pixel_variance16x32_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[33 * 32]; // Temp data buffer used in filtering
uint8_t temp2[36 * 32];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 33, 16, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 16, 16, 32, 16, vfilter);
return vp9_variance16x32(temp2, 16, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance16x32_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[33 * 32]; // Temp data buffer used in filtering
uint8_t temp2[36 * 32];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 16 * 32); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 33, 16, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 16, 16, 32, 16, vfilter);
comp_avg_pred(temp3, second_pred, 16, 32, temp2, 16);
return vp9_variance16x32(temp3, 16, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_variance64x64_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 64, 64, &var, &avg);
*sse = var;
return (var - (((int64_t)avg * avg) >> 12));
}
unsigned int vp9_variance32x32_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 32, 32, &var, &avg);
*sse = var;
return (var - (((int64_t)avg * avg) >> 10));
281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350
}
unsigned int vp9_variance16x16_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 16, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 8));
}
unsigned int vp9_variance8x16_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 8, 16, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 7));
}
unsigned int vp9_variance16x8_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 8, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 7));
}
void vp9_get_sse_sum_8x8_c(const uint8_t *src_ptr, int source_stride,
const uint8_t *ref_ptr, int ref_stride,
unsigned int *sse, int *sum) {
variance(src_ptr, source_stride, ref_ptr, ref_stride, 8, 8, sse, sum);
}
unsigned int vp9_variance8x8_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 8, 8, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 6));
}
unsigned int vp9_variance8x4_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 8, 4, &var, &avg);
*sse = var;
351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420
return (var - (((unsigned int)avg * avg) >> 5));
}
unsigned int vp9_variance4x8_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 4, 8, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 5));
}
unsigned int vp9_variance4x4_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 4, 4, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 4));
}
unsigned int vp9_mse16x16_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 16, &var, &avg);
*sse = var;
return var;
}
unsigned int vp9_mse16x8_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 8, &var, &avg);
*sse = var;
return var;
}
unsigned int vp9_mse8x16_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 8, 16, &var, &avg);
*sse = var;
return var;
}
unsigned int vp9_mse8x8_c(const uint8_t *src_ptr,
421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
variance(src_ptr, source_stride, ref_ptr, recon_stride, 8, 8, &var, &avg);
*sse = var;
return var;
}
unsigned int vp9_sub_pixel_variance4x4_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
uint16_t fdata3[5 * 4]; // Temp data buffer used in filtering
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
// First filter 1d Horizontal
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 5, 4, hfilter);
// Now filter Verticaly
var_filter_block2d_bil_second_pass(fdata3, temp2, 4, 4, 4, 4, vfilter);
return vp9_variance4x4(temp2, 4, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance4x4_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 4 * 4); // compound pred buffer
uint16_t fdata3[5 * 4]; // Temp data buffer used in filtering
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
// First filter 1d Horizontal
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 5, 4, hfilter);
// Now filter Verticaly
var_filter_block2d_bil_second_pass(fdata3, temp2, 4, 4, 4, 4, vfilter);
comp_avg_pred(temp3, second_pred, 4, 4, temp2, 4);
return vp9_variance4x4(temp3, 4, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_variance8x8_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560
uint16_t fdata3[9 * 8]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 9, 8, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 8, 8, 8, 8, vfilter);
return vp9_variance8x8(temp2, 8, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance8x8_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[9 * 8]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 8 * 8); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 9, 8, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 8, 8, 8, 8, vfilter);
comp_avg_pred(temp3, second_pred, 8, 8, temp2, 8);
return vp9_variance8x8(temp3, 8, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_variance16x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[17 * 16]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 17, 16, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 16, 16, 16, 16, vfilter);
return vp9_variance16x16(temp2, 16, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance16x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[17 * 16];
uint8_t temp2[20 * 16];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 16 * 16); // compound pred buffer
const int16_t *hfilter, *vfilter;
561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 17, 16, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 16, 16, 16, 16, vfilter);
comp_avg_pred(temp3, second_pred, 16, 16, temp2, 16);
return vp9_variance16x16(temp3, 16, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_variance64x64_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[65 * 64]; // Temp data buffer used in filtering
uint8_t temp2[68 * 64];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 65, 64, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 64, 64, 64, 64, vfilter);
return vp9_variance64x64(temp2, 64, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance64x64_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[65 * 64]; // Temp data buffer used in filtering
uint8_t temp2[68 * 64];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 64 * 64); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 65, 64, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 64, 64, 64, 64, vfilter);
comp_avg_pred(temp3, second_pred, 64, 64, temp2, 64);
return vp9_variance64x64(temp3, 64, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_variance32x32_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[33 * 32]; // Temp data buffer used in filtering
uint8_t temp2[36 * 32];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 33, 32, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 32, 32, 32, 32, vfilter);
return vp9_variance32x32(temp2, 32, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance32x32_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[33 * 32]; // Temp data buffer used in filtering
uint8_t temp2[36 * 32];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 32 * 32); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 33, 32, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 32, 32, 32, 32, vfilter);
comp_avg_pred(temp3, second_pred, 32, 32, temp2, 32);
return vp9_variance32x32(temp3, 32, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_variance_halfpixvar16x16_h_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance16x16_c(src_ptr, source_stride, 8, 0,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar32x32_h_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance32x32_c(src_ptr, source_stride, 8, 0,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar64x64_h_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance64x64_c(src_ptr, source_stride, 8, 0,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar16x16_v_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance16x16_c(src_ptr, source_stride, 0, 8,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar32x32_v_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770
unsigned int *sse) {
return vp9_sub_pixel_variance32x32_c(src_ptr, source_stride, 0, 8,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar64x64_v_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance64x64_c(src_ptr, source_stride, 0, 8,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar16x16_hv_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance16x16_c(src_ptr, source_stride, 8, 8,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar32x32_hv_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance32x32_c(src_ptr, source_stride, 8, 8,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_variance_halfpixvar64x64_hv_c(const uint8_t *src_ptr,
int source_stride,
const uint8_t *ref_ptr,
int recon_stride,
unsigned int *sse) {
return vp9_sub_pixel_variance64x64_c(src_ptr, source_stride, 8, 8,
ref_ptr, recon_stride, sse);
}
unsigned int vp9_sub_pixel_mse16x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
vp9_sub_pixel_variance16x16_c(src_ptr, src_pixels_per_line,
xoffset, yoffset, dst_ptr,
dst_pixels_per_line, sse);
return *sse;
}
unsigned int vp9_sub_pixel_mse32x32_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
vp9_sub_pixel_variance32x32_c(src_ptr, src_pixels_per_line,
xoffset, yoffset, dst_ptr,
dst_pixels_per_line, sse);
return *sse;
}
unsigned int vp9_sub_pixel_mse64x64_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
vp9_sub_pixel_variance64x64_c(src_ptr, src_pixels_per_line,
xoffset, yoffset, dst_ptr,
dst_pixels_per_line, sse);
return *sse;
}
unsigned int vp9_sub_pixel_variance16x8_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[16 * 9]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 9, 16, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 16, 16, 8, 16, vfilter);
return vp9_variance16x8(temp2, 16, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance16x8_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[16 * 9]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 16 * 8); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 9, 16, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 16, 16, 8, 16, vfilter);
comp_avg_pred(temp3, second_pred, 16, 8, temp2, 16);
return vp9_variance16x8(temp3, 16, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_variance8x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[9 * 16]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 17, 8, hfilter);
841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910
var_filter_block2d_bil_second_pass(fdata3, temp2, 8, 8, 16, 8, vfilter);
return vp9_variance8x16(temp2, 8, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance8x16_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[9 * 16]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 8 * 16); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 17, 8, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 8, 8, 16, 8, vfilter);
comp_avg_pred(temp3, second_pred, 8, 16, temp2, 8);
return vp9_variance8x16(temp3, 8, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_variance8x4_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[8 * 5]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 5, 8, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 8, 8, 4, 8, vfilter);
return vp9_variance8x4(temp2, 8, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance8x4_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[8 * 5]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 8 * 4); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 5, 8, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 8, 8, 4, 8, vfilter);
comp_avg_pred(temp3, second_pred, 8, 4, temp2, 8);
return vp9_variance8x4(temp3, 8, dst_ptr, dst_pixels_per_line, sse);
911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958
}
unsigned int vp9_sub_pixel_variance4x8_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse) {
uint16_t fdata3[5 * 8]; // Temp data buffer used in filtering
// FIXME(jingning,rbultje): this temp2 buffer probably doesn't need to be
// of this big? same issue appears in all other block size settings.
uint8_t temp2[20 * 16];
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 9, 4, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 4, 4, 8, 4, vfilter);
return vp9_variance4x8(temp2, 4, dst_ptr, dst_pixels_per_line, sse);
}
unsigned int vp9_sub_pixel_avg_variance4x8_c(const uint8_t *src_ptr,
int src_pixels_per_line,
int xoffset,
int yoffset,
const uint8_t *dst_ptr,
int dst_pixels_per_line,
unsigned int *sse,
const uint8_t *second_pred) {
uint16_t fdata3[5 * 8]; // Temp data buffer used in filtering
uint8_t temp2[20 * 16];
DECLARE_ALIGNED_ARRAY(16, uint8_t, temp3, 4 * 8); // compound pred buffer
const int16_t *hfilter, *vfilter;
hfilter = BILINEAR_FILTERS_2TAP(xoffset);
vfilter = BILINEAR_FILTERS_2TAP(yoffset);
var_filter_block2d_bil_first_pass(src_ptr, fdata3, src_pixels_per_line,
1, 9, 4, hfilter);
var_filter_block2d_bil_second_pass(fdata3, temp2, 4, 4, 8, 4, vfilter);
comp_avg_pred(temp3, second_pred, 4, 8, temp2, 4);
return vp9_variance4x8(temp3, 4, dst_ptr, dst_pixels_per_line, sse);
}