mirror of
https://github.com/ultravideo/uvg266.git
synced 2024-11-24 02:24:07 +00:00
Merge branch 'intra-sad-strategies'
This commit is contained in:
Ari Koivula
commit
153b1ee41f
|
|
@ -14,9 +14,9 @@
|
|||
<RuntimeLibrary>MultiThreadedDebugDLL</RuntimeLibrary>
|
||||
<PreprocessorDefinitions>WIN32_LEAN_AND_MEAN;WIN32;WIN64;_CONSOLE;_CRT_SECURE_NO_WARNINGS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
|
||||
<AdditionalIncludeDirectories>$(SolutionDir)..\..\pthreads.2\include;$(SolutionDir)..\src;$(SolutionDir)..\src\extras;$(SolutionDir)..\;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
|
||||
<DisableSpecificWarnings>4244;4204;4206;4028</DisableSpecificWarnings>
|
||||
<DisableSpecificWarnings>4244;4204;4206;4028;4152</DisableSpecificWarnings>
|
||||
<OpenMPSupport>false</OpenMPSupport>
|
||||
<TreatSpecificWarningsAsErrors>4013;4029</TreatSpecificWarningsAsErrors>
|
||||
<TreatSpecificWarningsAsErrors>4013;4029;4047</TreatSpecificWarningsAsErrors>
|
||||
</ClCompile>
|
||||
<Link>
|
||||
<AdditionalDependencies>%(AdditionalDependencies)</AdditionalDependencies>
|
||||
|
|
|
|||
|
|
@ -83,10 +83,6 @@
|
|||
<AdditionalLibraryDirectories>$(SolutionDir)..\..\pthreads.2\lib\x64</AdditionalLibraryDirectories>
|
||||
<AdditionalDependencies>pthreadVC2.lib</AdditionalDependencies>
|
||||
</Lib>
|
||||
<ClCompile>
|
||||
<DisableSpecificWarnings>
|
||||
</DisableSpecificWarnings>
|
||||
</ClCompile>
|
||||
</ItemDefinitionGroup>
|
||||
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
|
||||
<YASM>
|
||||
|
|
|
|||
278
src/image.c
278
src/image.c
|
|
@ -388,284 +388,6 @@ unsigned image_calc_sad(const image *pic, const image *ref, int pic_x, int pic_y
|
|||
}
|
||||
|
||||
|
||||
/**
|
||||
* \brief Calculate SATD between two 4x4 blocks inside bigger arrays.
|
||||
* From HM 13.0
|
||||
*/
|
||||
static unsigned satd_16bit_4x4(const pixel *piOrg, const pixel *piCur)
|
||||
{
|
||||
int32_t k, satd = 0, diff[16], m[16], d[16];
|
||||
for( k = 0; k < 16; ++k ) {
|
||||
diff[k] = piOrg[k] - piCur[k];
|
||||
}
|
||||
|
||||
/*===== hadamard transform =====*/
|
||||
m[ 0] = diff[ 0] + diff[12];
|
||||
m[ 1] = diff[ 1] + diff[13];
|
||||
m[ 2] = diff[ 2] + diff[14];
|
||||
m[ 3] = diff[ 3] + diff[15];
|
||||
m[ 4] = diff[ 4] + diff[ 8];
|
||||
m[ 5] = diff[ 5] + diff[ 9];
|
||||
m[ 6] = diff[ 6] + diff[10];
|
||||
m[ 7] = diff[ 7] + diff[11];
|
||||
m[ 8] = diff[ 4] - diff[ 8];
|
||||
m[ 9] = diff[ 5] - diff[ 9];
|
||||
m[10] = diff[ 6] - diff[10];
|
||||
m[11] = diff[ 7] - diff[11];
|
||||
m[12] = diff[ 0] - diff[12];
|
||||
m[13] = diff[ 1] - diff[13];
|
||||
m[14] = diff[ 2] - diff[14];
|
||||
m[15] = diff[ 3] - diff[15];
|
||||
|
||||
d[ 0] = m[ 0] + m[ 4];
|
||||
d[ 1] = m[ 1] + m[ 5];
|
||||
d[ 2] = m[ 2] + m[ 6];
|
||||
d[ 3] = m[ 3] + m[ 7];
|
||||
d[ 4] = m[ 8] + m[12];
|
||||
d[ 5] = m[ 9] + m[13];
|
||||
d[ 6] = m[10] + m[14];
|
||||
d[ 7] = m[11] + m[15];
|
||||
d[ 8] = m[ 0] - m[ 4];
|
||||
d[ 9] = m[ 1] - m[ 5];
|
||||
d[10] = m[ 2] - m[ 6];
|
||||
d[11] = m[ 3] - m[ 7];
|
||||
d[12] = m[12] - m[ 8];
|
||||
d[13] = m[13] - m[ 9];
|
||||
d[14] = m[14] - m[10];
|
||||
d[15] = m[15] - m[11];
|
||||
|
||||
m[ 0] = d[ 0] + d[ 3];
|
||||
m[ 1] = d[ 1] + d[ 2];
|
||||
m[ 2] = d[ 1] - d[ 2];
|
||||
m[ 3] = d[ 0] - d[ 3];
|
||||
m[ 4] = d[ 4] + d[ 7];
|
||||
m[ 5] = d[ 5] + d[ 6];
|
||||
m[ 6] = d[ 5] - d[ 6];
|
||||
m[ 7] = d[ 4] - d[ 7];
|
||||
m[ 8] = d[ 8] + d[11];
|
||||
m[ 9] = d[ 9] + d[10];
|
||||
m[10] = d[ 9] - d[10];
|
||||
m[11] = d[ 8] - d[11];
|
||||
m[12] = d[12] + d[15];
|
||||
m[13] = d[13] + d[14];
|
||||
m[14] = d[13] - d[14];
|
||||
m[15] = d[12] - d[15];
|
||||
|
||||
d[ 0] = m[ 0] + m[ 1];
|
||||
d[ 1] = m[ 0] - m[ 1];
|
||||
d[ 2] = m[ 2] + m[ 3];
|
||||
d[ 3] = m[ 3] - m[ 2];
|
||||
d[ 4] = m[ 4] + m[ 5];
|
||||
d[ 5] = m[ 4] - m[ 5];
|
||||
d[ 6] = m[ 6] + m[ 7];
|
||||
d[ 7] = m[ 7] - m[ 6];
|
||||
d[ 8] = m[ 8] + m[ 9];
|
||||
d[ 9] = m[ 8] - m[ 9];
|
||||
d[10] = m[10] + m[11];
|
||||
d[11] = m[11] - m[10];
|
||||
d[12] = m[12] + m[13];
|
||||
d[13] = m[12] - m[13];
|
||||
d[14] = m[14] + m[15];
|
||||
d[15] = m[15] - m[14];
|
||||
|
||||
for (k=0; k<16; ++k) {
|
||||
satd += abs(d[k]);
|
||||
}
|
||||
satd = ((satd+1)>>1);
|
||||
|
||||
return satd;
|
||||
}
|
||||
|
||||
/**
|
||||
* \brief Calculate SATD between two 8x8 blocks inside bigger arrays.
|
||||
*/
|
||||
unsigned satd_16bit_8x8_general(const pixel * piOrg, const int32_t iStrideOrg,
|
||||
const pixel * piCur, const int32_t iStrideCur)
|
||||
{
|
||||
int32_t k, i, j, jj, sad=0;
|
||||
int32_t diff[64], m1[8][8], m2[8][8], m3[8][8];
|
||||
|
||||
for (k = 0; k < 64; k += 8) {
|
||||
diff[k+0] = piOrg[0] - piCur[0];
|
||||
diff[k+1] = piOrg[1] - piCur[1];
|
||||
diff[k+2] = piOrg[2] - piCur[2];
|
||||
diff[k+3] = piOrg[3] - piCur[3];
|
||||
diff[k+4] = piOrg[4] - piCur[4];
|
||||
diff[k+5] = piOrg[5] - piCur[5];
|
||||
diff[k+6] = piOrg[6] - piCur[6];
|
||||
diff[k+7] = piOrg[7] - piCur[7];
|
||||
|
||||
piCur += iStrideCur;
|
||||
piOrg += iStrideOrg;
|
||||
}
|
||||
|
||||
// horizontal
|
||||
for (j = 0; j < 8; ++j) {
|
||||
jj = j << 3;
|
||||
m2[j][0] = diff[jj ] + diff[jj+4];
|
||||
m2[j][1] = diff[jj+1] + diff[jj+5];
|
||||
m2[j][2] = diff[jj+2] + diff[jj+6];
|
||||
m2[j][3] = diff[jj+3] + diff[jj+7];
|
||||
m2[j][4] = diff[jj ] - diff[jj+4];
|
||||
m2[j][5] = diff[jj+1] - diff[jj+5];
|
||||
m2[j][6] = diff[jj+2] - diff[jj+6];
|
||||
m2[j][7] = diff[jj+3] - diff[jj+7];
|
||||
|
||||
m1[j][0] = m2[j][0] + m2[j][2];
|
||||
m1[j][1] = m2[j][1] + m2[j][3];
|
||||
m1[j][2] = m2[j][0] - m2[j][2];
|
||||
m1[j][3] = m2[j][1] - m2[j][3];
|
||||
m1[j][4] = m2[j][4] + m2[j][6];
|
||||
m1[j][5] = m2[j][5] + m2[j][7];
|
||||
m1[j][6] = m2[j][4] - m2[j][6];
|
||||
m1[j][7] = m2[j][5] - m2[j][7];
|
||||
|
||||
m2[j][0] = m1[j][0] + m1[j][1];
|
||||
m2[j][1] = m1[j][0] - m1[j][1];
|
||||
m2[j][2] = m1[j][2] + m1[j][3];
|
||||
m2[j][3] = m1[j][2] - m1[j][3];
|
||||
m2[j][4] = m1[j][4] + m1[j][5];
|
||||
m2[j][5] = m1[j][4] - m1[j][5];
|
||||
m2[j][6] = m1[j][6] + m1[j][7];
|
||||
m2[j][7] = m1[j][6] - m1[j][7];
|
||||
}
|
||||
|
||||
// vertical
|
||||
for (i = 0; i < 8; ++i) {
|
||||
m3[0][i] = m2[0][i] + m2[4][i];
|
||||
m3[1][i] = m2[1][i] + m2[5][i];
|
||||
m3[2][i] = m2[2][i] + m2[6][i];
|
||||
m3[3][i] = m2[3][i] + m2[7][i];
|
||||
m3[4][i] = m2[0][i] - m2[4][i];
|
||||
m3[5][i] = m2[1][i] - m2[5][i];
|
||||
m3[6][i] = m2[2][i] - m2[6][i];
|
||||
m3[7][i] = m2[3][i] - m2[7][i];
|
||||
|
||||
m1[0][i] = m3[0][i] + m3[2][i];
|
||||
m1[1][i] = m3[1][i] + m3[3][i];
|
||||
m1[2][i] = m3[0][i] - m3[2][i];
|
||||
m1[3][i] = m3[1][i] - m3[3][i];
|
||||
m1[4][i] = m3[4][i] + m3[6][i];
|
||||
m1[5][i] = m3[5][i] + m3[7][i];
|
||||
m1[6][i] = m3[4][i] - m3[6][i];
|
||||
m1[7][i] = m3[5][i] - m3[7][i];
|
||||
|
||||
m2[0][i] = m1[0][i] + m1[1][i];
|
||||
m2[1][i] = m1[0][i] - m1[1][i];
|
||||
m2[2][i] = m1[2][i] + m1[3][i];
|
||||
m2[3][i] = m1[2][i] - m1[3][i];
|
||||
m2[4][i] = m1[4][i] + m1[5][i];
|
||||
m2[5][i] = m1[4][i] - m1[5][i];
|
||||
m2[6][i] = m1[6][i] + m1[7][i];
|
||||
m2[7][i] = m1[6][i] - m1[7][i];
|
||||
}
|
||||
|
||||
for (i = 0; i < 64; ++i) {
|
||||
sad += abs(((int*)m2)[i]);
|
||||
}
|
||||
|
||||
sad = (sad + 2) >> 2;
|
||||
|
||||
return sad;
|
||||
}
|
||||
|
||||
// Function macro for defining hadamard calculating functions
|
||||
// for fixed size blocks. They calculate hadamard for integer
|
||||
// multiples of 8x8 with the 8x8 hadamard function.
|
||||
#define SATD_NXN(n, pixel_type, suffix) \
|
||||
static unsigned satd_ ## suffix ## _ ## n ## x ## n( \
|
||||
const pixel_type * const block1, const pixel_type * const block2) \
|
||||
{ \
|
||||
unsigned x, y; \
|
||||
unsigned sum = 0; \
|
||||
for (y = 0; y < (n); y += 8) { \
|
||||
unsigned row = y * (n); \
|
||||
for (x = 0; x < (n); x += 8) { \
|
||||
sum += satd_16bit_8x8_general(&block1[row + x], (n), &block2[row + x], (n)); \
|
||||
} \
|
||||
} \
|
||||
return sum; \
|
||||
}
|
||||
|
||||
// These macros define sadt_16bit_NxN for N = 8, 16, 32, 64
|
||||
SATD_NXN(8, pixel, 16bit)
|
||||
SATD_NXN(16, pixel, 16bit)
|
||||
SATD_NXN(32, pixel, 16bit)
|
||||
SATD_NXN(64, pixel, 16bit)
|
||||
|
||||
// Function macro for defining SAD calculating functions
|
||||
// for fixed size blocks.
|
||||
#define SAD_NXN(n, pixel_type, suffix) \
|
||||
static unsigned sad_ ## suffix ## _ ## n ## x ## n( \
|
||||
const pixel_type * const block1, const pixel_type * const block2) \
|
||||
{ \
|
||||
unsigned i; \
|
||||
unsigned sum = 0; \
|
||||
for (i = 0; i < (n)*(n); ++i) { \
|
||||
sum += abs(block1[i] - block2[i]); \
|
||||
} \
|
||||
return sum; \
|
||||
}
|
||||
|
||||
// These macros define sad_16bit_nxn functions for n = 4, 8, 16, 32, 64
|
||||
// with function signatures of cost_16bit_nxn_func.
|
||||
// They are used through get_sad_16bit_nxn_func.
|
||||
SAD_NXN(4, pixel, 16bit)
|
||||
SAD_NXN(8, pixel, 16bit)
|
||||
SAD_NXN(16, pixel, 16bit)
|
||||
SAD_NXN(32, pixel, 16bit)
|
||||
SAD_NXN(64, pixel, 16bit)
|
||||
|
||||
/**
|
||||
* \brief Get a function that calculates SATD for NxN block.
|
||||
*
|
||||
* \param n Width of the region for which SATD is calculated.
|
||||
*
|
||||
* \returns Pointer to cost_16bit_nxn_func.
|
||||
*/
|
||||
cost_16bit_nxn_func get_satd_16bit_nxn_func(unsigned n)
|
||||
{
|
||||
switch (n) {
|
||||
case 4:
|
||||
return &satd_16bit_4x4;
|
||||
case 8:
|
||||
return &satd_16bit_8x8;
|
||||
case 16:
|
||||
return &satd_16bit_16x16;
|
||||
case 32:
|
||||
return &satd_16bit_32x32;
|
||||
case 64:
|
||||
return &satd_16bit_64x64;
|
||||
default:
|
||||
return NULL;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* \brief Get a function that calculates SAD for NxN block.
|
||||
*
|
||||
* \param n Width of the region for which SAD is calculated.
|
||||
*
|
||||
* \returns Pointer to cost_16bit_nxn_func.
|
||||
*/
|
||||
cost_16bit_nxn_func get_sad_16bit_nxn_func(unsigned n)
|
||||
{
|
||||
switch (n) {
|
||||
case 4:
|
||||
return &sad_16bit_4x4;
|
||||
case 8:
|
||||
return &sad_16bit_8x8;
|
||||
case 16:
|
||||
return &sad_16bit_16x16;
|
||||
case 32:
|
||||
return &sad_16bit_32x32;
|
||||
case 64:
|
||||
return &sad_16bit_64x64;
|
||||
default:
|
||||
return NULL;
|
||||
}
|
||||
}
|
||||
|
||||
unsigned pixels_calc_ssd(const pixel *const ref, const pixel *const rec,
|
||||
const int ref_stride, const int rec_stride,
|
||||
const int width)
|
||||
|
|
|
|||
|
|
@ -75,15 +75,6 @@ unsigned image_calc_sad(const image *pic, const image *ref, int pic_x, int pic_y
|
|||
int block_width, int block_height);
|
||||
|
||||
|
||||
typedef unsigned (*cost_16bit_nxn_func)(const pixel *block1, const pixel *block2);
|
||||
|
||||
|
||||
cost_16bit_nxn_func get_satd_16bit_nxn_func(unsigned n);
|
||||
cost_16bit_nxn_func get_sad_16bit_nxn_func(unsigned n);
|
||||
|
||||
unsigned pixels_satd_16bit_nxn(pixel *block1, pixel *block2, unsigned n);
|
||||
unsigned pixels_sad_16bit_nxn(pixel *block1, pixel *block2, unsigned n);
|
||||
|
||||
unsigned pixels_calc_ssd(const pixel *const ref, const pixel *const rec,
|
||||
const int ref_stride, const int rec_stride,
|
||||
const int width);
|
||||
|
|
|
|||
|
|
@ -31,6 +31,7 @@
|
|||
#include "config.h"
|
||||
#include "bitstream.h"
|
||||
#include "image.h"
|
||||
#include "strategies/strategies-picture.h"
|
||||
#include "intra.h"
|
||||
#include "inter.h"
|
||||
#include "filter.h"
|
||||
|
|
@ -825,7 +826,7 @@ static void search_intra_rough(encoder_state * const encoder_state,
|
|||
int8_t modes[35], uint32_t costs[35])
|
||||
{
|
||||
int16_t mode;
|
||||
cost_16bit_nxn_func cost_func = get_sad_16bit_nxn_func(width);
|
||||
cost_pixel_nxn_func * cost_func = pixels_get_sad_func(width);
|
||||
|
||||
// Temporary block arrays
|
||||
pixel pred[LCU_WIDTH * LCU_WIDTH + 1];
|
||||
|
|
|
|||
|
|
@ -26,6 +26,7 @@
|
|||
#include "strategyselector.h"
|
||||
|
||||
|
||||
|
||||
/**
|
||||
* \brief Calculate Sum of Absolute Differences (SAD)
|
||||
*
|
||||
|
|
@ -41,7 +42,7 @@
|
|||
* \returns Sum of Absolute Differences
|
||||
*/
|
||||
static unsigned reg_sad_generic(const pixel * const data1, const pixel * const data2,
|
||||
const int width, const int height, const unsigned stride1, const unsigned stride2)
|
||||
const int width, const int height, const unsigned stride1, const unsigned stride2)
|
||||
{
|
||||
int y, x;
|
||||
unsigned sad = 0;
|
||||
|
|
@ -55,6 +56,267 @@ static unsigned reg_sad_generic(const pixel * const data1, const pixel * const d
|
|||
return sad;
|
||||
}
|
||||
|
||||
static int strategy_register_picture_generic(void* opaque) {
|
||||
return strategyselector_register(opaque, "reg_sad", "generic", 0, ®_sad_generic);
|
||||
|
||||
/**
|
||||
* \brief Calculate SATD between two 4x4 blocks inside bigger arrays.
|
||||
* From HM 13.0
|
||||
*/
|
||||
static unsigned satd_8bit_4x4_generic(const pixel *piOrg, const pixel *piCur)
|
||||
{
|
||||
int32_t k, satd = 0, diff[16], m[16], d[16];
|
||||
for (k = 0; k < 16; ++k) {
|
||||
diff[k] = piOrg[k] - piCur[k];
|
||||
}
|
||||
|
||||
/*===== hadamard transform =====*/
|
||||
m[0] = diff[0] + diff[12];
|
||||
m[1] = diff[1] + diff[13];
|
||||
m[2] = diff[2] + diff[14];
|
||||
m[3] = diff[3] + diff[15];
|
||||
m[4] = diff[4] + diff[8];
|
||||
m[5] = diff[5] + diff[9];
|
||||
m[6] = diff[6] + diff[10];
|
||||
m[7] = diff[7] + diff[11];
|
||||
m[8] = diff[4] - diff[8];
|
||||
m[9] = diff[5] - diff[9];
|
||||
m[10] = diff[6] - diff[10];
|
||||
m[11] = diff[7] - diff[11];
|
||||
m[12] = diff[0] - diff[12];
|
||||
m[13] = diff[1] - diff[13];
|
||||
m[14] = diff[2] - diff[14];
|
||||
m[15] = diff[3] - diff[15];
|
||||
|
||||
d[0] = m[0] + m[4];
|
||||
d[1] = m[1] + m[5];
|
||||
d[2] = m[2] + m[6];
|
||||
d[3] = m[3] + m[7];
|
||||
d[4] = m[8] + m[12];
|
||||
d[5] = m[9] + m[13];
|
||||
d[6] = m[10] + m[14];
|
||||
d[7] = m[11] + m[15];
|
||||
d[8] = m[0] - m[4];
|
||||
d[9] = m[1] - m[5];
|
||||
d[10] = m[2] - m[6];
|
||||
d[11] = m[3] - m[7];
|
||||
d[12] = m[12] - m[8];
|
||||
d[13] = m[13] - m[9];
|
||||
d[14] = m[14] - m[10];
|
||||
d[15] = m[15] - m[11];
|
||||
|
||||
m[0] = d[0] + d[3];
|
||||
m[1] = d[1] + d[2];
|
||||
m[2] = d[1] - d[2];
|
||||
m[3] = d[0] - d[3];
|
||||
m[4] = d[4] + d[7];
|
||||
m[5] = d[5] + d[6];
|
||||
m[6] = d[5] - d[6];
|
||||
m[7] = d[4] - d[7];
|
||||
m[8] = d[8] + d[11];
|
||||
m[9] = d[9] + d[10];
|
||||
m[10] = d[9] - d[10];
|
||||
m[11] = d[8] - d[11];
|
||||
m[12] = d[12] + d[15];
|
||||
m[13] = d[13] + d[14];
|
||||
m[14] = d[13] - d[14];
|
||||
m[15] = d[12] - d[15];
|
||||
|
||||
d[0] = m[0] + m[1];
|
||||
d[1] = m[0] - m[1];
|
||||
d[2] = m[2] + m[3];
|
||||
d[3] = m[3] - m[2];
|
||||
d[4] = m[4] + m[5];
|
||||
d[5] = m[4] - m[5];
|
||||
d[6] = m[6] + m[7];
|
||||
d[7] = m[7] - m[6];
|
||||
d[8] = m[8] + m[9];
|
||||
d[9] = m[8] - m[9];
|
||||
d[10] = m[10] + m[11];
|
||||
d[11] = m[11] - m[10];
|
||||
d[12] = m[12] + m[13];
|
||||
d[13] = m[12] - m[13];
|
||||
d[14] = m[14] + m[15];
|
||||
d[15] = m[15] - m[14];
|
||||
|
||||
for (k = 0; k<16; ++k) {
|
||||
satd += abs(d[k]);
|
||||
}
|
||||
satd = ((satd + 1) >> 1);
|
||||
|
||||
return satd;
|
||||
}
|
||||
|
||||
/**
|
||||
* \brief Calculate SATD between two 8x8 blocks inside bigger arrays.
|
||||
*/
|
||||
static unsigned satd_16bit_8x8_general(const pixel * piOrg, const int32_t iStrideOrg,
|
||||
const pixel * piCur, const int32_t iStrideCur)
|
||||
{
|
||||
int32_t k, i, j, jj, sad = 0;
|
||||
int32_t diff[64], m1[8][8], m2[8][8], m3[8][8];
|
||||
|
||||
for (k = 0; k < 64; k += 8) {
|
||||
diff[k + 0] = piOrg[0] - piCur[0];
|
||||
diff[k + 1] = piOrg[1] - piCur[1];
|
||||
diff[k + 2] = piOrg[2] - piCur[2];
|
||||
diff[k + 3] = piOrg[3] - piCur[3];
|
||||
diff[k + 4] = piOrg[4] - piCur[4];
|
||||
diff[k + 5] = piOrg[5] - piCur[5];
|
||||
diff[k + 6] = piOrg[6] - piCur[6];
|
||||
diff[k + 7] = piOrg[7] - piCur[7];
|
||||
|
||||
piCur += iStrideCur;
|
||||
piOrg += iStrideOrg;
|
||||
}
|
||||
|
||||
// horizontal
|
||||
for (j = 0; j < 8; ++j) {
|
||||
jj = j << 3;
|
||||
m2[j][0] = diff[jj] + diff[jj + 4];
|
||||
m2[j][1] = diff[jj + 1] + diff[jj + 5];
|
||||
m2[j][2] = diff[jj + 2] + diff[jj + 6];
|
||||
m2[j][3] = diff[jj + 3] + diff[jj + 7];
|
||||
m2[j][4] = diff[jj] - diff[jj + 4];
|
||||
m2[j][5] = diff[jj + 1] - diff[jj + 5];
|
||||
m2[j][6] = diff[jj + 2] - diff[jj + 6];
|
||||
m2[j][7] = diff[jj + 3] - diff[jj + 7];
|
||||
|
||||
m1[j][0] = m2[j][0] + m2[j][2];
|
||||
m1[j][1] = m2[j][1] + m2[j][3];
|
||||
m1[j][2] = m2[j][0] - m2[j][2];
|
||||
m1[j][3] = m2[j][1] - m2[j][3];
|
||||
m1[j][4] = m2[j][4] + m2[j][6];
|
||||
m1[j][5] = m2[j][5] + m2[j][7];
|
||||
m1[j][6] = m2[j][4] - m2[j][6];
|
||||
m1[j][7] = m2[j][5] - m2[j][7];
|
||||
|
||||
m2[j][0] = m1[j][0] + m1[j][1];
|
||||
m2[j][1] = m1[j][0] - m1[j][1];
|
||||
m2[j][2] = m1[j][2] + m1[j][3];
|
||||
m2[j][3] = m1[j][2] - m1[j][3];
|
||||
m2[j][4] = m1[j][4] + m1[j][5];
|
||||
m2[j][5] = m1[j][4] - m1[j][5];
|
||||
m2[j][6] = m1[j][6] + m1[j][7];
|
||||
m2[j][7] = m1[j][6] - m1[j][7];
|
||||
}
|
||||
|
||||
// vertical
|
||||
for (i = 0; i < 8; ++i) {
|
||||
m3[0][i] = m2[0][i] + m2[4][i];
|
||||
m3[1][i] = m2[1][i] + m2[5][i];
|
||||
m3[2][i] = m2[2][i] + m2[6][i];
|
||||
m3[3][i] = m2[3][i] + m2[7][i];
|
||||
m3[4][i] = m2[0][i] - m2[4][i];
|
||||
m3[5][i] = m2[1][i] - m2[5][i];
|
||||
m3[6][i] = m2[2][i] - m2[6][i];
|
||||
m3[7][i] = m2[3][i] - m2[7][i];
|
||||
|
||||
m1[0][i] = m3[0][i] + m3[2][i];
|
||||
m1[1][i] = m3[1][i] + m3[3][i];
|
||||
m1[2][i] = m3[0][i] - m3[2][i];
|
||||
m1[3][i] = m3[1][i] - m3[3][i];
|
||||
m1[4][i] = m3[4][i] + m3[6][i];
|
||||
m1[5][i] = m3[5][i] + m3[7][i];
|
||||
m1[6][i] = m3[4][i] - m3[6][i];
|
||||
m1[7][i] = m3[5][i] - m3[7][i];
|
||||
|
||||
m2[0][i] = m1[0][i] + m1[1][i];
|
||||
m2[1][i] = m1[0][i] - m1[1][i];
|
||||
m2[2][i] = m1[2][i] + m1[3][i];
|
||||
m2[3][i] = m1[2][i] - m1[3][i];
|
||||
m2[4][i] = m1[4][i] + m1[5][i];
|
||||
m2[5][i] = m1[4][i] - m1[5][i];
|
||||
m2[6][i] = m1[6][i] + m1[7][i];
|
||||
m2[7][i] = m1[6][i] - m1[7][i];
|
||||
}
|
||||
|
||||
for (i = 0; i < 64; ++i) {
|
||||
sad += abs(((int*)m2)[i]);
|
||||
}
|
||||
|
||||
sad = (sad + 2) >> 2;
|
||||
|
||||
return sad;
|
||||
}
|
||||
|
||||
// Function macro for defining hadamard calculating functions
|
||||
// for fixed size blocks. They calculate hadamard for integer
|
||||
// multiples of 8x8 with the 8x8 hadamard function.
|
||||
#define SATD_NXN(n, pixel_type, suffix) \
|
||||
static unsigned satd_ ## suffix ## _ ## n ## x ## n ## _generic( \
|
||||
const pixel_type * const block1, const pixel_type * const block2) \
|
||||
{ \
|
||||
unsigned x, y; \
|
||||
unsigned sum = 0; \
|
||||
for (y = 0; y < (n); y += 8) { \
|
||||
unsigned row = y * (n); \
|
||||
for (x = 0; x < (n); x += 8) { \
|
||||
sum += satd_16bit_8x8_general(&block1[row + x], (n), &block2[row + x], (n)); \
|
||||
} \
|
||||
} \
|
||||
return sum; \
|
||||
}
|
||||
|
||||
// Declare these functions to make sure the signature of the macro matches.
|
||||
static cost_pixel_nxn_func satd_8bit_4x4_generic;
|
||||
static cost_pixel_nxn_func satd_8bit_8x8_generic;
|
||||
static cost_pixel_nxn_func satd_8bit_16x16_generic;
|
||||
static cost_pixel_nxn_func satd_8bit_32x32_generic;
|
||||
static cost_pixel_nxn_func satd_8bit_64x64_generic;
|
||||
|
||||
// These macros define sadt_16bit_NxN for N = 8, 16, 32, 64
|
||||
SATD_NXN(8, pixel, 8bit)
|
||||
SATD_NXN(16, pixel, 8bit)
|
||||
SATD_NXN(32, pixel, 8bit)
|
||||
SATD_NXN(64, pixel, 8bit)
|
||||
|
||||
// Function macro for defining SAD calculating functions
|
||||
// for fixed size blocks.
|
||||
#define SAD_NXN(n, pixel_type, suffix) \
|
||||
static unsigned sad_ ## suffix ## _ ## n ## x ## n ## _generic( \
|
||||
const pixel_type * const block1, const pixel_type * const block2) \
|
||||
{ \
|
||||
unsigned i; \
|
||||
unsigned sum = 0; \
|
||||
for (i = 0; i < (n)*(n); ++i) { \
|
||||
sum += abs(block1[i] - block2[i]); \
|
||||
} \
|
||||
return sum; \
|
||||
}
|
||||
|
||||
// Declare these functions to make sure the signature of the macro matches.
|
||||
static cost_pixel_nxn_func sad_8bit_4x4_generic;
|
||||
static cost_pixel_nxn_func sad_8bit_8x8_generic;
|
||||
static cost_pixel_nxn_func sad_8bit_16x16_generic;
|
||||
static cost_pixel_nxn_func sad_8bit_32x32_generic;
|
||||
static cost_pixel_nxn_func sad_8bit_64x64_generic;
|
||||
|
||||
// These macros define sad_16bit_nxn functions for n = 4, 8, 16, 32, 64
|
||||
// with function signatures of cost_16bit_nxn_func.
|
||||
// They are used through get_pixel_sad_func.
|
||||
SAD_NXN(4, pixel, 8bit)
|
||||
SAD_NXN(8, pixel, 8bit)
|
||||
SAD_NXN(16, pixel, 8bit)
|
||||
SAD_NXN(32, pixel, 8bit)
|
||||
SAD_NXN(64, pixel, 8bit)
|
||||
|
||||
|
||||
static int strategy_register_picture_generic(void* opaque)
|
||||
{
|
||||
bool success = true;
|
||||
|
||||
success &= strategyselector_register(opaque, "reg_sad", "generic", 0, ®_sad_generic);
|
||||
|
||||
success &= strategyselector_register(opaque, "sad_8bit_4x4", "generic", 0, &sad_8bit_4x4_generic);
|
||||
success &= strategyselector_register(opaque, "sad_8bit_8x8", "generic", 0, &sad_8bit_8x8_generic);
|
||||
success &= strategyselector_register(opaque, "sad_8bit_16x16", "generic", 0, &sad_8bit_16x16_generic);
|
||||
success &= strategyselector_register(opaque, "sad_8bit_32x32", "generic", 0, &sad_8bit_32x32_generic);
|
||||
success &= strategyselector_register(opaque, "sad_8bit_64x64", "generic", 0, &sad_8bit_64x64_generic);
|
||||
|
||||
success &= strategyselector_register(opaque, "satd_8bit_4x4", "generic", 0, &satd_8bit_4x4_generic);
|
||||
success &= strategyselector_register(opaque, "satd_8bit_8x8", "generic", 0, &satd_8bit_8x8_generic);
|
||||
success &= strategyselector_register(opaque, "satd_8bit_16x16", "generic", 0, &satd_8bit_16x16_generic);
|
||||
success &= strategyselector_register(opaque, "satd_8bit_32x32", "generic", 0, &satd_8bit_32x32_generic);
|
||||
success &= strategyselector_register(opaque, "satd_8bit_64x64", "generic", 0, &satd_8bit_64x64_generic);
|
||||
|
||||
return success;
|
||||
}
|
||||
|
|
|
|||
|
|
@ -39,6 +39,7 @@ extern array_checksum_func array_checksum;
|
|||
int strategy_register_nal(void* opaque);
|
||||
|
||||
|
||||
#define STRATEGIES_NAL_EXPORTS {"array_checksum", (void**) &array_checksum}
|
||||
#define STRATEGIES_NAL_EXPORTS \
|
||||
{"array_checksum", (void**) &array_checksum},
|
||||
|
||||
#endif //STRATEGIES_NAL_H_
|
||||
|
|
|
|||
|
|
@ -1,5 +1,22 @@
|
|||
#include "strategies-picture.h"
|
||||
|
||||
// Define function pointers.
|
||||
reg_sad_func * reg_sad = 0;
|
||||
|
||||
cost_pixel_nxn_func * sad_8bit_4x4 = 0;
|
||||
cost_pixel_nxn_func * sad_8bit_8x8 = 0;
|
||||
cost_pixel_nxn_func * sad_8bit_16x16 = 0;
|
||||
cost_pixel_nxn_func * sad_8bit_32x32 = 0;
|
||||
cost_pixel_nxn_func * sad_8bit_64x64 = 0;
|
||||
|
||||
cost_pixel_nxn_func * satd_8bit_4x4 = 0;
|
||||
cost_pixel_nxn_func * satd_8bit_8x8 = 0;
|
||||
cost_pixel_nxn_func * satd_8bit_16x16 = 0;
|
||||
cost_pixel_nxn_func * satd_8bit_32x32 = 0;
|
||||
cost_pixel_nxn_func * satd_8bit_64x64 = 0;
|
||||
|
||||
|
||||
// Include inline functions.
|
||||
#include "picture/picture-generic.c"
|
||||
#if COMPILE_INTEL_SSE2
|
||||
#include "picture/picture-sse2.c"
|
||||
|
|
@ -12,9 +29,6 @@
|
|||
#endif
|
||||
|
||||
|
||||
reg_sad_func reg_sad;
|
||||
|
||||
|
||||
int strategy_register_picture(void* opaque) {
|
||||
if (!strategy_register_picture_generic(opaque)) return 0;
|
||||
|
||||
|
|
@ -40,3 +54,55 @@ int strategy_register_picture(void* opaque) {
|
|||
#endif //COMPILE_POWERPC
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* \brief Get a function that calculates SATD for NxN block.
|
||||
*
|
||||
* \param n Width of the region for which SATD is calculated.
|
||||
*
|
||||
* \returns Pointer to cost_16bit_nxn_func.
|
||||
*/
|
||||
cost_pixel_nxn_func * pixels_get_satd_func(unsigned n)
|
||||
{
|
||||
switch (n) {
|
||||
case 4:
|
||||
return satd_8bit_4x4;
|
||||
case 8:
|
||||
return satd_8bit_8x8;
|
||||
case 16:
|
||||
return satd_8bit_16x16;
|
||||
case 32:
|
||||
return satd_8bit_32x32;
|
||||
case 64:
|
||||
return satd_8bit_64x64;
|
||||
default:
|
||||
return NULL;
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* \brief Get a function that calculates SAD for NxN block.
|
||||
*
|
||||
* \param n Width of the region for which SAD is calculated.
|
||||
*
|
||||
* \returns Pointer to cost_16bit_nxn_func.
|
||||
*/
|
||||
cost_pixel_nxn_func * pixels_get_sad_func(unsigned n)
|
||||
{
|
||||
switch (n) {
|
||||
case 4:
|
||||
return sad_8bit_4x4;
|
||||
case 8:
|
||||
return sad_8bit_8x8;
|
||||
case 16:
|
||||
return sad_8bit_16x16;
|
||||
case 32:
|
||||
return sad_8bit_32x32;
|
||||
case 64:
|
||||
return sad_8bit_64x64;
|
||||
default:
|
||||
return NULL;
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -21,16 +21,47 @@
|
|||
|
||||
#include "../image.h"
|
||||
|
||||
//Function pointer to reg_sad
|
||||
typedef unsigned(*reg_sad_func)(const pixel *const data1, const pixel *const data2,
|
||||
const int width, const int height,
|
||||
const unsigned stride1, const unsigned stride2);
|
||||
extern reg_sad_func reg_sad;
|
||||
|
||||
typedef unsigned(reg_sad_func)(const pixel *const data1, const pixel *const data2,
|
||||
const int width, const int height,
|
||||
const unsigned stride1, const unsigned stride2);
|
||||
typedef unsigned (cost_pixel_nxn_func)(const pixel *block1, const pixel *block2);
|
||||
|
||||
|
||||
// Declare function pointers.
|
||||
extern reg_sad_func * reg_sad;
|
||||
|
||||
extern cost_pixel_nxn_func * sad_8bit_4x4;
|
||||
extern cost_pixel_nxn_func * sad_8bit_8x8;
|
||||
extern cost_pixel_nxn_func * sad_8bit_16x16;
|
||||
extern cost_pixel_nxn_func * sad_8bit_32x32;
|
||||
extern cost_pixel_nxn_func * sad_8bit_64x64;
|
||||
|
||||
extern cost_pixel_nxn_func * satd_8bit_4x4;
|
||||
extern cost_pixel_nxn_func * satd_8bit_8x8;
|
||||
extern cost_pixel_nxn_func * satd_8bit_16x16;
|
||||
extern cost_pixel_nxn_func * satd_8bit_32x32;
|
||||
extern cost_pixel_nxn_func * satd_8bit_64x64;
|
||||
|
||||
|
||||
int strategy_register_picture(void* opaque);
|
||||
cost_pixel_nxn_func * pixels_get_satd_func(unsigned n);
|
||||
cost_pixel_nxn_func * pixels_get_sad_func(unsigned n);
|
||||
|
||||
|
||||
#define STRATEGIES_PICTURE_EXPORTS \
|
||||
{"reg_sad", (void**) ®_sad}, \
|
||||
{"sad_8bit_4x4", (void**) &sad_8bit_4x4}, \
|
||||
{"sad_8bit_8x8", (void**) &sad_8bit_8x8}, \
|
||||
{"sad_8bit_16x16", (void**) &sad_8bit_16x16}, \
|
||||
{"sad_8bit_32x32", (void**) &sad_8bit_32x32}, \
|
||||
{"sad_8bit_64x64", (void**) &sad_8bit_64x64}, \
|
||||
{"satd_8bit_4x4", (void**) &satd_8bit_4x4}, \
|
||||
{"satd_8bit_8x8", (void**) &satd_8bit_8x8}, \
|
||||
{"satd_8bit_16x16", (void**) &satd_8bit_16x16}, \
|
||||
{"satd_8bit_32x32", (void**) &satd_8bit_32x32}, \
|
||||
{"satd_8bit_64x64", (void**) &satd_8bit_64x64}, \
|
||||
|
||||
|
||||
#define STRATEGIES_PICTURE_EXPORTS {"reg_sad", (void**) ®_sad}
|
||||
|
||||
#endif //STRATEGIES_PICTURE_H_
|
||||
|
|
|
|||
|
|
@ -100,9 +100,9 @@ int strategyselector_register(void * const opaque, const char * const type, cons
|
|||
new_strategy->priority = priority;
|
||||
new_strategy->fptr = fptr;
|
||||
}
|
||||
#ifdef _DEBUG
|
||||
#ifdef DEBUG_STRATEGYSELECTOR
|
||||
fprintf(stderr, "Registered strategy %s:%s with priority %d (%p)\n", type, strategy_name, priority, fptr);
|
||||
#endif //_DEBUG
|
||||
#endif //DEBUG_STRATEGYSELECTOR
|
||||
|
||||
return 1;
|
||||
}
|
||||
|
|
@ -139,7 +139,7 @@ static void* strategyselector_choose_for(const strategy_list * const strategies,
|
|||
return NULL;
|
||||
}
|
||||
|
||||
#ifdef _DEBUG
|
||||
#ifdef DEBUG_STRATEGYSELECTOR
|
||||
fprintf(stderr, "Choosing strategy for %s:\n", strategy_type);
|
||||
for (i=0; i < strategies->count; ++i) {
|
||||
if (strcmp(strategies->strategies[i].type, strategy_type) == 0) {
|
||||
|
|
@ -150,7 +150,7 @@ static void* strategyselector_choose_for(const strategy_list * const strategies,
|
|||
}
|
||||
}
|
||||
}
|
||||
#endif //_DEBUG
|
||||
#endif //DEBUG_STRATEGYSELECTOR
|
||||
|
||||
|
||||
if (max_priority_i == -1) {
|
||||
|
|
|
|||
|
|
@ -19,6 +19,10 @@
|
|||
* along with Kvazaar. If not, see <http://www.gnu.org/licenses/>.
|
||||
****************************************************************************/
|
||||
|
||||
#if defined(_DEBUG) && !defined(DEBUG_STRATEGYSELECTOR)
|
||||
# define DEBUG_STRATEGYSELECTOR
|
||||
#endif
|
||||
|
||||
//Hardware data (abstraction of defines). Extend for other compilers
|
||||
|
||||
#if defined(_M_IX86) || defined(__i586__) || defined(__i686__) || defined(_M_X64) || defined(_M_AMD64) || defined(__amd64__) || defined(__x86_64__)
|
||||
|
|
@ -137,8 +141,8 @@ int strategyselector_register(void *opaque, const char *type, const char *strate
|
|||
#include "strategies/strategies-picture.h"
|
||||
|
||||
static const strategy_to_select strategies_to_select[] = {
|
||||
STRATEGIES_NAL_EXPORTS,
|
||||
STRATEGIES_PICTURE_EXPORTS,
|
||||
STRATEGIES_NAL_EXPORTS
|
||||
STRATEGIES_PICTURE_EXPORTS
|
||||
{NULL, NULL},
|
||||
};
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue