generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Be consistent with SAD values

Browse Source 
SAD returns unsigned values. Make all the declarations the same.

Remove bestsad initialization and check. It is always set to the
result of a SAD call so it will never remain UINT_MAX

Use ja instead of jg to test unsigned comparison instead of signed.

Update test.

Change-Id: I46336ab45f4e60fc37caf20bd36bc5782079c7a5
This commit is contained in:
Johann 2012-06-26 11:37:33 -07:00
parent aa4f9dd153
commit 1161055129
8 changed files with 202 additions and 238 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
test/sad_test.cc
View File

@ -28,7 +28,7 @@ typedef unsigned int (*sad_m_by_n_fn_t)(const unsigned char *source_ptr,
int source_stride,
const unsigned char *reference_ptr,
int reference_stride,
int max_sad);
unsigned int max_sad);
using libvpx_test::ACMRandom;
@ -45,7 +45,7 @@ class SADTest : public PARAMS(int, int, sad_m_by_n_fn_t) {
}
sad_m_by_n_fn_t sad_fn_;
virtual unsigned int SAD(int max_sad) {
virtual unsigned int SAD(unsigned int max_sad) {
return sad_fn_(source_data_, source_stride_,
reference_data_, reference_stride_,
max_sad);
@ -111,13 +111,13 @@ class SADTest : public PARAMS(int, int, sad_m_by_n_fn_t) {
TEST_P(SADTest, MaxRef) {
FillConstant(source_data_, source_stride_, 0);
FillConstant(reference_data_, reference_stride_, 255);
CheckSad(INT_MAX);
CheckSad(UINT_MAX);
}
TEST_P(SADTest, MaxSrc) {
FillConstant(source_data_, source_stride_, 255);
FillConstant(reference_data_, reference_stride_, 0);
CheckSad(INT_MAX);
CheckSad(UINT_MAX);
}
TEST_P(SADTest, ShortRef) {
@ -125,7 +125,7 @@ TEST_P(SADTest, ShortRef) {
reference_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
CheckSad(INT_MAX);
CheckSad(UINT_MAX);
reference_stride_ = tmp_stride;
}
@ -136,7 +136,7 @@ TEST_P(SADTest, UnalignedRef) {
reference_stride_ -= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
CheckSad(INT_MAX);
CheckSad(UINT_MAX);
reference_stride_ = tmp_stride;
}
@ -145,7 +145,7 @@ TEST_P(SADTest, ShortSrc) {
source_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
CheckSad(INT_MAX);
CheckSad(UINT_MAX);
source_stride_ = tmp_stride;
}

12
vp8/common/mfqe.c
View File

@ -160,9 +160,9 @@ static void multiframe_quality_enhance_block
vsad = (vp8_variance8x8(v, uv_stride, vd, uvd_stride, &sse));
vsad = (sse + 32)>>6;
#else
sad = (vp8_sad16x16(y, y_stride, yd, yd_stride, INT_MAX)+128)>>8;
usad = (vp8_sad8x8(u, uv_stride, ud, uvd_stride, INT_MAX)+32)>>6;
vsad = (vp8_sad8x8(v, uv_stride, vd, uvd_stride, INT_MAX)+32)>>6;
sad = (vp8_sad16x16(y, y_stride, yd, yd_stride, UINT_MAX) + 128) >> 8;
usad = (vp8_sad8x8(u, uv_stride, ud, uvd_stride, UINT_MAX) + 32) >> 6;
vsad = (vp8_sad8x8(v, uv_stride, vd, uvd_stride, UINT_MAX)+ 32) >> 6;
#endif
}
else /* if (blksize == 8) */
@ -177,9 +177,9 @@ static void multiframe_quality_enhance_block
vsad = (vp8_variance4x4(v, uv_stride, vd, uvd_stride, &sse));
vsad = (sse + 8)>>4;
#else
sad = (vp8_sad8x8(y, y_stride, yd, yd_stride, INT_MAX)+32)>>6;
usad = (vp8_sad4x4(u, uv_stride, ud, uvd_stride, INT_MAX)+8)>>4;
vsad = (vp8_sad4x4(v, uv_stride, vd, uvd_stride, INT_MAX)+8)>>4;
sad = (vp8_sad8x8(y, y_stride, yd, yd_stride, UINT_MAX) + 32) >> 6;
usad = (vp8_sad4x4(u, uv_stride, ud, uvd_stride, UINT_MAX) + 8) >> 4;
vsad = (vp8_sad4x4(v, uv_stride, vd, uvd_stride, UINT_MAX) + 8) >> 4;
#endif
}

10
vp8/common/rtcd_defs.sh
View File

@ -295,23 +295,23 @@ vp8_variance_halfpixvar16x16_hv_media=vp8_variance_halfpixvar16x16_hv_armv6
#
# Single block SAD
#
prototype unsigned int vp8_sad4x4 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
prototype unsigned int vp8_sad4x4 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad"
specialize vp8_sad4x4 mmx sse2 neon
vp8_sad4x4_sse2=vp8_sad4x4_wmt
prototype unsigned int vp8_sad8x8 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
prototype unsigned int vp8_sad8x8 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad"
specialize vp8_sad8x8 mmx sse2 neon
vp8_sad8x8_sse2=vp8_sad8x8_wmt
prototype unsigned int vp8_sad8x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
prototype unsigned int vp8_sad8x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad"
specialize vp8_sad8x16 mmx sse2 neon
vp8_sad8x16_sse2=vp8_sad8x16_wmt
prototype unsigned int vp8_sad16x8 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
prototype unsigned int vp8_sad16x8 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad"
specialize vp8_sad16x8 mmx sse2 neon
vp8_sad16x8_sse2=vp8_sad16x8_wmt
prototype unsigned int vp8_sad16x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
prototype unsigned int vp8_sad16x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad"
specialize vp8_sad16x16 mmx sse2 sse3 media neon
vp8_sad16x16_sse2=vp8_sad16x16_wmt
vp8_sad16x16_media=vp8_sad16x16_armv6

208
vp8/common/sad_c.c
View File

@ -9,13 +9,14 @@
*/
#include <limits.h>
#include <stdlib.h>
#include "vpx_config.h"
#include "vpx/vpx_integer.h"
static unsigned int sad_mx_n_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
int max_sad, int m, int n)
unsigned int max_sad, int m, int n)
{
int r, c;
unsigned int sad = 0;
@ -43,21 +44,21 @@ static unsigned int sad_mx_n_c(const unsigned char *src_ptr, int src_stride,
unsigned int vp8_sad16x16_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
int max_sad)
unsigned int max_sad)
{
return sad_mx_n_c(src_ptr, src_stride, ref_ptr, ref_stride, max_sad, 16, 16);
}
unsigned int vp8_sad8x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
int max_sad)
unsigned int max_sad)
{
return sad_mx_n_c(src_ptr, src_stride, ref_ptr, ref_stride, max_sad, 8, 8);
}
unsigned int vp8_sad16x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
int max_sad)
unsigned int max_sad)
{
return sad_mx_n_c(src_ptr, src_stride, ref_ptr, ref_stride, max_sad, 16, 8);
@ -65,196 +66,181 @@ unsigned int vp8_sad16x8_c(const unsigned char *src_ptr, int src_stride,
unsigned int vp8_sad8x16_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
int max_sad)
unsigned int max_sad)
{
return sad_mx_n_c(src_ptr, src_stride, ref_ptr, ref_stride, max_sad, 8, 16);
}
unsigned int vp8_sad4x4_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
int max_sad)
unsigned int max_sad)
{
return sad_mx_n_c(src_ptr, src_stride, ref_ptr, ref_stride, max_sad, 4, 4);
}
void vp8_sad16x16x3_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
}
void vp8_sad16x16x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned short *sad_array
)
unsigned short *sad_array)
{
sad_array[0] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[3] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, 0x7fffffff);
sad_array[4] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, 0x7fffffff);
sad_array[5] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, 0x7fffffff);
sad_array[6] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, 0x7fffffff);
sad_array[7] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, 0x7fffffff);
sad_array[0] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
sad_array[3] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, UINT_MAX);
sad_array[4] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, UINT_MAX);
sad_array[5] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, UINT_MAX);
sad_array[6] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, UINT_MAX);
sad_array[7] = (unsigned short)vp8_sad16x16_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, UINT_MAX);
}
void vp8_sad16x8x3_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
}
void vp8_sad16x8x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned short *sad_array
)
unsigned short *sad_array)
{
sad_array[0] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[3] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, 0x7fffffff);
sad_array[4] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, 0x7fffffff);
sad_array[5] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, 0x7fffffff);
sad_array[6] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, 0x7fffffff);
sad_array[7] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, 0x7fffffff);
sad_array[0] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
sad_array[3] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, UINT_MAX);
sad_array[4] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, UINT_MAX);
sad_array[5] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, UINT_MAX);
sad_array[6] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, UINT_MAX);
sad_array[7] = (unsigned short)vp8_sad16x8_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, UINT_MAX);
}
void vp8_sad8x8x3_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
}
void vp8_sad8x8x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned short *sad_array
)
unsigned short *sad_array)
{
sad_array[0] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[3] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, 0x7fffffff);
sad_array[4] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, 0x7fffffff);
sad_array[5] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, 0x7fffffff);
sad_array[6] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, 0x7fffffff);
sad_array[7] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, 0x7fffffff);
sad_array[0] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
sad_array[3] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, UINT_MAX);
sad_array[4] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, UINT_MAX);
sad_array[5] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, UINT_MAX);
sad_array[6] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, UINT_MAX);
sad_array[7] = (unsigned short)vp8_sad8x8_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, UINT_MAX);
}
void vp8_sad8x16x3_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
}
void vp8_sad8x16x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned short *sad_array
)
unsigned short *sad_array)
{
sad_array[0] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[3] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, 0x7fffffff);
sad_array[4] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, 0x7fffffff);
sad_array[5] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, 0x7fffffff);
sad_array[6] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, 0x7fffffff);
sad_array[7] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, 0x7fffffff);
sad_array[0] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
sad_array[3] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, UINT_MAX);
sad_array[4] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, UINT_MAX);
sad_array[5] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, UINT_MAX);
sad_array[6] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, UINT_MAX);
sad_array[7] = (unsigned short)vp8_sad8x16_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, UINT_MAX);
}
void vp8_sad4x4x3_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
}
void vp8_sad4x4x8_c(const unsigned char *src_ptr, int src_stride,
const unsigned char *ref_ptr, int ref_stride,
unsigned short *sad_array
)
unsigned short *sad_array)
{
sad_array[0] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, 0x7fffffff);
sad_array[1] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, 0x7fffffff);
sad_array[2] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, 0x7fffffff);
sad_array[3] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, 0x7fffffff);
sad_array[4] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, 0x7fffffff);
sad_array[5] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, 0x7fffffff);
sad_array[6] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, 0x7fffffff);
sad_array[7] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, 0x7fffffff);
sad_array[0] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 0, ref_stride, UINT_MAX);
sad_array[1] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 1, ref_stride, UINT_MAX);
sad_array[2] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 2, ref_stride, UINT_MAX);
sad_array[3] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 3, ref_stride, UINT_MAX);
sad_array[4] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 4, ref_stride, UINT_MAX);
sad_array[5] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 5, ref_stride, UINT_MAX);
sad_array[6] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 6, ref_stride, UINT_MAX);
sad_array[7] = (unsigned short)vp8_sad4x4_c(src_ptr, src_stride, ref_ptr + 7, ref_stride, UINT_MAX);
}
void vp8_sad16x16x4d_c(const unsigned char *src_ptr, int src_stride,
const unsigned char * const ref_ptr[], int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[0], ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[1], ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[2], ref_stride, 0x7fffffff);
sad_array[3] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[3], ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[0], ref_stride, UINT_MAX);
sad_array[1] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[1], ref_stride, UINT_MAX);
sad_array[2] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[2], ref_stride, UINT_MAX);
sad_array[3] = vp8_sad16x16_c(src_ptr, src_stride, ref_ptr[3], ref_stride, UINT_MAX);
}
void vp8_sad16x8x4d_c(const unsigned char *src_ptr, int src_stride,
const unsigned char * const ref_ptr[], int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[0], ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[1], ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[2], ref_stride, 0x7fffffff);
sad_array[3] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[3], ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[0], ref_stride, UINT_MAX);
sad_array[1] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[1], ref_stride, UINT_MAX);
sad_array[2] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[2], ref_stride, UINT_MAX);
sad_array[3] = vp8_sad16x8_c(src_ptr, src_stride, ref_ptr[3], ref_stride, UINT_MAX);
}
void vp8_sad8x8x4d_c(const unsigned char *src_ptr, int src_stride,
const unsigned char * const ref_ptr[], int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[0], ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[1], ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[2], ref_stride, 0x7fffffff);
sad_array[3] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[3], ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[0], ref_stride, UINT_MAX);
sad_array[1] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[1], ref_stride, UINT_MAX);
sad_array[2] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[2], ref_stride, UINT_MAX);
sad_array[3] = vp8_sad8x8_c(src_ptr, src_stride, ref_ptr[3], ref_stride, UINT_MAX);
}
void vp8_sad8x16x4d_c(const unsigned char *src_ptr, int src_stride,
const unsigned char * const ref_ptr[], int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[0], ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[1], ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[2], ref_stride, 0x7fffffff);
sad_array[3] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[3], ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[0], ref_stride, UINT_MAX);
sad_array[1] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[1], ref_stride, UINT_MAX);
sad_array[2] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[2], ref_stride, UINT_MAX);
sad_array[3] = vp8_sad8x16_c(src_ptr, src_stride, ref_ptr[3], ref_stride, UINT_MAX);
}
void vp8_sad4x4x4d_c(const unsigned char *src_ptr, int src_stride,
const unsigned char * const ref_ptr[], int ref_stride,
unsigned int *sad_array
)
unsigned int *sad_array)
{
sad_array[0] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[0], ref_stride, 0x7fffffff);
sad_array[1] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[1], ref_stride, 0x7fffffff);
sad_array[2] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[2], ref_stride, 0x7fffffff);
sad_array[3] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[3], ref_stride, 0x7fffffff);
sad_array[0] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[0], ref_stride, UINT_MAX);
sad_array[1] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[1], ref_stride, UINT_MAX);
sad_array[2] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[2], ref_stride, UINT_MAX);
sad_array[3] = vp8_sad4x4_c(src_ptr, src_stride, ref_ptr[3], ref_stride, UINT_MAX);
}
/* Copy 2 macroblocks to a buffer */

6
vp8/common/variance.h
View File

@ -12,14 +12,12 @@
#ifndef VARIANCE_H
#define VARIANCE_H
typedef unsigned int(*vp8_sad_fn_t)
(
typedef unsigned int(*vp8_sad_fn_t)(
const unsigned char *src_ptr,
int source_stride,
const unsigned char *ref_ptr,
int ref_stride,
int max_sad
);
unsigned int max_sad);
typedef void (*vp8_copy32xn_fn_t)(
const unsigned char *src_ptr,

6
vp8/common/x86/sad_sse2.asm
View File

@ -115,7 +115,7 @@ sym(vp8_sad8x16_wmt):
movq rax, mm7
cmp eax, arg(4)
jg .x8x16sad_wmt_early_exit
ja .x8x16sad_wmt_early_exit
movq mm0, QWORD PTR [rsi]
movq mm1, QWORD PTR [rdi]
@ -176,7 +176,7 @@ sym(vp8_sad8x8_wmt):
movq rax, mm7
cmp eax, arg(4)
jg .x8x8sad_wmt_early_exit
ja .x8x8sad_wmt_early_exit
movq mm0, QWORD PTR [rsi]
movq mm1, QWORD PTR [rdi]
@ -285,7 +285,7 @@ sym(vp8_sad16x8_wmt):
movq rax, mm7
cmp eax, arg(4)
jg .x16x8sad_wmt_early_exit
ja .x16x8sad_wmt_early_exit
movq mm0, QWORD PTR [rsi]
movq mm2, QWORD PTR [rsi+8]

164
vp8/encoder/mcomp.c
View File

@ -198,7 +198,7 @@ void vp8_init3smotion_compensation(MACROBLOCK *x, int stride)
/* returns distortion + motion vector cost */
#define ERR(r,c) (MVC(r,c)+DIST(r,c))
/* checks if (r,c) has better score than previous best */
#define CHECK_BETTER(v,r,c) IFMVCV(r,c,{thismse = DIST(r,c); if((v = (MVC(r,c)+thismse)) < besterr) { besterr = v; br=r; bc=c; *distortion = thismse; *sse1 = sse; }}, v=INT_MAX;)
#define CHECK_BETTER(v,r,c) IFMVCV(r,c,{thismse = DIST(r,c); if((v = (MVC(r,c)+thismse)) < besterr) { besterr = v; br=r; bc=c; *distortion = thismse; *sse1 = sse; }}, v=UINT_MAX;)
int vp8_find_best_sub_pixel_step_iteratively(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
int_mv *bestmv, int_mv *ref_mv,
@ -212,7 +212,7 @@ int vp8_find_best_sub_pixel_step_iteratively(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
int rr = ref_mv->as_mv.row >> 1, rc = ref_mv->as_mv.col >> 1;
int br = bestmv->as_mv.row << 2, bc = bestmv->as_mv.col << 2;
int tr = br, tc = bc;
unsigned int besterr = INT_MAX;
unsigned int besterr;
unsigned int left, right, up, down, diag;
unsigned int sse;
unsigned int whichdir;
@ -870,7 +870,7 @@ int vp8_hex_search
int in_what_stride = pre_stride;
int br, bc;
int_mv this_mv;
unsigned int bestsad = 0x7fffffff;
unsigned int bestsad;
unsigned int thissad;
unsigned char *base_offset;
unsigned char *this_offset;
@ -894,8 +894,7 @@ int vp8_hex_search
this_offset = base_offset + (br * (pre_stride)) + bc;
this_mv.as_mv.row = br;
this_mv.as_mv.col = bc;
bestsad = vfp->sdf( what, what_stride, this_offset,
in_what_stride, 0x7fffffff)
bestsad = vfp->sdf(what, what_stride, this_offset, in_what_stride, UINT_MAX)
+ mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
#if CONFIG_MULTI_RES_ENCODING
@ -920,7 +919,7 @@ int vp8_hex_search
this_mv.as_mv.row = br + hex[i].row;
this_mv.as_mv.col = bc + hex[i].col;
this_offset = base_offset + (this_mv.as_mv.row * in_what_stride) + this_mv.as_mv.col;
thissad=vfp->sdf( what, what_stride, this_offset, in_what_stride, bestsad);
thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride, bestsad);
CHECK_BETTER
}
}else
@ -931,7 +930,7 @@ int vp8_hex_search
this_mv.as_mv.col = bc + hex[i].col;
CHECK_POINT
this_offset = base_offset + (this_mv.as_mv.row * in_what_stride) + this_mv.as_mv.col;
thissad=vfp->sdf( what, what_stride, this_offset, in_what_stride, bestsad);
thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride, bestsad);
CHECK_BETTER
}
}
@ -957,7 +956,7 @@ int vp8_hex_search
this_mv.as_mv.row = br + next_chkpts[k][i].row;
this_mv.as_mv.col = bc + next_chkpts[k][i].col;
this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
thissad = vfp->sdf( what, what_stride, this_offset, in_what_stride, bestsad);
thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride, bestsad);
CHECK_BETTER
}
}else
@ -968,7 +967,7 @@ int vp8_hex_search
this_mv.as_mv.col = bc + next_chkpts[k][i].col;
CHECK_POINT
this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
thissad = vfp->sdf( what, what_stride, this_offset, in_what_stride, bestsad);
thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride, bestsad);
CHECK_BETTER
}
}
@ -999,7 +998,7 @@ cal_neighbors:
this_mv.as_mv.row = br + neighbors[i].row;
this_mv.as_mv.col = bc + neighbors[i].col;
this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
thissad = vfp->sdf( what, what_stride, this_offset, in_what_stride, bestsad);
thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride, bestsad);
CHECK_BETTER
}
}else
@ -1010,7 +1009,7 @@ cal_neighbors:
this_mv.as_mv.col = bc + neighbors[i].col;
CHECK_POINT
this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
thissad = vfp->sdf( what, what_stride, this_offset, in_what_stride, bestsad);
thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride, bestsad);
CHECK_BETTER
}
}
@ -1061,7 +1060,8 @@ int vp8_diamond_search_sad_c
int tot_steps;
int_mv this_mv;
int bestsad = INT_MAX;
unsigned int bestsad;
unsigned int thissad;
int best_site = 0;
int last_site = 0;
@ -1072,7 +1072,6 @@ int vp8_diamond_search_sad_c
search_site *ss;
unsigned char *check_here;
int thissad;
int *mvsadcost[2];
int_mv fcenter_mv;
@ -1094,9 +1093,8 @@ int vp8_diamond_search_sad_c
best_address = in_what;
/* Check the starting position */
bestsad = fn_ptr->sdf(what, what_stride, in_what,
in_what_stride, 0x7fffffff)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
bestsad = fn_ptr->sdf(what, what_stride, in_what, in_what_stride, UINT_MAX)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
/* search_param determines the length of the initial step and hence
* the number of iterations 0 = initial step (MAX_FIRST_STEP) pel :
@ -1120,14 +1118,14 @@ int vp8_diamond_search_sad_c
{
check_here = ss[i].offset + best_address;
thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride, bestsad);
thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride, bestsad);
if (thissad < bestsad)
{
this_mv.as_mv.row = this_row_offset;
this_mv.as_mv.col = this_col_offset;
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
mvsadcost, sad_per_bit);
if (thissad < bestsad)
{
@ -1154,11 +1152,8 @@ int vp8_diamond_search_sad_c
this_mv.as_mv.row = best_mv->as_mv.row << 3;
this_mv.as_mv.col = best_mv->as_mv.col << 3;
if (bestsad == INT_MAX)
return INT_MAX;
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
int vp8_diamond_search_sadx4
@ -1189,7 +1184,8 @@ int vp8_diamond_search_sadx4
int tot_steps;
int_mv this_mv;
unsigned int bestsad = UINT_MAX;
unsigned int bestsad;
unsigned int thissad;
int best_site = 0;
int last_site = 0;
@ -1200,7 +1196,6 @@ int vp8_diamond_search_sadx4
search_site *ss;
unsigned char *check_here;
unsigned int thissad;
int *mvsadcost[2];
int_mv fcenter_mv;
@ -1222,9 +1217,8 @@ int vp8_diamond_search_sadx4
best_address = in_what;
/* Check the starting position */
bestsad = fn_ptr->sdf(what, what_stride,
in_what, in_what_stride, 0x7fffffff)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
bestsad = fn_ptr->sdf(what, what_stride, in_what, in_what_stride, UINT_MAX)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
/* search_param determines the length of the initial step and hence the
* number of iterations 0 = initial step (MAX_FIRST_STEP) pel : 1 =
@ -1291,14 +1285,14 @@ int vp8_diamond_search_sadx4
(this_row_offset > x->mv_row_min) && (this_row_offset < x->mv_row_max))
{
check_here = ss[i].offset + best_address;
thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride, bestsad);
thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride, bestsad);
if (thissad < bestsad)
{
this_mv.as_mv.row = this_row_offset;
this_mv.as_mv.col = this_col_offset;
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
mvsadcost, sad_per_bit);
if (thissad < bestsad)
{
@ -1325,11 +1319,8 @@ int vp8_diamond_search_sadx4
this_mv.as_mv.row = best_mv->as_mv.row << 3;
this_mv.as_mv.col = best_mv->as_mv.col << 3;
if (bestsad == INT_MAX)
return INT_MAX;
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
int vp8_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
@ -1347,11 +1338,11 @@ int vp8_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
unsigned char *bestaddress;
int_mv *best_mv = &d->bmi.mv;
int_mv this_mv;
int bestsad = INT_MAX;
unsigned int bestsad;
unsigned int thissad;
int r, c;
unsigned char *check_here;
int thissad;
int ref_row = ref_mv->as_mv.row;
int ref_col = ref_mv->as_mv.col;
@ -1378,8 +1369,8 @@ int vp8_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
/* Baseline value at the centre */
bestsad = fn_ptr->sdf(what, what_stride, bestaddress,
in_what_stride, 0x7fffffff)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
in_what_stride, UINT_MAX)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
/* Apply further limits to prevent us looking using vectors that
* stretch beyiond the UMV border
@ -1403,11 +1394,11 @@ int vp8_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
for (c = col_min; c < col_max; c++)
{
thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride, bestsad);
thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride, bestsad);
this_mv.as_mv.col = c;
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
if (thissad < bestsad)
{
@ -1424,11 +1415,8 @@ int vp8_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
this_mv.as_mv.row = best_mv->as_mv.row << 3;
this_mv.as_mv.col = best_mv->as_mv.col << 3;
if (bestsad < INT_MAX)
return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
else
return INT_MAX;
return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
@ -1446,11 +1434,11 @@ int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
unsigned char *bestaddress;
int_mv *best_mv = &d->bmi.mv;
int_mv this_mv;
unsigned int bestsad = UINT_MAX;
unsigned int bestsad;
unsigned int thissad;
int r, c;
unsigned char *check_here;
unsigned int thissad;
int ref_row = ref_mv->as_mv.row;
int ref_col = ref_mv->as_mv.col;
@ -1478,9 +1466,9 @@ int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
best_mv->as_mv.col = ref_col;
/* Baseline value at the centre */
bestsad = fn_ptr->sdf(what, what_stride,
bestaddress, in_what_stride, 0x7fffffff)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
bestsad = fn_ptr->sdf(what, what_stride, bestaddress,
in_what_stride, UINT_MAX)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
/* Apply further limits to prevent us looking using vectors that stretch
* beyond the UMV border
@ -1507,7 +1495,7 @@ int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
{
int i;
fn_ptr->sdx3f(what, what_stride, check_here , in_what_stride, sad_array);
fn_ptr->sdx3f(what, what_stride, check_here, in_what_stride, sad_array);
for (i = 0; i < 3; i++)
{
@ -1516,8 +1504,8 @@ int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
if (thissad < bestsad)
{
this_mv.as_mv.col = c;
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
if (thissad < bestsad)
{
@ -1535,13 +1523,13 @@ int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
while (c < col_max)
{
thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride, bestsad);
thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride, bestsad);
if (thissad < bestsad)
{
this_mv.as_mv.col = c;
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
if (thissad < bestsad)
{
@ -1561,11 +1549,8 @@ int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
this_mv.as_mv.row = best_mv->as_mv.row << 3;
this_mv.as_mv.col = best_mv->as_mv.col << 3;
if (bestsad < INT_MAX)
return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
else
return INT_MAX;
return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
int vp8_full_search_sadx8(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
@ -1583,11 +1568,11 @@ int vp8_full_search_sadx8(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
unsigned char *bestaddress;
int_mv *best_mv = &d->bmi.mv;
int_mv this_mv;
unsigned int bestsad = UINT_MAX;
unsigned int bestsad;
unsigned int thissad;
int r, c;
unsigned char *check_here;
unsigned int thissad;
int ref_row = ref_mv->as_mv.row;
int ref_col = ref_mv->as_mv.col;
@ -1617,8 +1602,8 @@ int vp8_full_search_sadx8(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
/* Baseline value at the centre */
bestsad = fn_ptr->sdf(what, what_stride,
bestaddress, in_what_stride, 0x7fffffff)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
bestaddress, in_what_stride, UINT_MAX)
+ mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
/* Apply further limits to prevent us looking using vectors that stretch
* beyond the UMV border
@ -1645,17 +1630,17 @@ int vp8_full_search_sadx8(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
{
int i;
fn_ptr->sdx8f(what, what_stride, check_here , in_what_stride, sad_array8);
fn_ptr->sdx8f(what, what_stride, check_here, in_what_stride, sad_array8);
for (i = 0; i < 8; i++)
{
thissad = (unsigned int)sad_array8[i];
thissad = sad_array8[i];
if (thissad < bestsad)
{
this_mv.as_mv.col = c;
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
mvsadcost, sad_per_bit);
if (thissad < bestsad)
{
@ -1728,11 +1713,8 @@ int vp8_full_search_sadx8(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
this_mv.as_mv.row = best_mv->as_mv.row << 3;
this_mv.as_mv.col = best_mv->as_mv.col << 3;
if (bestsad < INT_MAX)
return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
else
return INT_MAX;
return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
int vp8_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
@ -1752,9 +1734,9 @@ int vp8_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv
unsigned char *best_address = (unsigned char *)(base_pre + d->offset +
(ref_mv->as_mv.row * pre_stride) + ref_mv->as_mv.col);
unsigned char *check_here;
unsigned int thissad;
int_mv this_mv;
unsigned int bestsad = INT_MAX;
unsigned int bestsad;
unsigned int thissad;
int *mvsadcost[2];
int_mv fcenter_mv;
@ -1764,7 +1746,9 @@ int vp8_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv
fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
bestsad = fn_ptr->sdf(what, what_stride, best_address, in_what_stride, 0x7fffffff) + mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
bestsad = fn_ptr->sdf(what, what_stride, best_address,
in_what_stride, UINT_MAX)
+ mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
for (i=0; i<search_range; i++)
{
@ -1809,11 +1793,8 @@ int vp8_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv
this_mv.as_mv.row = ref_mv->as_mv.row << 3;
this_mv.as_mv.col = ref_mv->as_mv.col << 3;
if (bestsad < INT_MAX)
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
else
return INT_MAX;
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
int vp8_refining_search_sadx4(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
@ -1833,9 +1814,9 @@ int vp8_refining_search_sadx4(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
unsigned char *best_address = (unsigned char *)(base_pre + d->offset +
(ref_mv->as_mv.row * pre_stride) + ref_mv->as_mv.col);
unsigned char *check_here;
unsigned int thissad;
int_mv this_mv;
unsigned int bestsad = INT_MAX;
unsigned int bestsad;
unsigned int thissad;
int *mvsadcost[2];
int_mv fcenter_mv;
@ -1845,7 +1826,9 @@ int vp8_refining_search_sadx4(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
bestsad = fn_ptr->sdf(what, what_stride, best_address, in_what_stride, 0x7fffffff) + mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
bestsad = fn_ptr->sdf(what, what_stride, best_address,
in_what_stride, UINT_MAX)
+ mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
for (i=0; i<search_range; i++)
{
@ -1926,11 +1909,8 @@ int vp8_refining_search_sadx4(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
this_mv.as_mv.row = ref_mv->as_mv.row << 3;
this_mv.as_mv.col = ref_mv->as_mv.col << 3;
if (bestsad < INT_MAX)
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, (unsigned int *)(&thissad))
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
else
return INT_MAX;
return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
+ mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
}
#ifdef ENTROPY_STATS

20
vp8/encoder/rdopt.c
View File

@ -1683,16 +1683,16 @@ void vp8_cal_sad(VP8_COMP *cpi, MACROBLOCKD *xd, MACROBLOCK *x, int recon_yoffse
}else if(xd->mb_to_top_edge==0)
{ /* only has left MB for sad calculation. */
near_sad[0] = near_sad[2] = INT_MAX;
near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - 16,xd->dst.y_stride, 0x7fffffff);
near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - 16,xd->dst.y_stride, UINT_MAX);
}else if(xd->mb_to_left_edge ==0)
{ /* only has left MB for sad calculation. */
near_sad[1] = near_sad[2] = INT_MAX;
near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16,xd->dst.y_stride, 0x7fffffff);
near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16,xd->dst.y_stride, UINT_MAX);
}else
{
near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16,xd->dst.y_stride, 0x7fffffff);
near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - 16,xd->dst.y_stride, 0x7fffffff);
near_sad[2] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16 -16,xd->dst.y_stride, 0x7fffffff);
near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16,xd->dst.y_stride, UINT_MAX);
near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - 16,xd->dst.y_stride, UINT_MAX);
near_sad[2] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16 -16,xd->dst.y_stride, UINT_MAX);
}
if(cpi->common.last_frame_type != KEY_FRAME)
@ -1707,14 +1707,14 @@ void vp8_cal_sad(VP8_COMP *cpi, MACROBLOCKD *xd, MACROBLOCK *x, int recon_yoffse
if(xd->mb_to_bottom_edge==0) near_sad[7] = INT_MAX;
if(near_sad[4] != INT_MAX)
near_sad[4] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer - pre_y_stride *16, pre_y_stride, 0x7fffffff);
near_sad[4] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer - pre_y_stride *16, pre_y_stride, UINT_MAX);
if(near_sad[5] != INT_MAX)
near_sad[5] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer - 16, pre_y_stride, 0x7fffffff);
near_sad[3] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer, pre_y_stride, 0x7fffffff);
near_sad[5] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer - 16, pre_y_stride, UINT_MAX);
near_sad[3] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer, pre_y_stride, UINT_MAX);
if(near_sad[6] != INT_MAX)
near_sad[6] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer + 16, pre_y_stride, 0x7fffffff);
near_sad[6] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer + 16, pre_y_stride, UINT_MAX);
if(near_sad[7] != INT_MAX)
near_sad[7] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer + pre_y_stride *16, pre_y_stride, 0x7fffffff);
near_sad[7] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer + pre_y_stride *16, pre_y_stride, UINT_MAX);
}
if(cpi->common.last_frame_type != KEY_FRAME)