generic-library/openh264
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
b1fc94e314
openh264/codec/decoder/core/src/deblocking.cpp

726 lines
28 KiB
C++
Raw Normal View History

Initial Commit
2013-12-09 13:51:09 +01:00
/*!
* \copy
* Copyright (c) 2010-2013, Cisco Systems
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
* FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
* COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
* ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*
*
* \file deblocking.c
*
* \brief Interfaces introduced in frame deblocking filtering
*
* \date 08/02/2010
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
*
Initial Commit
2013-12-09 13:51:09 +01:00
*************************************************************************************
*/
#include "deblocking.h"
#include "cpu_core.h"
#include "fmo.h"
namespace WelsDec {
#define NO_SUPPORTED_FILTER_IDX (-1)
#define LEFT_FLAG_BIT 0
#define TOP_FLAG_BIT 1
#define LEFT_FLAG_MASK 0x01
#define TOP_FLAG_MASK 0x02
#define SAME_MB_DIFF_REFIDX
#define g_kuiAlphaTable(x) g_kuiAlphaTable[(x)+12]
#define g_kiBetaTable(x) g_kiBetaTable[(x)+12]
#define g_kiTc0Table(x) g_kiTc0Table[(x)+12]
#define MB_BS_MV(iRefIndex, iMotionVector, iMbXy, iMbBn, iIndex, iNeighIndex) \
(\
( iRefIndex[iMbXy][iIndex] - iRefIndex[iMbBn][iNeighIndex] )||\
( WELS_ABS( iMotionVector[iMbXy][iIndex][0] - iMotionVector[iMbBn][iNeighIndex][0] ) >= 4 ) ||\
( WELS_ABS( iMotionVector[iMbXy][iIndex][1] - iMotionVector[iMbBn][iNeighIndex][1] ) >= 4 )\
)
#if defined(SAME_MB_DIFF_REFIDX)
#define SMB_EDGE_MV(iRefIndex, iMotionVector, iIndex, iNeighIndex) \
(\
( iRefIndex[iIndex] - iRefIndex[iNeighIndex] )||(\
( WELS_ABS( iMotionVector[iIndex][0] - iMotionVector[iNeighIndex][0] ) &(~3) ) |\
( WELS_ABS( iMotionVector[iIndex][1] - iMotionVector[iNeighIndex][1] ) &(~3) ))\
)
#else
#define SMB_EDGE_MV(iRefIndex, iMotionVector, iIndex, iNeighIndex) \
(\
!!(( WELS_ABS( iMotionVector[iIndex][0] - iMotionVector[iNeighIndex][0] ) &(~3) ) |( WELS_ABS( iMotionVector[iIndex][1] - iMotionVector[iNeighIndex][1] ) &(~3) ))\
)
#endif
#define BS_EDGE(bsx1, iRefIndex, iMotionVector, iIndex, iNeighIndex) \
( (bsx1|SMB_EDGE_MV(iRefIndex, iMotionVector, iIndex, iNeighIndex))<<((uint8_t)(!!bsx1)))
#define GET_ALPHA_BETA_FROM_QP(iQp, iAlphaOffset, iBetaOffset, iIndex, iAlpha, iBeta) \
{\
iIndex = (iQp + iAlphaOffset);\
iAlpha = g_kuiAlphaTable(iIndex);\
iBeta = g_kiBetaTable((iQp + iBetaOffset));\
}
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
static const uint8_t g_kuiAlphaTable[52 + 24] = { //this table refers to Table 8-16 in H.264/AVC standard
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 4, 4, 5, 6,
7, 8, 9, 10, 12, 13, 15, 17, 20, 22,
25, 28, 32, 36, 40, 45, 50, 56, 63, 71,
80, 90, 101, 113, 127, 144, 162, 182, 203, 226,
255, 255
, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255
Initial Commit
2013-12-09 13:51:09 +01:00
};
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
static const int8_t g_kiBetaTable[52 + 24] = { //this table refers to Table 8-16 in H.264/AVC standard
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 2, 2, 2, 3,
3, 3, 3, 4, 4, 4, 6, 6, 7, 7,
8, 8, 9, 9, 10, 10, 11, 11, 12, 12,
13, 13, 14, 14, 15, 15, 16, 16, 17, 17,
18, 18
, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18
Initial Commit
2013-12-09 13:51:09 +01:00
};
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
static const int8_t g_kiTc0Table[52 + 24][4] = { //this table refers Table 8-17 in H.264/AVC standard
{ -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
{ -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
{ -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
{ -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
{ -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 1 },
{ -1, 0, 0, 1 }, { -1, 0, 0, 1 }, { -1, 0, 0, 1 }, { -1, 0, 1, 1 }, { -1, 0, 1, 1 }, { -1, 1, 1, 1 },
{ -1, 1, 1, 1 }, { -1, 1, 1, 1 }, { -1, 1, 1, 1 }, { -1, 1, 1, 2 }, { -1, 1, 1, 2 }, { -1, 1, 1, 2 },
{ -1, 1, 1, 2 }, { -1, 1, 2, 3 }, { -1, 1, 2, 3 }, { -1, 2, 2, 3 }, { -1, 2, 2, 4 }, { -1, 2, 3, 4 },
{ -1, 2, 3, 4 }, { -1, 3, 3, 5 }, { -1, 3, 4, 6 }, { -1, 3, 4, 6 }, { -1, 4, 5, 7 }, { -1, 4, 5, 8 },
{ -1, 4, 6, 9 }, { -1, 5, 7, 10 }, { -1, 6, 8, 11 }, { -1, 6, 8, 13 }, { -1, 7, 10, 14 }, { -1, 8, 11, 16 },
{ -1, 9, 12, 18 }, { -1, 10, 13, 20 }, { -1, 11, 15, 23 }, { -1, 13, 17, 25 }
, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }
, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }
Initial Commit
2013-12-09 13:51:09 +01:00
};
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
static const uint8_t g_kuiTableBIdx[2][8] = {
{
0, 4, 8, 12,
3, 7, 11, 15
},
Initial Commit
2013-12-09 13:51:09 +01:00
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
{
0, 1, 2, 3 ,
12, 13, 14, 15
},
Initial Commit
2013-12-09 13:51:09 +01:00
};
#define TC0_TBL_LOOKUP(tc, iIndexA, pBS, bChroma) \
{\
tc[0] = g_kiTc0Table(iIndexA)[pBS[0]] + bChroma;\
tc[1] = g_kiTc0Table(iIndexA)[pBS[1]] + bChroma;\
tc[2] = g_kiTc0Table(iIndexA)[pBS[2]] + bChroma;\
tc[3] = g_kiTc0Table(iIndexA)[pBS[3]] + bChroma;\
}
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
void_t inline DeblockingBSInsideMBAvsbase (int8_t* pNnzTab, uint8_t nBS[2][4][4], int32_t iLShiftFactor) {
uint32_t uiNnz32b0, uiNnz32b1, uiNnz32b2, uiNnz32b3;
FORCE_STACK_ALIGN_1D (uint8_t, uiBsx3, 4, 4);
uiNnz32b0 = * (uint32_t*) (pNnzTab + 0);
uiNnz32b1 = * (uint32_t*) (pNnzTab + 4);
uiNnz32b2 = * (uint32_t*) (pNnzTab + 8);
uiNnz32b3 = * (uint32_t*) (pNnzTab + 12);
* (uint32_t*)uiBsx3 = (uiNnz32b0 | (uiNnz32b0 >> 8)) << iLShiftFactor;
nBS[0][1][0] = uiBsx3[0];
nBS[0][2][0] = uiBsx3[1];
nBS[0][3][0] = uiBsx3[2];
* (uint32_t*)uiBsx3 = (uiNnz32b1 | (uiNnz32b1 >> 8)) << iLShiftFactor;
nBS[0][1][1] = uiBsx3[0];
nBS[0][2][1] = uiBsx3[1];
nBS[0][3][1] = uiBsx3[2];
* (uint32_t*)nBS[1][1] = (uiNnz32b0 | uiNnz32b1) << iLShiftFactor;
* (uint32_t*)uiBsx3 = (uiNnz32b2 | (uiNnz32b2 >> 8)) << iLShiftFactor;
nBS[0][1][2] = uiBsx3[0];
nBS[0][2][2] = uiBsx3[1];
nBS[0][3][2] = uiBsx3[2];
* (uint32_t*)nBS[1][2] = (uiNnz32b1 | uiNnz32b2) << iLShiftFactor;
* (uint32_t*)uiBsx3 = (uiNnz32b3 | (uiNnz32b3 >> 8)) << iLShiftFactor;
nBS[0][1][3] = uiBsx3[0];
nBS[0][2][3] = uiBsx3[1];
nBS[0][3][3] = uiBsx3[2];
* (uint32_t*)nBS[1][3] = (uiNnz32b2 | uiNnz32b3) << iLShiftFactor;
}
void_t static inline DeblockingBSInsideMBNormal (PDqLayer pCurDqLayer, uint8_t nBS[2][4][4], int8_t* pNnzTab,
int32_t iMbXy) {
uint32_t uiNnz32b0, uiNnz32b1, uiNnz32b2, uiNnz32b3;
int8_t* iRefIndex = pCurDqLayer->pRefIndex[LIST_0][iMbXy];
FORCE_STACK_ALIGN_1D (uint8_t, uiBsx4, 4, 4);
uiNnz32b0 = * (uint32_t*) (pNnzTab + 0);
uiNnz32b1 = * (uint32_t*) (pNnzTab + 4);
uiNnz32b2 = * (uint32_t*) (pNnzTab + 8);
uiNnz32b3 = * (uint32_t*) (pNnzTab + 12);
* (uint32_t*)uiBsx4 = (uiNnz32b0 | (uiNnz32b0 >> 8));
nBS[0][1][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 1, 0);
nBS[0][2][0] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 2, 1);
nBS[0][3][0] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 3, 2);
* (uint32_t*)uiBsx4 = (uiNnz32b1 | (uiNnz32b1 >> 8));
nBS[0][1][1] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 5, 4);
nBS[0][2][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 6, 5);
nBS[0][3][1] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 7, 6);
* (uint32_t*)uiBsx4 = (uiNnz32b2 | (uiNnz32b2 >> 8));
nBS[0][1][2] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 9, 8);
nBS[0][2][2] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 10, 9);
nBS[0][3][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 11, 10);
* (uint32_t*)uiBsx4 = (uiNnz32b3 | (uiNnz32b3 >> 8));
nBS[0][1][3] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 13, 12);
nBS[0][2][3] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 14, 13);
nBS[0][3][3] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 15, 14);
// horizontal
* (uint32_t*)uiBsx4 = (uiNnz32b0 | uiNnz32b1);
nBS[1][1][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 4, 0);
nBS[1][1][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 5, 1);
nBS[1][1][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 6, 2);
nBS[1][1][3] = BS_EDGE (uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 7, 3);
* (uint32_t*)uiBsx4 = (uiNnz32b1 | uiNnz32b2);
nBS[1][2][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 8, 4);
nBS[1][2][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 9, 5);
nBS[1][2][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 10, 6);
nBS[1][2][3] = BS_EDGE (uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 11, 7);
* (uint32_t*)uiBsx4 = (uiNnz32b2 | uiNnz32b3);
nBS[1][3][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 12, 8);
nBS[1][3][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 13, 9);
nBS[1][3][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 14, 10);
nBS[1][3][3] = BS_EDGE (uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 15, 11);
}
uint32_t DeblockingBsMarginalMBAvcbase (PDqLayer pCurDqLayer, int32_t iEdge, int32_t iNeighMb, int32_t iMbXy) {
int32_t i;
uint32_t uiBSx4;
//uint8_t* bS = static_cast<uint8_t*>(&uiBSx4);
uint8_t* pBS = (uint8_t*) (&uiBSx4);
uint32_t uiBIdx = * (uint32_t*) (&g_kuiTableBIdx[iEdge][0]);
uint32_t uiBnIdx = * (uint32_t*) (&g_kuiTableBIdx[iEdge][4]);
for (i = 0; i < 4; i++) {
if (pCurDqLayer->pNzc[iMbXy][uiBIdx & 0xff] | pCurDqLayer->pNzc[iNeighMb][uiBnIdx & 0xff]) {
pBS[i] = 2;
} else {
pBS[i] = MB_BS_MV (pCurDqLayer->pRefIndex[LIST_0], pCurDqLayer->pMv[LIST_0], iMbXy, iNeighMb, (uiBIdx & 0xff),
(uiBnIdx & 0xff));
}
uiBIdx = uiBIdx >> 8;
uiBnIdx = uiBnIdx >> 8;
}
return uiBSx4;
}
int32_t DeblockingAvailableNoInterlayer (PDqLayer pCurDqLayer, int32_t iFilterIdc) {
int32_t iMbY = pCurDqLayer->iMbY;
int32_t iMbX = pCurDqLayer->iMbX;
int32_t iMbXy = pCurDqLayer->iMbXyIndex;
BOOL_T bLeftFlag = FALSE;
BOOL_T bTopFlag = FALSE;
if (2 == iFilterIdc) {
bLeftFlag = (iMbX > 0) && (pCurDqLayer->pSliceIdc[iMbXy] == pCurDqLayer->pSliceIdc[iMbXy - 1]);
bTopFlag = (iMbY > 0) && (pCurDqLayer->pSliceIdc[iMbXy] == pCurDqLayer->pSliceIdc[iMbXy - pCurDqLayer->iMbWidth]);
} else { //if ( 0 == iFilterIdc )
bLeftFlag = (iMbX > 0);
bTopFlag = (iMbY > 0);
}
return (bLeftFlag << LEFT_FLAG_BIT) | (bTopFlag << TOP_FLAG_BIT);
}
void_t FilteringEdgeLumaH (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
TC0_TBL_LOOKUP (tc, iIndexA, pBS, 0);
pFilter->pLoopf->pfLumaDeblockingLT4Ver (pPix, iStride, iAlpha, iBeta, tc);
}
return;
}
void_t FilteringEdgeLumaV (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
TC0_TBL_LOOKUP (tc, iIndexA, pBS, 0);
pFilter->pLoopf->pfLumaDeblockingLT4Hor (pPix, iStride, iAlpha, iBeta, tc);
}
return;
}
void_t FilteringEdgeLumaIntraH (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
pFilter->pLoopf->pfLumaDeblockingEQ4Ver (pPix, iStride, iAlpha, iBeta);
}
return;
}
void_t FilteringEdgeLumaIntraV (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
pFilter->pLoopf->pfLumaDeblockingEQ4Hor (pPix, iStride, iAlpha, iBeta);
}
return;
}
void_t FilteringEdgeChromaH (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
TC0_TBL_LOOKUP (tc, iIndexA, pBS, 1);
pFilter->pLoopf->pfChromaDeblockingLT4Ver (pPixCb, pPixCr, iStride, iAlpha, iBeta, tc);
}
return;
}
void_t FilteringEdgeChromaV (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
TC0_TBL_LOOKUP (tc, iIndexA, pBS, 1);
pFilter->pLoopf->pfChromaDeblockingLT4Hor (pPixCb, pPixCr, iStride, iAlpha, iBeta, tc);
}
return;
}
void_t FilteringEdgeChromaIntraH (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
pFilter->pLoopf->pfChromaDeblockingEQ4Ver (pPixCb, pPixCr, iStride, iAlpha, iBeta);
}
return;
}
void_t FilteringEdgeChromaIntraV (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
uint8_t* pBS) {
int32_t iIndexA;
int32_t iAlpha;
int32_t iBeta;
GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
pFilter->pLoopf->pfChromaDeblockinEQ4Hor (pPixCb, pPixCr, iStride, iAlpha, iBeta);
}
return;
}
void_t DeblockingInterMb (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, uint8_t nBS[2][4][4],
int32_t iBoundryFlag) {
int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
int32_t iMbX = pCurDqLayer->iMbX;
int32_t iMbY = pCurDqLayer->iMbY;
int32_t iCurLumaQp = pCurDqLayer->pLumaQp[iMbXyIndex];
int32_t iCurChromaQp = pCurDqLayer->pChromaQp[iMbXyIndex];
int32_t iLineSize = pFilter->iCsStride[0];
int32_t iLineSizeUV = pFilter->iCsStride[1];
uint8_t* pDestY, * pDestCb, * pDestCr;
pDestY = pFilter->pCsData[0] + ((iMbY * iLineSize + iMbX) << 4);
pDestCb = pFilter->pCsData[1] + ((iMbY * iLineSizeUV + iMbX) << 3);
pDestCr = pFilter->pCsData[2] + ((iMbY * iLineSizeUV + iMbX) << 3);
if (iBoundryFlag & LEFT_FLAG_MASK) {
int32_t iLeftXyIndex = iMbXyIndex - 1;
pFilter->iLumaQP = (iCurLumaQp + pCurDqLayer->pLumaQp[iLeftXyIndex] + 1) >> 1;
pFilter->iChromaQP = (iCurChromaQp + pCurDqLayer->pChromaQp[iLeftXyIndex] + 1) >> 1;
if (nBS[0][0][0] == 0x04) {
FilteringEdgeLumaIntraV (pFilter, pDestY, iLineSize, NULL);
FilteringEdgeChromaIntraV (pFilter, pDestCb, pDestCr, iLineSizeUV, NULL);
} else {
if (* (uint32_t*)nBS[0][0] != 0) {
FilteringEdgeLumaV (pFilter, pDestY, iLineSize, nBS[0][0]);
FilteringEdgeChromaV (pFilter, pDestCb, pDestCr, iLineSizeUV, nBS[0][0]);
}
}
}
pFilter->iLumaQP = iCurLumaQp;
pFilter->iChromaQP = iCurChromaQp;
if (* (uint32_t*)nBS[0][1] != 0) {
FilteringEdgeLumaV (pFilter, &pDestY[1 << 2], iLineSize, nBS[0][1]);
}
if (* (uint32_t*)nBS[0][2] != 0) {
FilteringEdgeLumaV (pFilter, &pDestY[2 << 2], iLineSize, nBS[0][2]);
FilteringEdgeChromaV (pFilter, &pDestCb[2 << 1], &pDestCr[2 << 1], iLineSizeUV, nBS[0][2]);
}
if (* (uint32_t*)nBS[0][3] != 0) {
FilteringEdgeLumaV (pFilter, &pDestY[3 << 2], iLineSize, nBS[0][3]);
}
if (iBoundryFlag & TOP_FLAG_MASK) {
int32_t iTopXyIndex = iMbXyIndex - pCurDqLayer->iMbWidth;
pFilter->iLumaQP = (iCurLumaQp + pCurDqLayer->pLumaQp[iTopXyIndex] + 1) >> 1;
pFilter->iChromaQP = (iCurChromaQp + pCurDqLayer->pChromaQp[iTopXyIndex] + 1) >> 1;
if (nBS[1][0][0] == 0x04) {
FilteringEdgeLumaIntraH (pFilter, pDestY, iLineSize, NULL);
FilteringEdgeChromaIntraH (pFilter, pDestCb, pDestCr, iLineSizeUV, NULL);
} else {
if (* (uint32_t*)nBS[1][0] != 0) {
FilteringEdgeLumaH (pFilter, pDestY, iLineSize, nBS[1][0]);
FilteringEdgeChromaH (pFilter, pDestCb, pDestCr, iLineSizeUV, nBS[1][0]);
}
}
}
pFilter->iLumaQP = iCurLumaQp;
pFilter->iChromaQP = iCurChromaQp;
if (* (uint32_t*)nBS[1][1] != 0) {
FilteringEdgeLumaH (pFilter, &pDestY[ (1 << 2)*iLineSize], iLineSize, nBS[1][1]);
}
if (* (uint32_t*)nBS[1][2] != 0) {
FilteringEdgeLumaH (pFilter, &pDestY[ (2 << 2)*iLineSize], iLineSize, nBS[1][2]);
FilteringEdgeChromaH (pFilter, &pDestCb[ (2 << 1)*iLineSizeUV], &pDestCr[ (2 << 1)*iLineSizeUV], iLineSizeUV,
nBS[1][2]);
}
if (* (uint32_t*)nBS[1][3] != 0) {
FilteringEdgeLumaH (pFilter, &pDestY[ (3 << 2)*iLineSize], iLineSize, nBS[1][3]);
}
}
void_t /*__FASTCALL*/ FilteringEdgeLumaHV (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
int32_t iMbX = pCurDqLayer->iMbX;
int32_t iMbY = pCurDqLayer->iMbY;
int32_t iMbWidth = pCurDqLayer->iMbWidth;
int32_t iLineSize = pFilter->iCsStride[0];
uint8_t* pDestY;
int32_t iCurQp;
int32_t iIndexA, iAlpha, iBeta;
FORCE_STACK_ALIGN_1D (int8_t, iTc, 4, 16);
FORCE_STACK_ALIGN_1D (uint8_t, uiBSx4, 4, 4);
pDestY = pFilter->pCsData[0] + ((iMbY * iLineSize + iMbX) << 4);
iCurQp = pCurDqLayer->pLumaQp[iMbXyIndex];
* (uint32_t*)uiBSx4 = 0x03030303;
// luma v
if (iBoundryFlag & LEFT_FLAG_MASK) {
pFilter->iLumaQP = (iCurQp + pCurDqLayer->pLumaQp[iMbXyIndex - 1] + 1) >> 1;
FilteringEdgeLumaIntraV (pFilter, pDestY, iLineSize, NULL);
}
pFilter->iLumaQP = iCurQp;
GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
TC0_TBL_LOOKUP (iTc, iIndexA, uiBSx4, 0);
pFilter->pLoopf->pfLumaDeblockingLT4Hor (&pDestY[1 << 2], iLineSize, iAlpha, iBeta, iTc);
pFilter->pLoopf->pfLumaDeblockingLT4Hor (&pDestY[2 << 2], iLineSize, iAlpha, iBeta, iTc);
pFilter->pLoopf->pfLumaDeblockingLT4Hor (&pDestY[3 << 2], iLineSize, iAlpha, iBeta, iTc);
}
// luma h
if (iBoundryFlag & TOP_FLAG_MASK) {
pFilter->iLumaQP = (iCurQp + pCurDqLayer->pLumaQp[iMbXyIndex - iMbWidth] + 1) >> 1;
FilteringEdgeLumaIntraH (pFilter, pDestY, iLineSize, NULL);
}
pFilter->iLumaQP = iCurQp;
if (iAlpha | iBeta) {
pFilter->pLoopf->pfLumaDeblockingLT4Ver (&pDestY[ (1 << 2)*iLineSize], iLineSize, iAlpha, iBeta, iTc);
pFilter->pLoopf->pfLumaDeblockingLT4Ver (&pDestY[ (2 << 2)*iLineSize], iLineSize, iAlpha, iBeta, iTc);
pFilter->pLoopf->pfLumaDeblockingLT4Ver (&pDestY[ (3 << 2)*iLineSize], iLineSize, iAlpha, iBeta, iTc);
}
}
void_t /*__FASTCALL*/ FilteringEdgeChromaHV (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
int32_t iMbX = pCurDqLayer->iMbX;
int32_t iMbY = pCurDqLayer->iMbY;
int32_t iMbWidth = pCurDqLayer->iMbWidth;
int32_t iLineSize = pFilter->iCsStride[1];
uint8_t* pDestCb, *pDestCr;
int32_t iCurQp;
int32_t iIndexA, iAlpha, iBeta;
FORCE_STACK_ALIGN_1D (int8_t, iTc, 4, 16);
FORCE_STACK_ALIGN_1D (uint8_t, uiBSx4, 4, 4);
pDestCb = pFilter->pCsData[1] + ((iMbY * iLineSize + iMbX) << 3);
pDestCr = pFilter->pCsData[2] + ((iMbY * iLineSize + iMbX) << 3);
iCurQp = pCurDqLayer->pChromaQp[iMbXyIndex];
* (uint32_t*)uiBSx4 = 0x03030303;
// chroma v
if (iBoundryFlag & LEFT_FLAG_MASK) {
pFilter->iChromaQP = (iCurQp + pCurDqLayer->pChromaQp[iMbXyIndex - 1] + 1) >> 1;
FilteringEdgeChromaIntraV (pFilter, pDestCb, pDestCr, iLineSize, NULL);
}
pFilter->iChromaQP = iCurQp;
GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
iBeta);
if (iAlpha | iBeta) {
TC0_TBL_LOOKUP (iTc, iIndexA, uiBSx4, 1);
pFilter->pLoopf->pfChromaDeblockingLT4Hor (&pDestCb[2 << 1], &pDestCr[2 << 1], iLineSize, iAlpha, iBeta, iTc);
}
// chroma h
if (iBoundryFlag & TOP_FLAG_MASK) {
pFilter->iChromaQP = (iCurQp + pCurDqLayer->pChromaQp[iMbXyIndex - iMbWidth] + 1) >> 1;
FilteringEdgeChromaIntraH (pFilter, pDestCb, pDestCr, iLineSize, NULL);
}
pFilter->iChromaQP = iCurQp;
if (iAlpha | iBeta) {
pFilter->pLoopf->pfChromaDeblockingLT4Ver (&pDestCb[ (2 << 1)*iLineSize], &pDestCr[ (2 << 1)*iLineSize], iLineSize,
iAlpha, iBeta, iTc);
}
Initial Commit
2013-12-09 13:51:09 +01:00
}
// merge h&v lookup table operation to save performance
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
void_t DeblockingIntraMb (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
FilteringEdgeLumaHV (pCurDqLayer, pFilter, iBoundryFlag);
FilteringEdgeChromaHV (pCurDqLayer, pFilter, iBoundryFlag);
}
void_t WelsDeblockingMb (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
uint8_t nBS[2][4][4] = { 0 };
int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
int32_t iCurMbType = pCurDqLayer->pMbType[iMbXyIndex];
int32_t iMbNb;
switch (iCurMbType) {
case MB_TYPE_INTRA4x4:
case MB_TYPE_INTRA16x16:
case MB_TYPE_INTRA_PCM:
DeblockingIntraMb (pCurDqLayer, pFilter, iBoundryFlag);
break;
default:
if (iBoundryFlag & LEFT_FLAG_MASK) {
iMbNb = iMbXyIndex - 1;
* (uint32_t*)nBS[0][0] = IS_INTRA (pCurDqLayer->pMbType[iMbNb]) ? 0x04040404 : DeblockingBsMarginalMBAvcbase (
pCurDqLayer, 0, iMbNb, iMbXyIndex);
} else {
* (uint32_t*)nBS[0][0] = 0;
}
if (iBoundryFlag & TOP_FLAG_MASK) {
iMbNb = iMbXyIndex - pCurDqLayer->iMbWidth;
* (uint32_t*)nBS[1][0] = IS_INTRA (pCurDqLayer->pMbType[iMbNb]) ? 0x04040404 : DeblockingBsMarginalMBAvcbase (
pCurDqLayer, 1, iMbNb, iMbXyIndex);
} else {
* (uint32_t*)nBS[1][0] = 0;
}
//SKIP MB_16x16 or others
if (iCurMbType != MB_TYPE_SKIP) {
if (iCurMbType == MB_TYPE_16x16) {
DeblockingBSInsideMBAvsbase (pCurDqLayer->pNzc[iMbXyIndex], nBS, 1);
} else {
DeblockingBSInsideMBNormal (pCurDqLayer, nBS, pCurDqLayer->pNzc[iMbXyIndex], iMbXyIndex);
}
} else {
* (uint32_t*)nBS[0][1] = * (uint32_t*)nBS[0][2] = * (uint32_t*)nBS[0][3] =
* (uint32_t*)nBS[1][1] = * (uint32_t*)nBS[1][2] = * (uint32_t*)nBS[1][3] = 0;
}
DeblockingInterMb (pCurDqLayer, pFilter, nBS, iBoundryFlag);
break;
}
}
Initial Commit
2013-12-09 13:51:09 +01:00
/*!
* \brief AVC slice deblocking filtering target layer
*
* \param dec Wels avc decoder context
*
* \return NONE
*/
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
void_t WelsDeblockingFilterSlice (PWelsDecoderContext pCtx, PDeblockingFilterMbFunc pDeblockMb) {
PDqLayer pCurDqLayer = pCtx->pCurDqLayer;
PSliceHeaderExt pSliceHeaderExt = &pCurDqLayer->sLayerInfo.sSliceInLayer.sSliceHeaderExt;
int32_t iMbWidth = pCurDqLayer->iMbWidth;
int32_t iTotalMbCount = pSliceHeaderExt->sSliceHeader.pSps->uiTotalMbCount;
SDeblockingFilter pFilter = {0};
PFmo pFmo = pCtx->pFmo;
int32_t iNextMbXyIndex = 0;
int32_t iTotalNumMb = pCurDqLayer->sLayerInfo.sSliceInLayer.iTotalMbInCurSlice;
int32_t iCountNumMb = 0;
int32_t iBoundryFlag;
int32_t iFilterIdc = pCurDqLayer->sLayerInfo.sSliceInLayer.sSliceHeaderExt.sSliceHeader.uiDisableDeblockingFilterIdc;
/* Step1: parameters set */
pFilter.pCsData[0] = pCtx->pDec->pData[0];
pFilter.pCsData[1] = pCtx->pDec->pData[1];
pFilter.pCsData[2] = pCtx->pDec->pData[2];
pFilter.iCsStride[0] = pCtx->pDec->iLinesize[0];
pFilter.iCsStride[1] = pCtx->pDec->iLinesize[1];
pFilter.eSliceType = (ESliceType) pCurDqLayer->sLayerInfo.sSliceInLayer.eSliceType;
pFilter.iSliceAlphaC0Offset = pSliceHeaderExt->sSliceHeader.iSliceAlphaC0Offset;
pFilter.iSliceBetaOffset = pSliceHeaderExt->sSliceHeader.iSliceBetaOffset;
pFilter.pLoopf = &pCtx->sDeblockingFunc;
/* Step2: macroblock deblocking */
if (0 == iFilterIdc || 2 == iFilterIdc) {
iNextMbXyIndex = pSliceHeaderExt->sSliceHeader.iFirstMbInSlice;
pCurDqLayer->iMbX = iNextMbXyIndex % iMbWidth;
pCurDqLayer->iMbY = iNextMbXyIndex / iMbWidth;
pCurDqLayer->iMbXyIndex = iNextMbXyIndex;
do {
iBoundryFlag = DeblockingAvailableNoInterlayer (pCurDqLayer, iFilterIdc);
pDeblockMb (pCurDqLayer, &pFilter, iBoundryFlag);
++iCountNumMb;
if (iCountNumMb >= iTotalNumMb) {
break;
}
if (pSliceHeaderExt->sSliceHeader.pPps->uiNumSliceGroups > 1) {
iNextMbXyIndex = FmoNextMb (pFmo, iNextMbXyIndex);
} else {
++iNextMbXyIndex;
}
if (-1 == iNextMbXyIndex || iNextMbXyIndex >= iTotalMbCount) { // slice group boundary or end of a frame
break;
}
pCurDqLayer->iMbX = iNextMbXyIndex % iMbWidth;
pCurDqLayer->iMbY = iNextMbXyIndex / iMbWidth;
pCurDqLayer->iMbXyIndex = iNextMbXyIndex;
} while (1);
}
Initial Commit
2013-12-09 13:51:09 +01:00
}
/*!
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
* \brief deblocking module initialize
Initial Commit
2013-12-09 13:51:09 +01:00
*
* \param pf
* cpu
*
* \return NONE
*/
Pretty printed the C++ code with astyle
2013-12-12 23:21:12 +01:00
void_t DeblockingInit (SDeblockingFunc* pFunc, int32_t iCpu) {
pFunc->pfLumaDeblockingLT4Ver = DeblockLumaLt4V_c;
pFunc->pfLumaDeblockingEQ4Ver = DeblockLumaEq4V_c;
pFunc->pfLumaDeblockingLT4Hor = DeblockLumaLt4H_c;
pFunc->pfLumaDeblockingEQ4Hor = DeblockLumaEq4H_c;
pFunc->pfChromaDeblockingLT4Ver = DeblockChromaLt4V_c;
pFunc->pfChromaDeblockingEQ4Ver = DeblockChromaEq4V_c;
pFunc->pfChromaDeblockingLT4Hor = DeblockChromaLt4H_c;
pFunc->pfChromaDeblockinEQ4Hor = DeblockChromaEq4H_c;
#ifdef X86_ASM
if (iCpu & WELS_CPU_SSE2) {
pFunc->pfLumaDeblockingLT4Ver = DeblockLumaLt4V_sse2;
pFunc->pfLumaDeblockingEQ4Ver = DeblockLumaEq4V_sse2;
pFunc->pfLumaDeblockingLT4Hor = DeblockLumaLt4H_sse2;
pFunc->pfLumaDeblockingEQ4Hor = DeblockLumaEq4H_sse2;
pFunc->pfChromaDeblockingLT4Ver = DeblockChromaLt4V_sse2;
pFunc->pfChromaDeblockingEQ4Ver = DeblockChromaEq4V_sse2;
pFunc->pfChromaDeblockingLT4Hor = DeblockChromaLt4H_sse2;
pFunc->pfChromaDeblockinEQ4Hor = DeblockChromaEq4H_sse2;
}
Initial Commit
2013-12-09 13:51:09 +01:00
#endif
}
file a new commit Squashed commit of the following: commit 26be4c66686d2d75ba1eb381ae562d30474429b2 Merge: 91bde82 8df0566 Author: Licai Guo <guolicai@gmail.com> Date: Wed Jan 8 16:56:49 2014 -0800 Merge remote-tracking branch 'upstream/master' into move_to_common Conflicts: codec/common/targets.mk codec/console/dec/targets.mk codec/decoder/targets.mk codec/encoder/targets.mk codec/processing/targets.mk commit 91bde8209dad40eaa0b26cc75467e4ce52b4c8ae Merge: c827054 cc2ffee Author: Licai Guo <guolicai@gmail.com> Date: Tue Jan 7 20:16:11 2014 -0800 Merge remote-tracking branch 'upstream/master' into move_to_common commit c8270543fab8e40cb4aebf62d7646068afe19484 Author: Licai Guo <guolicai@gmail.com> Date: Mon Jan 6 19:12:39 2014 -0800 enable WelsScan4x4DcAc_ssse3 commit c8d2823d54e54cab146098302441ef44a4638382 Author: Licai Guo <guolicai@gmail.com> Date: Mon Jan 6 18:46:06 2014 -0800 remove trailing white spaces commit 39866a9cf54db6120aabd6a01afd4aedbead8af0 Author: Licai Guo <guolicai@gmail.com> Date: Mon Jan 6 18:06:10 2014 -0800 move common code(deblocking and cpu) to common, rename decoder's Intra-prediction functions prefix.
2014-01-09 02:03:00 +01:00
} // namespace WelsDec
Reference in New Issue Copy Permalink