TEncCu::compressCU分析
TEncCu::compressCU对应的代码如下：
Void TEncCu::compressCU( TComDataCU*& rpcCU )
{
// initialize CU data
m_ppcBestCU[0]->initCU( rpcCU->getPic(), rpcCU->getAddr() );
m_ppcTempCU[0]->initCU( rpcCU->getPic(), rpcCU->getAddr() );
// analysis of CU
xCompressCU( m_ppcBestCU[0], m_ppcTempCU[0], 0 );
#if ADAPTIVE_QP_SELECTION
if( m_pcEncCfg->getUseAdaptQpSelect() )
{
if(rpcCU->getSlice()->getSliceType()!=I_SLICE) //IIII
{
xLcuCollectARLStats( rpcCU);
}
}
#endif
}
此段代码主要包括三个部分：
一是Input：
TComDataCU*& rpcCU指向CU的参数。
二是initialize CU data：
m_ppcBestCU[0]->initCU( rpcCU->getPic(), rpcCU->getAddr() );
m_ppcTempCU[0]->initCU( rpcCU->getPic(), rpcCU->getAddr() );TempCU表示当前CU的切法，BestCU表示前面算出的最好的CU切法。
m_ppcBestCU / m_ppcTempCU：存储最好的/当前的QP和在每一个深度的预测模式决策。
下面进入initCU：TComDataCU::initCU，对应的代码如下：
Void TComDataCU::initCU( TComPic* pcPic, UInt iCUAddr )
{
m_pcPic = pcPic;
m_pcSlice = pcPic->getSlice(pcPic->getCurrSliceIdx());
m_uiCUAddr = iCUAddr;
m_uiCUPelX = ( iCUAddr % pcPic->getFrameWidthInCU() ) * g_uiMaxCUWidth;
m_uiCUPelY = ( iCUAddr / pcPic->getFrameWidthInCU() ) * g_uiMaxCUHeight;
m_uiAbsIdxInLCU = 0;
m_dTotalCost = MAX_DOUBLE;
m_uiTotalDistortion = 0;
m_uiTotalBits = 0;
m_uiTotalBins = 0;
m_uiNumPartition = pcPic->getNumPartInCU();
for(Int i=0; i<pcPic->getNumPartInCU(); i++)
{
if(pcPic->getPicSym()->getInverseCUOrderMap(iCUAddr)*pcPic->getNumPartInCU()+i>=getSlice()->getSliceCurStartCUAddr())
{
m_sliceStartCU[i]=getSlice()->getSliceCurStartCUAddr();
}
else
{
m_sliceStartCU[i]=pcPic->getCU(getAddr())->m_sliceStartCU[i];
}
}
for(Int i=0; i<pcPic->getNumPartInCU(); i++)
{
if(pcPic->getPicSym()->getInverseCUOrderMap(iCUAddr)*pcPic->getNumPartInCU()+i>=getSlice()->getSliceSegmentCurStartCUAddr())
{
m_sliceSegmentStartCU[i]=getSlice()->getSliceSegmentCurStartCUAddr();
}
else
{
m_sliceSegmentStartCU[i]=pcPic->getCU(getAddr())->m_sliceSegmentStartCU[i];
}
}
Int partStartIdx = getSlice()->getSliceSegmentCurStartCUAddr() - pcPic->getPicSym()->getInverseCUOrderMap(iCUAddr) * pcPic->getNumPartInCU();
Int numElements = min<Int>( partStartIdx, m_uiNumPartition );
for ( Int ui = 0; ui < numElements; ui++ )
{
TComDataCU * pcFrom = pcPic->getCU(getAddr());
m_skipFlag[ui] = pcFrom->getSkipFlag(ui);
m_pePartSize[ui] = pcFrom->getPartitionSize(ui);
m_pePredMode[ui] = pcFrom->getPredictionMode(ui);
m_CUTransquantBypass[ui] = pcFrom->getCUTransquantBypass(ui);
m_puhDepth[ui] = pcFrom->getDepth(ui);
m_puhWidth [ui] = pcFrom->getWidth(ui);
m_puhHeight [ui] = pcFrom->getHeight(ui);
m_puhTrIdx [ui] = pcFrom->getTransformIdx(ui);
m_puhTransformSkip[0][ui] = pcFrom->getTransformSkip(ui,TEXT_LUMA);
m_puhTransformSkip[1][ui] = pcFrom->getTransformSkip(ui,TEXT_CHROMA_U);
m_puhTransformSkip[2][ui] = pcFrom->getTransformSkip(ui,TEXT_CHROMA_V);
m_apiMVPIdx[0][ui] = pcFrom->m_apiMVPIdx[0][ui];;
m_apiMVPIdx[1][ui] = pcFrom->m_apiMVPIdx[1][ui];
m_apiMVPNum[0][ui] = pcFrom->m_apiMVPNum[0][ui];
m_apiMVPNum[1][ui] = pcFrom->m_apiMVPNum[1][ui];
m_phQP[ui]=pcFrom->m_phQP[ui];
m_pbMergeFlag[ui]=pcFrom->m_pbMergeFlag[ui];
m_puhMergeIndex[ui]=pcFrom->m_puhMergeIndex[ui];
m_puhLumaIntraDir[ui]=pcFrom->m_puhLumaIntraDir[ui];
m_puhChromaIntraDir[ui]=pcFrom->m_puhChromaIntraDir[ui];
m_puhInterDir[ui]=pcFrom->m_puhInterDir[ui];
m_puhCbf[0][ui]=pcFrom->m_puhCbf[0][ui];
m_puhCbf[1][ui]=pcFrom->m_puhCbf[1][ui];
m_puhCbf[2][ui]=pcFrom->m_puhCbf[2][ui];
m_pbIPCMFlag[ui] = pcFrom->m_pbIPCMFlag[ui];
}
Int firstElement = max<Int>( partStartIdx, 0 );
numElements = m_uiNumPartition - firstElement;
if ( numElements > 0 )
{
memset( m_skipFlag + firstElement, false, numElements * sizeof( *m_skipFlag ) );
memset( m_pePartSize + firstElement, SIZE_NONE, numElements * sizeof( *m_pePartSize ) );
memset( m_pePredMode + firstElement, MODE_NONE, numElements * sizeof( *m_pePredMode ) );
memset( m_CUTransquantBypass+ firstElement, false, numElements * sizeof( *m_CUTransquantBypass) );
memset( m_puhDepth + firstElement, 0, numElements * sizeof( *m_puhDepth ) );
memset( m_puhTrIdx + firstElement, 0, numElements * sizeof( *m_puhTrIdx ) );
memset( m_puhTransformSkip[0] + firstElement, 0, numElements * sizeof( *m_puhTransformSkip[0]) );
memset( m_puhTransformSkip[1] + firstElement, 0, numElements * sizeof( *m_puhTransformSkip[1]) );
memset( m_puhTransformSkip[2] + firstElement, 0, numElements * sizeof( *m_puhTransformSkip[2]) );
memset( m_puhWidth + firstElement, g_uiMaxCUWidth, numElements * sizeof( *m_puhWidth ) );
memset( m_puhHeight + firstElement, g_uiMaxCUHeight, numElements * sizeof( *m_puhHeight ) );
memset( m_apiMVPIdx[0] + firstElement, -1, numElements * sizeof( *m_apiMVPIdx[0] ) );
memset( m_apiMVPIdx[1] + firstElement, -1, numElements * sizeof( *m_apiMVPIdx[1] ) );
memset( m_apiMVPNum[0] + firstElement, -1, numElements * sizeof( *m_apiMVPNum[0] ) );
memset( m_apiMVPNum[1] + firstElement, -1, numElements * sizeof( *m_apiMVPNum[1] ) );
memset( m_phQP + firstElement, getSlice()->getSliceQp(), numElements * sizeof( *m_phQP ) );
memset( m_pbMergeFlag + firstElement, false, numElements * sizeof( *m_pbMergeFlag ) );
memset( m_puhMergeIndex + firstElement, 0, numElements * sizeof( *m_puhMergeIndex ) );
memset( m_puhLumaIntraDir + firstElement, DC_IDX, numElements * sizeof( *m_puhLumaIntraDir ) );
memset( m_puhChromaIntraDir + firstElement, 0, numElements * sizeof( *m_puhChromaIntraDir ) );
memset( m_puhInterDir + firstElement, 0, numElements * sizeof( *m_puhInterDir ) );
memset( m_puhCbf[0] + firstElement, 0, numElements * sizeof( *m_puhCbf[0] ) );
memset( m_puhCbf[1] + firstElement, 0, numElements * sizeof( *m_puhCbf[1] ) );
memset( m_puhCbf[2] + firstElement, 0, numElements * sizeof( *m_puhCbf[2] ) );
memset( m_pbIPCMFlag + firstElement, false, numElements * sizeof( *m_pbIPCMFlag ) );
}
UInt uiTmp = g_uiMaxCUWidth*g_uiMaxCUHeight;
if ( 0 >= partStartIdx )
{
m_acCUMvField[0].clearMvField();
m_acCUMvField[1].clearMvField();
memset( m_pcTrCoeffY , 0, sizeof( TCoeff ) * uiTmp );
#if ADAPTIVE_QP_SELECTION
memset( m_pcArlCoeffY , 0, sizeof( Int ) * uiTmp );
#endif
memset( m_pcIPCMSampleY , 0, sizeof( Pel ) * uiTmp );
uiTmp >>= 2;
memset( m_pcTrCoeffCb, 0, sizeof( TCoeff ) * uiTmp );
memset( m_pcTrCoeffCr, 0, sizeof( TCoeff ) * uiTmp );
#if ADAPTIVE_QP_SELECTION
memset( m_pcArlCoeffCb, 0, sizeof( Int ) * uiTmp );
memset( m_pcArlCoeffCr, 0, sizeof( Int ) * uiTmp );
#endif
memset( m_pcIPCMSampleCb , 0, sizeof( Pel ) * uiTmp );
memset( m_pcIPCMSampleCr , 0, sizeof( Pel ) * uiTmp );
}
else
{
TComDataCU * pcFrom = pcPic->getCU(getAddr());
m_acCUMvField[0].copyFrom(&pcFrom->m_acCUMvField[0],m_uiNumPartition,0);
m_acCUMvField[1].copyFrom(&pcFrom->m_acCUMvField[1],m_uiNumPartition,0);
for(Int i=0; i<uiTmp; i++)
{
m_pcTrCoeffY[i]=pcFrom->m_pcTrCoeffY[i];
#if ADAPTIVE_QP_SELECTION
m_pcArlCoeffY[i]=pcFrom->m_pcArlCoeffY[i];
#endif
m_pcIPCMSampleY[i]=pcFrom->m_pcIPCMSampleY[i];
}
for(Int i=0; i<(uiTmp>>2); i++)
{
m_pcTrCoeffCb[i]=pcFrom->m_pcTrCoeffCb[i];
m_pcTrCoeffCr[i]=pcFrom->m_pcTrCoeffCr[i];
#if ADAPTIVE_QP_SELECTION
m_pcArlCoeffCb[i]=pcFrom->m_pcArlCoeffCb[i];
m_pcArlCoeffCr[i]=pcFrom->m_pcArlCoeffCr[i];
#endif
m_pcIPCMSampleCb[i]=pcFrom->m_pcIPCMSampleCb[i];
m_pcIPCMSampleCr[i]=pcFrom->m_pcIPCMSampleCr[i];
}
}
// Setting neighbor CU
m_pcCULeft = NULL;
m_pcCUAbove = NULL;
m_pcCUAboveLeft = NULL;
m_pcCUAboveRight = NULL;
m_apcCUColocated[0] = NULL;
m_apcCUColocated[1] = NULL;
UInt uiWidthInCU = pcPic->getFrameWidthInCU();
if ( m_uiCUAddr % uiWidthInCU )
{
m_pcCULeft = pcPic->getCU( m_uiCUAddr - 1 );
}
if ( m_uiCUAddr / uiWidthInCU )
{
m_pcCUAbove = pcPic->getCU( m_uiCUAddr - uiWidthInCU );
}
if ( m_pcCULeft && m_pcCUAbove )
{
m_pcCUAboveLeft = pcPic->getCU( m_uiCUAddr - uiWidthInCU - 1 );
}
if ( m_pcCUAbove && ( (m_uiCUAddr%uiWidthInCU) < (uiWidthInCU-1) ) )
{
m_pcCUAboveRight = pcPic->getCU( m_uiCUAddr - uiWidthInCU + 1 );
}
if ( getSlice()->getNumRefIdx( REF_PIC_LIST_0 ) > 0 )
{
m_apcCUColocated[0] = getSlice()->getRefPic( REF_PIC_LIST_0, 0)->getCU( m_uiCUAddr );
}
if ( getSlice()->getNumRefIdx( REF_PIC_LIST_1 ) > 0 )
{
m_apcCUColocated[1] = getSlice()->getRefPic( REF_PIC_LIST_1, 0)->getCU( m_uiCUAddr );
}
}此函数传入的是Pic和CU的参数，此段程序主要完成以下几个任务：
1、Initialize top-level CU
m_pcPic = pcPic; m_pcSlice = pcPic->getSlice(pcPic->getCurrSliceIdx()); m_uiCUAddr = iCUAddr; m_uiCUPelX = ( iCUAddr % pcPic->getFrameWidthInCU() ) * g_uiMaxCUWidth; m_uiCUPelY = ( iCUAddr / pcPic->getFrameWidthInCU() ) * g_uiMaxCUHeight; m_uiAbsIdxInLCU = 0; m_dTotalCost = MAX_DOUBLE; m_uiTotalDistortion = 0; m_uiTotalBits = 0; m_uiTotalBins = 0; m_uiNumPartition = pcPic->getNumPartInCU();
指明了当前CU属于哪一个Pic，属于哪一个Slice以及CU的地址。
2、Set neighbor CU
m_pcCULeft = NULL; m_pcCUAbove = NULL; m_pcCUAboveLeft = NULL; m_pcCUAboveRight = NULL;
左方、上方、左上和右上的位置和值，接下来的也可以参考这一块的值。
三是analysis of CU：
xCompressCU( m_ppcBestCU[0], m_ppcTempCU[0], 0 );
从深度0开始一直往上加，选择最好的预测模式和QP
xCompressCU的函数位于TEncCu::xCompressCU，对应的代码如下：
#if AMP_ENC_SPEEDUP
Void TEncCu::xCompressCU( TComDataCU*& rpcBestCU, TComDataCU*& rpcTempCU, UInt uiDepth, PartSize eParentPartSize )
#else
Void TEncCu::xCompressCU( TComDataCU*& rpcBestCU, TComDataCU*& rpcTempCU, UInt uiDepth )
#endif
{
TComPic* pcPic = rpcBestCU->getPic();
// get Original YUV data from picture
m_ppcOrigYuv[uiDepth]->copyFromPicYuv( pcPic->getPicYuvOrg(), rpcBestCU->getAddr(), rpcBestCU->getZorderIdxInCU() );
// variable for Early CU determination
Bool bSubBranch = true;
// variable for Cbf fast mode PU decision
Bool doNotBlockPu = true;
Bool earlyDetectionSkipMode = false;
Bool bBoundary = false;
UInt uiLPelX = rpcBestCU->getCUPelX();
UInt uiRPelX = uiLPelX + rpcBestCU->getWidth(0) - 1;
UInt uiTPelY = rpcBestCU->getCUPelY();
UInt uiBPelY = uiTPelY + rpcBestCU->getHeight(0) - 1;
Int iBaseQP = xComputeQP( rpcBestCU, uiDepth );
Int iMinQP;
Int iMaxQP;
Bool isAddLowestQP = false;
if( (g_uiMaxCUWidth>>uiDepth) >= rpcTempCU->getSlice()->getPPS()->getMinCuDQPSize() )
{
Int idQP = m_pcEncCfg->getMaxDeltaQP();
iMinQP = Clip3( -rpcTempCU->getSlice()->getSPS()->getQpBDOffsetY(), MAX_QP, iBaseQP-idQP );
iMaxQP = Clip3( -rpcTempCU->getSlice()->getSPS()->getQpBDOffsetY(), MAX_QP, iBaseQP+idQP );
}
else
{
iMinQP = rpcTempCU->getQP(0);
iMaxQP = rpcTempCU->getQP(0);
}
if ( m_pcEncCfg->getUseRateCtrl() )
{
iMinQP = m_pcRateCtrl->getRCQP();
iMaxQP = m_pcRateCtrl->getRCQP();
}
// transquant-bypass (TQB) processing loop variable initialisation ---
const Int lowestQP = iMinQP; // For TQB, use this QP which is the lowest non TQB QP tested (rather than QP'=0) - that way delta QPs are smaller, and TQB can be tested at all CU levels.
if ( (rpcTempCU->getSlice()->getPPS()->getTransquantBypassEnableFlag()) )
{
isAddLowestQP = true; // mark that the first iteration is to cost TQB mode.
iMinQP = iMinQP - 1; // increase loop variable range by 1, to allow testing of TQB mode along with other QPs
if ( m_pcEncCfg->getCUTransquantBypassFlagForceValue() )
{
iMaxQP = iMinQP;
}
}
// If slice start or slice end is within this cu...
TComSlice * pcSlice = rpcTempCU->getPic()->getSlice(rpcTempCU->getPic()->getCurrSliceIdx());
Bool bSliceStart = pcSlice->getSliceSegmentCurStartCUAddr()>rpcTempCU->getSCUAddr()&&pcSlice->getSliceSegmentCurStartCUAddr()<rpcTempCU->getSCUAddr()+rpcTempCU->getTotalNumPart();
Bool bSliceEnd = (pcSlice->getSliceSegmentCurEndCUAddr()>rpcTempCU->getSCUAddr()&&pcSlice->getSliceSegmentCurEndCUAddr()<rpcTempCU->getSCUAddr()+rpcTempCU->getTotalNumPart());
Bool bInsidePicture = ( uiRPelX < rpcBestCU->getSlice()->getSPS()->getPicWidthInLumaSamples() ) && ( uiBPelY < rpcBestCU->getSlice()->getSPS()->getPicHeightInLumaSamples() );
// We need to split, so don't try these modes.
if(!bSliceEnd && !bSliceStart && bInsidePicture )
{
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
const Bool bIsLosslessMode = isAddLowestQP && (iQP == iMinQP);
if (bIsLosslessMode)
{
iQP = lowestQP;
}
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
// do inter modes, SKIP and 2Nx2N
if( rpcBestCU->getSlice()->getSliceType() != I_SLICE )
{
// 2Nx2N
if(m_pcEncCfg->getUseEarlySkipDetection())
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );//by Competition for inter_2Nx2N
}
// SKIP
xCheckRDCostMerge2Nx2N( rpcBestCU, rpcTempCU, &earlyDetectionSkipMode );//by Merge for inter_2Nx2N
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(!m_pcEncCfg->getUseEarlySkipDetection())
{
// 2Nx2N, NxN
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode())
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
}
if (bIsLosslessMode)
{
iQP = iMinQP;
}
}
if(!earlyDetectionSkipMode)
{
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
const Bool bIsLosslessMode = isAddLowestQP && (iQP == iMinQP);
if (bIsLosslessMode)
{
iQP = lowestQP;
}
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
// do inter modes, NxN, 2NxN, and Nx2N
if( rpcBestCU->getSlice()->getSliceType() != I_SLICE )
{
// 2Nx2N, NxN
if(!( (rpcBestCU->getWidth(0)==8) && (rpcBestCU->getHeight(0)==8) ))
{
if( uiDepth == g_uiMaxCUDepth - g_uiAddCUDepth && doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_NxN );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
// 2NxN, Nx2N
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_Nx2N )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter ( rpcBestCU, rpcTempCU, SIZE_2NxN );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxN)
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
#if 1
//! Try AMP (SIZE_2NxnU, SIZE_2NxnD, SIZE_nLx2N, SIZE_nRx2N)
if( pcPic->getSlice(0)->getSPS()->getAMPAcc(uiDepth) )
{
#if AMP_ENC_SPEEDUP
Bool bTestAMP_Hor = false, bTestAMP_Ver = false;
#if AMP_MRG
Bool bTestMergeAMP_Hor = false, bTestMergeAMP_Ver = false;
deriveTestModeAMP (rpcBestCU, eParentPartSize, bTestAMP_Hor, bTestAMP_Ver, bTestMergeAMP_Hor, bTestMergeAMP_Ver);
#else
deriveTestModeAMP (rpcBestCU, eParentPartSize, bTestAMP_Hor, bTestAMP_Ver);
#endif
//! Do horizontal AMP
if ( bTestAMP_Hor )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnU );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnU )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnD );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnD )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
}
#if AMP_MRG
else if ( bTestMergeAMP_Hor )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnU, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnU )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnD, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnD )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
}
#endif
//! Do horizontal AMP
if ( bTestAMP_Ver )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nLx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_nLx2N )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nRx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
#if AMP_MRG
else if ( bTestMergeAMP_Ver )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nLx2N, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_nLx2N )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nRx2N, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
#endif
#else
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnU );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnD );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nLx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nRx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
#endif
}
#endif
}
// do normal intra modes
// speedup for inter frames
if( rpcBestCU->getSlice()->getSliceType() == I_SLICE ||
rpcBestCU->getCbf( 0, TEXT_LUMA ) != 0 ||
rpcBestCU->getCbf( 0, TEXT_CHROMA_U ) != 0 ||
rpcBestCU->getCbf( 0, TEXT_CHROMA_V ) != 0 ) // avoid very complex intra if it is unlikely
{
xCheckRDCostIntra( rpcBestCU, rpcTempCU, SIZE_2Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if( uiDepth == g_uiMaxCUDepth - g_uiAddCUDepth )
{
if( rpcTempCU->getWidth(0) > ( 1 << rpcTempCU->getSlice()->getSPS()->getQuadtreeTULog2MinSize() ) )
{
xCheckRDCostIntra( rpcBestCU, rpcTempCU, SIZE_NxN );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
}
// test PCM
if(pcPic->getSlice(0)->getSPS()->getUsePCM()
&& rpcTempCU->getWidth(0) <= (1<<pcPic->getSlice(0)->getSPS()->getPCMLog2MaxSize())
&& rpcTempCU->getWidth(0) >= (1<<pcPic->getSlice(0)->getSPS()->getPCMLog2MinSize()) )
{
UInt uiRawBits = (2 * g_bitDepthY + g_bitDepthC) * rpcBestCU->getWidth(0) * rpcBestCU->getHeight(0) / 2;
UInt uiBestBits = rpcBestCU->getTotalBits();
if((uiBestBits > uiRawBits) || (rpcBestCU->getTotalCost() > m_pcRdCost->calcRdCost(uiRawBits, 0)))
{
xCheckIntraPCM (rpcBestCU, rpcTempCU);
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
if (bIsLosslessMode)
{
iQP = iMinQP;
}
}
}
m_pcEntropyCoder->resetBits();
m_pcEntropyCoder->encodeSplitFlag( rpcBestCU, 0, uiDepth, true );
rpcBestCU->getTotalBits() += m_pcEntropyCoder->getNumberOfWrittenBits(); // split bits
rpcBestCU->getTotalBins() += ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded();
rpcBestCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcBestCU->getTotalBits(), rpcBestCU->getTotalDistortion() );
// Early CU determination
if( m_pcEncCfg->getUseEarlyCU() && rpcBestCU->isSkipped(0) )
{
bSubBranch = false;
}
else
{
bSubBranch = true;
}
}
else if(!(bSliceEnd && bInsidePicture))
{
bBoundary = true;
}
// copy orginal YUV samples to PCM buffer
if( rpcBestCU->isLosslessCoded(0) && (rpcBestCU->getIPCMFlag(0) == false))
{
xFillPCMBuffer(rpcBestCU, m_ppcOrigYuv[uiDepth]);
}
if( (g_uiMaxCUWidth>>uiDepth) == rpcTempCU->getSlice()->getPPS()->getMinCuDQPSize() )
{
Int idQP = m_pcEncCfg->getMaxDeltaQP();
iMinQP = Clip3( -rpcTempCU->getSlice()->getSPS()->getQpBDOffsetY(), MAX_QP, iBaseQP-idQP );
iMaxQP = Clip3( -rpcTempCU->getSlice()->getSPS()->getQpBDOffsetY(), MAX_QP, iBaseQP+idQP );
}
else if( (g_uiMaxCUWidth>>uiDepth) > rpcTempCU->getSlice()->getPPS()->getMinCuDQPSize() )
{
iMinQP = iBaseQP;
iMaxQP = iBaseQP;
}
else
{
Int iStartQP;
if( pcPic->getCU( rpcTempCU->getAddr() )->getSliceSegmentStartCU(rpcTempCU->getZorderIdxInCU()) == pcSlice->getSliceSegmentCurStartCUAddr())
{
iStartQP = rpcTempCU->getQP(0);
}
else
{
UInt uiCurSliceStartPartIdx = pcSlice->getSliceSegmentCurStartCUAddr() % pcPic->getNumPartInCU() - rpcTempCU->getZorderIdxInCU();
iStartQP = rpcTempCU->getQP(uiCurSliceStartPartIdx);
}
iMinQP = iStartQP;
iMaxQP = iStartQP;
}
if ( m_pcEncCfg->getUseRateCtrl() )
{
iMinQP = m_pcRateCtrl->getRCQP();
iMaxQP = m_pcRateCtrl->getRCQP();
}
if ( m_pcEncCfg->getCUTransquantBypassFlagForceValue() )
{
iMaxQP = iMinQP; // If all blocks are forced into using transquant bypass, do not loop here.
}
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
const Bool bIsLosslessMode = false; // False at this level. Next level down may set it to true.
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
// further split
if( bSubBranch && uiDepth < g_uiMaxCUDepth - g_uiAddCUDepth )
{
UChar uhNextDepth = uiDepth+1;
TComDataCU* pcSubBestPartCU = m_ppcBestCU[uhNextDepth];
TComDataCU* pcSubTempPartCU = m_ppcTempCU[uhNextDepth];
for ( UInt uiPartUnitIdx = 0; uiPartUnitIdx < 4; uiPartUnitIdx++ )
{
pcSubBestPartCU->initSubCU( rpcTempCU, uiPartUnitIdx, uhNextDepth, iQP ); // clear sub partition datas or init.
pcSubTempPartCU->initSubCU( rpcTempCU, uiPartUnitIdx, uhNextDepth, iQP ); // clear sub partition datas or init.
Bool bInSlice = pcSubBestPartCU->getSCUAddr()+pcSubBestPartCU->getTotalNumPart()>pcSlice->getSliceSegmentCurStartCUAddr()&&pcSubBestPartCU->getSCUAddr()<pcSlice->getSliceSegmentCurEndCUAddr();
if(bInSlice && ( pcSubBestPartCU->getCUPelX() < pcSlice->getSPS()->getPicWidthInLumaSamples() ) && ( pcSubBestPartCU->getCUPelY() < pcSlice->getSPS()->getPicHeightInLumaSamples() ) )
{
if ( 0 == uiPartUnitIdx) //initialize RD with previous depth buffer
{
m_pppcRDSbacCoder[uhNextDepth][CI_CURR_BEST]->load(m_pppcRDSbacCoder[uiDepth][CI_CURR_BEST]);
}
else
{
m_pppcRDSbacCoder[uhNextDepth][CI_CURR_BEST]->load(m_pppcRDSbacCoder[uhNextDepth][CI_NEXT_BEST]);
}
#if AMP_ENC_SPEEDUP
if ( rpcBestCU->isIntra(0) )
{
xCompressCU( pcSubBestPartCU, pcSubTempPartCU, uhNextDepth, SIZE_NONE );
}
else
{
xCompressCU( pcSubBestPartCU, pcSubTempPartCU, uhNextDepth, rpcBestCU->getPartitionSize(0) );
}
#else
xCompressCU( pcSubBestPartCU, pcSubTempPartCU, uhNextDepth );
#endif
rpcTempCU->copyPartFrom( pcSubBestPartCU, uiPartUnitIdx, uhNextDepth ); // Keep best part data to current temporary data.
xCopyYuv2Tmp( pcSubBestPartCU->getTotalNumPart()*uiPartUnitIdx, uhNextDepth );
}
else if (bInSlice)
{
pcSubBestPartCU->copyToPic( uhNextDepth );
rpcTempCU->copyPartFrom( pcSubBestPartCU, uiPartUnitIdx, uhNextDepth );
}
}
if( !bBoundary )
{
m_pcEntropyCoder->resetBits();
m_pcEntropyCoder->encodeSplitFlag( rpcTempCU, 0, uiDepth, true );
rpcTempCU->getTotalBits() += m_pcEntropyCoder->getNumberOfWrittenBits(); // split bits
rpcTempCU->getTotalBins() += ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded();
}
rpcTempCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcTempCU->getTotalBits(), rpcTempCU->getTotalDistortion() );
if( (g_uiMaxCUWidth>>uiDepth) == rpcTempCU->getSlice()->getPPS()->getMinCuDQPSize() && rpcTempCU->getSlice()->getPPS()->getUseDQP())
{
Bool hasResidual = false;
for( UInt uiBlkIdx = 0; uiBlkIdx < rpcTempCU->getTotalNumPart(); uiBlkIdx ++)
{
if( ( pcPic->getCU( rpcTempCU->getAddr() )->getSliceSegmentStartCU(uiBlkIdx+rpcTempCU->getZorderIdxInCU()) == rpcTempCU->getSlice()->getSliceSegmentCurStartCUAddr() ) &&
( rpcTempCU->getCbf( uiBlkIdx, TEXT_LUMA ) || rpcTempCU->getCbf( uiBlkIdx, TEXT_CHROMA_U ) || rpcTempCU->getCbf( uiBlkIdx, TEXT_CHROMA_V ) ) )
{
hasResidual = true;
break;
}
}
UInt uiTargetPartIdx;
if ( pcPic->getCU( rpcTempCU->getAddr() )->getSliceSegmentStartCU(rpcTempCU->getZorderIdxInCU()) != pcSlice->getSliceSegmentCurStartCUAddr() )
{
uiTargetPartIdx = pcSlice->getSliceSegmentCurStartCUAddr() % pcPic->getNumPartInCU() - rpcTempCU->getZorderIdxInCU();
}
else
{
uiTargetPartIdx = 0;
}
if ( hasResidual )
{
#if !RDO_WITHOUT_DQP_BITS
m_pcEntropyCoder->resetBits();
m_pcEntropyCoder->encodeQP( rpcTempCU, uiTargetPartIdx, false );
rpcTempCU->getTotalBits() += m_pcEntropyCoder->getNumberOfWrittenBits(); // dQP bits
rpcTempCU->getTotalBins() += ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded();
rpcTempCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcTempCU->getTotalBits(), rpcTempCU->getTotalDistortion() );
#endif
Bool foundNonZeroCbf = false;
rpcTempCU->setQPSubCUs( rpcTempCU->getRefQP( uiTargetPartIdx ), rpcTempCU, 0, uiDepth, foundNonZeroCbf );
assert( foundNonZeroCbf );
}
else
{
rpcTempCU->setQPSubParts( rpcTempCU->getRefQP( uiTargetPartIdx ), 0, uiDepth ); // set QP to default QP
}
}
m_pppcRDSbacCoder[uhNextDepth][CI_NEXT_BEST]->store(m_pppcRDSbacCoder[uiDepth][CI_TEMP_BEST]);
Bool isEndOfSlice = rpcBestCU->getSlice()->getSliceMode()==FIXED_NUMBER_OF_BYTES
&& (rpcBestCU->getTotalBits()>rpcBestCU->getSlice()->getSliceArgument()<<3);
Bool isEndOfSliceSegment = rpcBestCU->getSlice()->getSliceSegmentMode()==FIXED_NUMBER_OF_BYTES
&& (rpcBestCU->getTotalBits()>rpcBestCU->getSlice()->getSliceSegmentArgument()<<3);
if(isEndOfSlice||isEndOfSliceSegment)
{
rpcBestCU->getTotalCost()=rpcTempCU->getTotalCost()+1;
}
xCheckBestMode( rpcBestCU, rpcTempCU, uiDepth); // RD compare current larger prediction
} // with sub partitioned prediction.
}
rpcBestCU->copyToPic(uiDepth); // Copy Best data to Picture for next partition prediction.
xCopyYuv2Pic( rpcBestCU->getPic(), rpcBestCU->getAddr(), rpcBestCU->getZorderIdxInCU(), uiDepth, uiDepth, rpcBestCU, uiLPelX, uiTPelY ); // Copy Yuv data to picture Yuv
if( bBoundary ||(bSliceEnd && bInsidePicture))
{
return;
}
// Assert if Best prediction mode is NONE
// Selected mode's RD-cost must be not MAX_DOUBLE.
assert( rpcBestCU->getPartitionSize ( 0 ) != SIZE_NONE );
assert( rpcBestCU->getPredictionMode( 0 ) != MODE_NONE );
assert( rpcBestCU->getTotalCost ( ) != MAX_DOUBLE );
}
xCompressCU是主要做切割的地方，传入的参数是BestCU，TempCU和Depth，主要完成以下几个任务：
1、Get picture YUV data
TComPic* pcPic = rpcBestCU->getPic(); // get Original YUV data from picture m_ppcOrigYuv[uiDepth]->copyFromPicYuv( pcPic->getPicYuvOrg(), rpcBestCU->getAddr(), rpcBestCU->getZorderIdxInCU() );
2、Compute BaseQP
通过iBaseQP计算iMinQP和iMaxQP，对应的函数是：
Int iBaseQP = xComputeQP( rpcBestCU, uiDepth );
3、Check every possible QP
在一个for循环中完成，主要功能为：尝试每一个可能的QP对应的每一种预测模式，得到QP和预测模式。
其基本模型为：
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
try all kinds of prediction modes for every possible QP
}
对应的代码如下：
// If slice start or slice end is within this cu...
TComSlice * pcSlice = rpcTempCU->getPic()->getSlice(rpcTempCU->getPic()->getCurrSliceIdx());
Bool bSliceStart = pcSlice->getSliceSegmentCurStartCUAddr()>rpcTempCU->getSCUAddr()&&pcSlice->getSliceSegmentCurStartCUAddr()<rpcTempCU->getSCUAddr()+rpcTempCU->getTotalNumPart();
Bool bSliceEnd = (pcSlice->getSliceSegmentCurEndCUAddr()>rpcTempCU->getSCUAddr()&&pcSlice->getSliceSegmentCurEndCUAddr()<rpcTempCU->getSCUAddr()+rpcTempCU->getTotalNumPart());
Bool bInsidePicture = ( uiRPelX < rpcBestCU->getSlice()->getSPS()->getPicWidthInLumaSamples() ) && ( uiBPelY < rpcBestCU->getSlice()->getSPS()->getPicHeightInLumaSamples() );
// We need to split, so don't try these modes.
<span style="color:#cc0000;"> <strong>if(!bSliceEnd && !bSliceStart && bInsidePicture )</strong></span>
{
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
const Bool bIsLosslessMode = isAddLowestQP && (iQP == iMinQP);
if (bIsLosslessMode)
{
iQP = lowestQP;
}
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
// do inter modes, SKIP and 2Nx2N
if( rpcBestCU->getSlice()->getSliceType() != I_SLICE )
{
// 2Nx2N
if(m_pcEncCfg->getUseEarlySkipDetection())
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );//by Competition for inter_2Nx2N
}
// SKIP
xCheckRDCostMerge2Nx2N( rpcBestCU, rpcTempCU, &earlyDetectionSkipMode );//by Merge for inter_2Nx2N
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(!m_pcEncCfg->getUseEarlySkipDetection())
{
// 2Nx2N, NxN
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode())
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
}
if (bIsLosslessMode)
{
iQP = iMinQP;
}
}
if(!earlyDetectionSkipMode)
{
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
const Bool bIsLosslessMode = isAddLowestQP && (iQP == iMinQP);
if (bIsLosslessMode)
{
iQP = lowestQP;
}
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
// do inter modes, NxN, 2NxN, and Nx2N
if( rpcBestCU->getSlice()->getSliceType() != I_SLICE )
{
// 2Nx2N, NxN
if(!( (rpcBestCU->getWidth(0)==8) && (rpcBestCU->getHeight(0)==8) ))
{
if( uiDepth == g_uiMaxCUDepth - g_uiAddCUDepth && doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_NxN );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
// 2NxN, Nx2N
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_Nx2N )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter ( rpcBestCU, rpcTempCU, SIZE_2NxN );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxN)
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
#if 1
//! Try AMP (SIZE_2NxnU, SIZE_2NxnD, SIZE_nLx2N, SIZE_nRx2N)
if( pcPic->getSlice(0)->getSPS()->getAMPAcc(uiDepth) )
{
#if AMP_ENC_SPEEDUP
Bool bTestAMP_Hor = false, bTestAMP_Ver = false;
#if AMP_MRG
Bool bTestMergeAMP_Hor = false, bTestMergeAMP_Ver = false;
deriveTestModeAMP (rpcBestCU, eParentPartSize, bTestAMP_Hor, bTestAMP_Ver, bTestMergeAMP_Hor, bTestMergeAMP_Ver);
#else
deriveTestModeAMP (rpcBestCU, eParentPartSize, bTestAMP_Hor, bTestAMP_Ver);
#endif
//! Do horizontal AMP
if ( bTestAMP_Hor )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnU );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnU )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnD );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnD )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
}
#if AMP_MRG
else if ( bTestMergeAMP_Hor )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnU, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnU )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnD, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_2NxnD )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
}
#endif
//! Do horizontal AMP
if ( bTestAMP_Ver )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nLx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_nLx2N )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nRx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
#if AMP_MRG
else if ( bTestMergeAMP_Ver )
{
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nLx2N, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if(m_pcEncCfg->getUseCbfFastMode() && rpcBestCU->getPartitionSize(0) == SIZE_nLx2N )
{
doNotBlockPu = rpcBestCU->getQtRootCbf( 0 ) != 0;
}
}
if(doNotBlockPu)
{
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nRx2N, true );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
#endif
#else
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnU );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_2NxnD );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nLx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
xCheckRDCostInter( rpcBestCU, rpcTempCU, SIZE_nRx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
#endif
}
#endif
}
// do normal intra modes
// speedup for inter frames
if( rpcBestCU->getSlice()->getSliceType() == I_SLICE ||
rpcBestCU->getCbf( 0, TEXT_LUMA ) != 0 ||
rpcBestCU->getCbf( 0, TEXT_CHROMA_U ) != 0 ||
rpcBestCU->getCbf( 0, TEXT_CHROMA_V ) != 0 ) // avoid very complex intra if it is unlikely
{
xCheckRDCostIntra( rpcBestCU, rpcTempCU, SIZE_2Nx2N );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
if( uiDepth == g_uiMaxCUDepth - g_uiAddCUDepth )
{
if( rpcTempCU->getWidth(0) > ( 1 << rpcTempCU->getSlice()->getSPS()->getQuadtreeTULog2MinSize() ) )
{
xCheckRDCostIntra( rpcBestCU, rpcTempCU, SIZE_NxN );
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
}
// test PCM
if(pcPic->getSlice(0)->getSPS()->getUsePCM()
&& rpcTempCU->getWidth(0) <= (1<<pcPic->getSlice(0)->getSPS()->getPCMLog2MaxSize())
&& rpcTempCU->getWidth(0) >= (1<<pcPic->getSlice(0)->getSPS()->getPCMLog2MinSize()) )
{
UInt uiRawBits = (2 * g_bitDepthY + g_bitDepthC) * rpcBestCU->getWidth(0) * rpcBestCU->getHeight(0) / 2;
UInt uiBestBits = rpcBestCU->getTotalBits();
if((uiBestBits > uiRawBits) || (rpcBestCU->getTotalCost() > m_pcRdCost->calcRdCost(uiRawBits, 0)))
{
xCheckIntraPCM (rpcBestCU, rpcTempCU);
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
}
}
if (bIsLosslessMode)
{
iQP = iMinQP;
}
}
}
m_pcEntropyCoder->resetBits();
m_pcEntropyCoder->encodeSplitFlag( rpcBestCU, 0, uiDepth, true );
rpcBestCU->getTotalBits() += m_pcEntropyCoder->getNumberOfWrittenBits(); // split bits
rpcBestCU->getTotalBins() += ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded();
rpcBestCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcBestCU->getTotalBits(), rpcBestCU->getTotalDistortion() );
// Early CU determination
if( m_pcEncCfg->getUseEarlyCU() && rpcBestCU->isSkipped(0) )
{
bSubBranch = false;
}
else
{
bSubBranch = true;
}
}
<span style="color:#cc0000;"> <strong>else if(!(bSliceEnd && bInsidePicture))</strong></span>
{
bBoundary = true;
}
4、Check if Slice starts or Slice ends within this CU
对应的函数功能代码如下：
Bool bInsidePicture = ( uiRPelX < rpcBestCU->getSlice()->getSPS()->getPicWidthInLumaSamples() ) && ( uiBPelY < rpcBestCU->getSlice()->getSPS()->getPicHeightInLumaSamples() );
True：try every prediction mode
False：further split
//try every prediction mode
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
try all kinds of prediction modes for every possible QP
}
//further split在此函数的功能：即检查当前CU是否在当前Slice的头或尾，若没有跨越Slice，则尝试每一种预测模式；若跨越Slice，则进一步分割。
对应的代码如下：
for (Int iQP=iMinQP; iQP<=iMaxQP; iQP++)
{
const Bool bIsLosslessMode = false; // False at this level. Next level down may set it to true.
rpcTempCU->initEstData( uiDepth, iQP, bIsLosslessMode );
// further split
if( bSubBranch && uiDepth < g_uiMaxCUDepth - g_uiAddCUDepth )
{
UChar uhNextDepth = uiDepth+1;
TComDataCU* pcSubBestPartCU = m_ppcBestCU[uhNextDepth];
TComDataCU* pcSubTempPartCU = m_ppcTempCU[uhNextDepth];
for ( UInt uiPartUnitIdx = 0; uiPartUnitIdx < 4; uiPartUnitIdx++ )
{
pcSubBestPartCU->initSubCU( rpcTempCU, uiPartUnitIdx, uhNextDepth, iQP ); // clear sub partition datas or init.
pcSubTempPartCU->initSubCU( rpcTempCU, uiPartUnitIdx, uhNextDepth, iQP ); // clear sub partition datas or init.
Bool bInSlice = pcSubBestPartCU->getSCUAddr()+pcSubBestPartCU->getTotalNumPart()>pcSlice->getSliceSegmentCurStartCUAddr()&&pcSubBestPartCU->getSCUAddr()<pcSlice->getSliceSegmentCurEndCUAddr();
if(bInSlice && ( pcSubBestPartCU->getCUPelX() < pcSlice->getSPS()->getPicWidthInLumaSamples() ) && ( pcSubBestPartCU->getCUPelY() < pcSlice->getSPS()->getPicHeightInLumaSamples() ) )
{
if ( 0 == uiPartUnitIdx) //initialize RD with previous depth buffer
{
m_pppcRDSbacCoder[uhNextDepth][CI_CURR_BEST]->load(m_pppcRDSbacCoder[uiDepth][CI_CURR_BEST]);
}
else
{
m_pppcRDSbacCoder[uhNextDepth][CI_CURR_BEST]->load(m_pppcRDSbacCoder[uhNextDepth][CI_NEXT_BEST]);
}
#if AMP_ENC_SPEEDUP
if ( rpcBestCU->isIntra(0) )
{
xCompressCU( pcSubBestPartCU, pcSubTempPartCU, uhNextDepth, SIZE_NONE );
}
else
{
xCompressCU( pcSubBestPartCU, pcSubTempPartCU, uhNextDepth, rpcBestCU->getPartitionSize(0) );
}
#else
xCompressCU( pcSubBestPartCU, pcSubTempPartCU, uhNextDepth );
#endif
rpcTempCU->copyPartFrom( pcSubBestPartCU, uiPartUnitIdx, uhNextDepth ); // Keep best part data to current temporary data.
xCopyYuv2Tmp( pcSubBestPartCU->getTotalNumPart()*uiPartUnitIdx, uhNextDepth );
}
else if (bInSlice)
{
pcSubBestPartCU->copyToPic( uhNextDepth );
rpcTempCU->copyPartFrom( pcSubBestPartCU, uiPartUnitIdx, uhNextDepth );
}
}
if( !bBoundary )
{
m_pcEntropyCoder->resetBits();
m_pcEntropyCoder->encodeSplitFlag( rpcTempCU, 0, uiDepth, true );
rpcTempCU->getTotalBits() += m_pcEntropyCoder->getNumberOfWrittenBits(); // split bits
rpcTempCU->getTotalBins() += ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded();
}
rpcTempCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcTempCU->getTotalBits(), rpcTempCU->getTotalDistortion() );
if( (g_uiMaxCUWidth>>uiDepth) == rpcTempCU->getSlice()->getPPS()->getMinCuDQPSize() && rpcTempCU->getSlice()->getPPS()->getUseDQP())
{
Bool hasResidual = false;
for( UInt uiBlkIdx = 0; uiBlkIdx < rpcTempCU->getTotalNumPart(); uiBlkIdx ++)
{
if( ( pcPic->getCU( rpcTempCU->getAddr() )->getSliceSegmentStartCU(uiBlkIdx+rpcTempCU->getZorderIdxInCU()) == rpcTempCU->getSlice()->getSliceSegmentCurStartCUAddr() ) &&
( rpcTempCU->getCbf( uiBlkIdx, TEXT_LUMA ) || rpcTempCU->getCbf( uiBlkIdx, TEXT_CHROMA_U ) || rpcTempCU->getCbf( uiBlkIdx, TEXT_CHROMA_V ) ) )
{
hasResidual = true;
break;
}
}
UInt uiTargetPartIdx;
if ( pcPic->getCU( rpcTempCU->getAddr() )->getSliceSegmentStartCU(rpcTempCU->getZorderIdxInCU()) != pcSlice->getSliceSegmentCurStartCUAddr() )
{
uiTargetPartIdx = pcSlice->getSliceSegmentCurStartCUAddr() % pcPic->getNumPartInCU() - rpcTempCU->getZorderIdxInCU();
}
else
{
uiTargetPartIdx = 0;
}
if ( hasResidual )
{
#if !RDO_WITHOUT_DQP_BITS
m_pcEntropyCoder->resetBits();
m_pcEntropyCoder->encodeQP( rpcTempCU, uiTargetPartIdx, false );
rpcTempCU->getTotalBits() += m_pcEntropyCoder->getNumberOfWrittenBits(); // dQP bits
rpcTempCU->getTotalBins() += ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded();
rpcTempCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcTempCU->getTotalBits(), rpcTempCU->getTotalDistortion() );
#endif
Bool foundNonZeroCbf = false;
rpcTempCU->setQPSubCUs( rpcTempCU->getRefQP( uiTargetPartIdx ), rpcTempCU, 0, uiDepth, foundNonZeroCbf );
assert( foundNonZeroCbf );
}
else
{
rpcTempCU->setQPSubParts( rpcTempCU->getRefQP( uiTargetPartIdx ), 0, uiDepth ); // set QP to default QP
}
}
m_pppcRDSbacCoder[uhNextDepth][CI_NEXT_BEST]->store(m_pppcRDSbacCoder[uiDepth][CI_TEMP_BEST]);
Bool isEndOfSlice = rpcBestCU->getSlice()->getSliceMode()==FIXED_NUMBER_OF_BYTES
&& (rpcBestCU->getTotalBits()>rpcBestCU->getSlice()->getSliceArgument()<<3);
Bool isEndOfSliceSegment = rpcBestCU->getSlice()->getSliceSegmentMode()==FIXED_NUMBER_OF_BYTES
&& (rpcBestCU->getTotalBits()>rpcBestCU->getSlice()->getSliceSegmentArgument()<<3);
if(isEndOfSlice||isEndOfSliceSegment)
{
rpcBestCU->getTotalCost()=rpcTempCU->getTotalCost()+1;
}
xCheckBestMode( rpcBestCU, rpcTempCU, uiDepth); // RD compare current larger prediction
} // with sub partitioned prediction.
}
5、xCompressCU's Structure
