HM16.0之帧内模式——xCheckRDCostIntra()函数
参考：https://blog.csdn.net/nb_vol_1/article/category/6179825/1?
1、源代码：
Void TEncCu::xCheckRDCostIntra( TComDataCU *&rpcBestCU, TComDataCU *&rpcTempCU, Double &cost, PartSize eSize DEBUG_STRING_FN_DECLARE(sDebug) ) { DEBUG_STRING_NEW(sTest) UInt uiDepth = rpcTempCU->getDepth( 0 ); rpcTempCU->setSkipFlagSubParts( false, 0, uiDepth ); rpcTempCU->setPartSizeSubParts( eSize, 0, uiDepth ); rpcTempCU->setPredModeSubParts( MODE_INTRA, 0, uiDepth ); rpcTempCU->setChromaQpAdjSubParts( rpcTempCU->getCUTransquantBypass(0) ? 0 : m_ChromaQpAdjIdc, 0, uiDepth ); Bool bSeparateLumaChroma = true; // choose estimation mode Distortion uiPreCalcDistC = 0; if (rpcBestCU->getPic()->getChromaFormat()==CHROMA_400) { bSeparateLumaChroma=true; } Pel resiLuma[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE]; if( !bSeparateLumaChroma ) { // after this function, the direction will be PLANAR, DC, HOR or VER // however, if Luma ends up being one of those, the chroma dir must be later changed to DM_CHROMA. m_pcPredSearch->preestChromaPredMode( rpcTempCU, m_ppcOrigYuv[uiDepth], m_ppcPredYuvTemp[uiDepth] ); }
// 亮度块的帧内预测 m_pcPredSearch->estIntraPredQT( rpcTempCU, m_ppcOrigYuv[uiDepth], m_ppcPredYuvTemp[uiDepth], m_ppcResiYuvTemp[uiDepth], m_ppcRecoYuvTemp[uiDepth], resiLuma, uiPreCalcDistC, bSeparateLumaChroma DEBUG_STRING_PASS_INTO(sTest) ); m_ppcRecoYuvTemp[uiDepth]->copyToPicComponent(COMPONENT_Y, rpcTempCU->getPic()->getPicYuvRec(), rpcTempCU->getAddr(), rpcTempCU->getZorderIdxInCU() ); if (rpcBestCU->getPic()->getChromaFormat()!=CHROMA_400) {
// 色度块的帧内预测 m_pcPredSearch->estIntraPredChromaQT( rpcTempCU, m_ppcOrigYuv[uiDepth], m_ppcPredYuvTemp[uiDepth], m_ppcResiYuvTemp[uiDepth], m_ppcRecoYuvTemp[uiDepth], resiLuma, uiPreCalcDistC DEBUG_STRING_PASS_INTO(sTest) ); } m_pcEntropyCoder->resetBits(); if ( rpcTempCU->getSlice()->getPPS()->getTransquantBypassEnableFlag()) { m_pcEntropyCoder->encodeCUTransquantBypassFlag( rpcTempCU, 0, true ); } m_pcEntropyCoder->encodeSkipFlag ( rpcTempCU, 0, true ); m_pcEntropyCoder->encodePredMode( rpcTempCU, 0, true ); m_pcEntropyCoder->encodePartSize( rpcTempCU, 0, uiDepth, true ); m_pcEntropyCoder->encodePredInfo( rpcTempCU, 0 ); m_pcEntropyCoder->encodeIPCMInfo(rpcTempCU, 0, true ); // Encode Coefficients Bool bCodeDQP = getdQPFlag(); Bool codeChromaQpAdjFlag = getCodeChromaQpAdjFlag(); m_pcEntropyCoder->encodeCoeff( rpcTempCU, 0, uiDepth, bCodeDQP, codeChromaQpAdjFlag ); setCodeChromaQpAdjFlag( codeChromaQpAdjFlag ); setdQPFlag( bCodeDQP ); m_pcRDGoOnSbacCoder->store(m_pppcRDSbacCoder[uiDepth][CI_TEMP_BEST]); rpcTempCU->getTotalBits() = m_pcEntropyCoder->getNumberOfWrittenBits(); rpcTempCU->getTotalBins() = ((TEncBinCABAC *)((TEncSbac*)m_pcEntropyCoder->m_pcEntropyCoderIf)->getEncBinIf())->getBinsCoded(); rpcTempCU->getTotalCost() = m_pcRdCost->calcRdCost( rpcTempCU->getTotalBits(), rpcTempCU->getTotalDistortion() ); xCheckDQP( rpcTempCU ); cost = rpcTempCU->getTotalCost(); xCheckBestMode(rpcBestCU, rpcTempCU, uiDepth DEBUG_STRING_PASS_INTO(sDebug) DEBUG_STRING_PASS_INTO(sTest)); }
2、estIntraPredQT（亮度块的帧内预测）：
Void TEncSearch::estIntraPredQT(TComDataCU* pcCU, TComYuv* pcOrgYuv, TComYuv* pcPredYuv, TComYuv* pcResiYuv, TComYuv* pcRecoYuv, Pel resiLuma[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE], Distortion& ruiDistC, Bool bLumaOnly DEBUG_STRING_FN_DECLARE(sDebug)) { const UInt uiDepth = pcCU->getDepth(0); const UInt uiInitTrDepth = pcCU->getPartitionSize(0) == SIZE_2Nx2N ? 0 : 1; const UInt uiInitTrDepthC = pcCU->getPartitionSize(0) != SIZE_2Nx2N && enable4ChromaPUsInIntraNxNCU(pcOrgYuv->getChromaFormat()) ? 1 : 0; const UInt uiNumPU = 1<<(2*uiInitTrDepth); const UInt uiQNumParts = pcCU->getTotalNumPart() >> 2; const UInt uiWidthBit = pcCU->getIntraSizeIdx(0); const ChromaFormat chFmt = pcCU->getPic()->getChromaFormat(); const UInt numberValidComponents = getNumberValidComponents(chFmt); Distortion uiOverallDistY = 0; Distortion uiOverallDistC = 0; UInt CandNum; Double CandCostList[ FAST_UDI_MAX_RDMODE_NUM ]; // 候选的Cost列表 Pel resiLumaPU[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE]; Bool bMaintainResidual[NUMBER_OF_STORED_RESIDUAL_TYPES]; for (UInt residualTypeIndex = 0; residualTypeIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; residualTypeIndex++) { bMaintainResidual[residualTypeIndex] = true; //assume true unless specified otherwise } bMaintainResidual[RESIDUAL_ENCODER_SIDE] = !(m_pcEncCfg->getUseReconBasedCrossCPredictionEstimate()); //NOTE: RExt - Lambda calculation at equivalent Qp of 4 is recommended because at that Qp, the quantisation divisor is 1. #if FULL_NBIT const Double sqrtLambdaForFirstPass= (m_pcEncCfg->getCostMode()==COST_MIXED_LOSSLESS_LOSSY_CODING && pcCU->getCUTransquantBypass(0)) ? sqrt(0.57 * pow(2.0, ((RExt__LOSSLESS_AND_MIXED_LOSSLESS_RD_COST_TEST_QP_PRIME - 12) / 3.0))) : m_pcRdCost->getSqrtLambda(); #else const Double sqrtLambdaForFirstPass= (m_pcEncCfg->getCostMode()==COST_MIXED_LOSSLESS_LOSSY_CODING && pcCU->getCUTransquantBypass(0)) ? sqrt(0.57 * pow(2.0, ((RExt__LOSSLESS_AND_MIXED_LOSSLESS_RD_COST_TEST_QP_PRIME - 12 - 6 * (g_bitDepth[CHANNEL_TYPE_LUMA] - 8)) / 3.0))) : m_pcRdCost->getSqrtLambda(); #endif //===== set QP and clear Cbf =====
// 设置QP参数，清理Cbf
if ( pcCU->getSlice()->getPPS()->getUseDQP() == true) { pcCU->setQPSubParts( pcCU->getQP(0), 0, uiDepth ); } else {
// 进入此处 pcCU->setQPSubParts( pcCU->getSlice()->getSliceQp(), 0, uiDepth ); } //===== loop over partitions ===== TComTURecurse tuRecurseCU(pcCU, 0); TComTURecurse tuRecurseWithPU(tuRecurseCU, false, (uiInitTrDepth==0)?TComTU::DONT_SPLIT : TComTU::QUAD_SPLIT); do { const UInt uiPartOffset=tuRecurseWithPU.GetAbsPartIdxTU(); // for( UInt uiPU = 0, uiPartOffset=0; uiPU < uiNumPU; uiPU++, uiPartOffset += uiQNumParts ) //{ //===== init pattern for luma prediction ===== Bool bAboveAvail = false; // 上面的块是否有效 Bool bLeftAvail = false; // 左边的块是否有效 DEBUG_STRING_NEW(sTemp2) //===== determine set of modes to be tested (using prediction signal only) =====
// 35种帧内预测模式 Int numModesAvailable = 35; //total number of Intra modes UInt uiRdModeList[FAST_UDI_MAX_RDMODE_NUM]; Int numModesForFullRD = g_aucIntraModeNumFast[ uiWidthBit ]; if (tuRecurseWithPU.ProcessComponentSection(COMPONENT_Y))
// 使用重建后的YUV图像对当前PU的相邻样点进行滤波，提供参考样本值 initAdiPatternChType( tuRecurseWithPU, bAboveAvail, bLeftAvail, COMPONENT_Y, true DEBUG_STRING_PASS_INTO(sTemp2) ); Bool doFastSearch = (numModesForFullRD != numModesAvailable); if (doFastSearch) { assert(numModesForFullRD < numModesAvailable); for( Int i=0; i < numModesForFullRD; i++ ) {
// 用于存储每一种模式的消耗 CandCostList[ i ] = MAX_DOUBLE; } CandNum = 0; const TComRectangle &puRect=tuRecurseWithPU.getRect(COMPONENT_Y); const UInt uiAbsPartIdx=tuRecurseWithPU.GetAbsPartIdxTU();
// 在原始的YUV中获取亮度的地址 Pel* piOrg = pcOrgYuv ->getAddr( COMPONENT_Y, uiAbsPartIdx );
// 在预测的YUV中获取亮度的地址 Pel* piPred = pcPredYuv->getAddr( COMPONENT_Y, uiAbsPartIdx ); UInt uiStride = pcPredYuv->getStride( COMPONENT_Y ); // 偏移 DistParam distParam; const Bool bUseHadamard=pcCU->getCUTransquantBypass(0) == 0; m_pcRdCost->setDistParam(distParam, g_bitDepth[CHANNEL_TYPE_LUMA], piOrg, uiStride, piPred, uiStride, puRect.width, puRect.height, bUseHadamard); distParam.bApplyWeight = false;
// 遍历35种帧内预测模式，选取若干个代价比较小的模式作为后续处理的模式 for( Int modeIdx = 0; modeIdx < numModesAvailable; modeIdx++ ) { UInt uiMode = modeIdx; Distortion uiSad = 0; const Bool bUseFilter=TComPrediction::filteringIntraReferenceSamples(COMPONENT_Y, uiMode, puRect.width, puRect.height, chFmt, pcCU->getSlice()->getSPS()->getDisableIntraReferenceSmoothing());
// 对亮度块进行预测 predIntraAng( COMPONENT_Y, uiMode, piOrg, uiStride, piPred, uiStride, tuRecurseWithPU, bAboveAvail, bLeftAvail, bUseFilter, TComPrediction::UseDPCMForFirstPassIntraEstimation(tuRecurseWithPU, uiMode) ); // use hadamard transform here
// 使用hadamard变换，计算SATD的值 uiSad+=distParam.DistFunc(&distParam); UInt iModeBits = 0; // NB xModeBitsIntra will not affect the mode for chroma that may have already been pre-estimated. iModeBits+=xModeBitsIntra( pcCU, uiMode, uiPartOffset, uiDepth, uiInitTrDepth, CHANNEL_TYPE_LUMA );
// 计算此种模式的代价 Double cost = (Double)uiSad + (Double)iModeBits * sqrtLambdaForFirstPass; #ifdef DEBUG_INTRA_SEARCH_COSTS std::cout << "1st pass mode " << uiMode << " SAD = " << uiSad << ", mode bits = " << iModeBits << ", cost = " << cost << "\n"; #endif // 更新候选列表 CandNum += xUpdateCandList( uiMode, cost, numModesForFullRD, uiRdModeList, CandCostList ); } #if FAST_UDI_USE_MPM Int uiPreds[NUM_MOST_PROBABLE_MODES] = {-1, -1, -1}; Int iMode = -1;
// 根据相邻块的预测模式来对当前块的模式进行预测，得到若干模式（称为预测模式），存放在uiPreds中 Int numCand = pcCU->getIntraDirPredictor( uiPartOffset, uiPreds, COMPONENT_Y, &iMode ); if( iMode >= 0 ) {
// 将候选列表的索引设置为此模式 numCand = iMode; }
// 遍历预测模式，如果它不在模式候选列表中，那么把它添加到其中 for( Int j=0; j < numCand; j++) { Bool mostProbableModeIncluded = false; Int mostProbableMode = uiPreds[j]; for( Int i=0; i < numModesForFullRD; i++) { mostProbableModeIncluded |= (mostProbableMode == uiRdModeList[i]); } if (!mostProbableModeIncluded) { uiRdModeList[numModesForFullRD++] = mostProbableMode; } } #endif // FAST_UDI_USE_MPM } else { for( Int i=0; i < numModesForFullRD; i++) { uiRdModeList[i] = i; } } //===== check modes (using r-d costs) ===== #if HHI_RQT_INTRA_SPEEDUP_MOD UInt uiSecondBestMode = MAX_UINT; Double dSecondBestPUCost = MAX_DOUBLE; #endif DEBUG_STRING_NEW(sPU) UInt uiBestPUMode = 0; Distortion uiBestPUDistY = 0; Distortion uiBestPUDistC = 0; Double dBestPUCost = MAX_DOUBLE; #if RExt__ENVIRONMENT_VARIABLE_DEBUG_AND_TEST UInt max=numModesForFullRD; if (DebugOptionList::ForceLumaMode.isSet()) max=0; // we are forcing a direction, so don't bother with mode check for ( UInt uiMode = 0; uiMode < max; uiMode++) #else
// 遍历候选集中的模式 for( UInt uiMode = 0; uiMode < numModesForFullRD; uiMode++ ) #endif { // set luma prediction mode UInt uiOrgMode = uiRdModeList[uiMode]; pcCU->setIntraDirSubParts ( CHANNEL_TYPE_LUMA, uiOrgMode, uiPartOffset, uiDepth + uiInitTrDepth ); DEBUG_STRING_NEW(sMode) // set context models
// 设置上下文模型 m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[uiDepth][CI_CURR_BEST] ); // determine residual for partition Distortion uiPUDistY = 0; Distortion uiPUDistC = 0; Double dPUCost = 0.0; #if HHI_RQT_INTRA_SPEEDUP
// 通过多候选模式进行预测、变换、量化等操作来计算代价
// 注意倒数第三个参数bCheckFirst是true，表示会继续按照四叉树的方式向下划分 xRecurIntraCodingQT( bLumaOnly, pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaPU, uiPUDistY, uiPUDistC, true, dPUCost, tuRecurseWithPU DEBUG_STRING_PASS_INTO(sMode) ); #else xRecurIntraCodingQT( bLumaOnly, pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaPU, uiPUDistY, uiPUDistC, dPUCost, tuRecurseWithPU DEBUG_STRING_PASS_INTO(sMode) ); #endif #ifdef DEBUG_INTRA_SEARCH_COSTS std::cout << "2nd pass [luma,chroma] mode [" << Int(pcCU->getIntraDir(CHANNEL_TYPE_LUMA, uiPartOffset)) << "," << Int(pcCU->getIntraDir(CHANNEL_TYPE_CHROMA, uiPartOffset)) << "] cost = " << dPUCost << "\n"; #endif // check r-d cost
// 从候选列表中选取最优的模式 if( dPUCost < dBestPUCost ) { DEBUG_STRING_SWAP(sPU, sMode) #if HHI_RQT_INTRA_SPEEDUP_MOD uiSecondBestMode = uiBestPUMode; dSecondBestPUCost = dBestPUCost; #endif uiBestPUMode = uiOrgMode; uiBestPUDistY = uiPUDistY; uiBestPUDistC = uiPUDistC; dBestPUCost = dPUCost; xSetIntraResultQT( bLumaOnly, pcRecoYuv, tuRecurseWithPU ); if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { const Int xOffset = tuRecurseWithPU.getRect( COMPONENT_Y ).x0; const Int yOffset = tuRecurseWithPU.getRect( COMPONENT_Y ).y0; for (UInt storedResidualIndex = 0; storedResidualIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; storedResidualIndex++) { if (bMaintainResidual[storedResidualIndex]) { xStoreCrossComponentPredictionResult(resiLuma[storedResidualIndex], resiLumaPU[storedResidualIndex], tuRecurseWithPU, xOffset, yOffset, MAX_CU_SIZE, MAX_CU_SIZE ); } } } UInt uiQPartNum = tuRecurseWithPU.GetAbsPartIdxNumParts(); ::memcpy( m_puhQTTempTrIdx, pcCU->getTransformIdx() + uiPartOffset, uiQPartNum * sizeof( UChar ) ); for (UInt component = 0; component < numberValidComponents; component++) { const ComponentID compID = ComponentID(component); ::memcpy( m_puhQTTempCbf[compID], pcCU->getCbf( compID ) + uiPartOffset, uiQPartNum * sizeof( UChar ) ); ::memcpy( m_puhQTTempTransformSkipFlag[compID], pcCU->getTransformSkip(compID) + uiPartOffset, uiQPartNum * sizeof( UChar ) ); } } #if HHI_RQT_INTRA_SPEEDUP_MOD else if( dPUCost < dSecondBestPUCost ) { uiSecondBestMode = uiOrgMode; dSecondBestPUCost = dPUCost; } #endif } // Mode loop #if HHI_RQT_INTRA_SPEEDUP #if HHI_RQT_INTRA_SPEEDUP_MOD for( UInt ui =0; ui < 2; ++ui ) #endif { #if HHI_RQT_INTRA_SPEEDUP_MOD UInt uiOrgMode = ui ? uiSecondBestMode : uiBestPUMode; if( uiOrgMode == MAX_UINT ) { break; } #else UInt uiOrgMode = uiBestPUMode; #endif #if RExt__ENVIRONMENT_VARIABLE_DEBUG_AND_TEST if (DebugOptionList::ForceLumaMode.isSet()) uiOrgMode = DebugOptionList::ForceLumaMode.getInt(); #endif pcCU->setIntraDirSubParts ( CHANNEL_TYPE_LUMA, uiOrgMode, uiPartOffset, uiDepth + uiInitTrDepth ); DEBUG_STRING_NEW(sModeTree) // set context models m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[uiDepth][CI_CURR_BEST] ); // determine residual for partition Distortion uiPUDistY = 0; Distortion uiPUDistC = 0; Double dPUCost = 0.0;
// 使用最优模式对PU进行预测，然后变换量化等，计算代价
// 注意倒数第三个参数bCheckFirst是false，表示当前PU不再进行划分，即只处理当前深度的PU xRecurIntraCodingQT( bLumaOnly, pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaPU, uiPUDistY, uiPUDistC, false, dPUCost, tuRecurseWithPU DEBUG_STRING_PASS_INTO(sModeTree)); // check r-d cost
// 检测同一种模式下，bCheckFirst为true和false的情况下，哪个的代价更低 if( dPUCost < dBestPUCost ) { DEBUG_STRING_SWAP(sPU, sModeTree) uiBestPUMode = uiOrgMode; uiBestPUDistY = uiPUDistY; uiBestPUDistC = uiPUDistC; dBestPUCost = dPUCost; xSetIntraResultQT( bLumaOnly, pcRecoYuv, tuRecurseWithPU ); if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { const Int xOffset = tuRecurseWithPU.getRect( COMPONENT_Y ).x0; const Int yOffset = tuRecurseWithPU.getRect( COMPONENT_Y ).y0; for (UInt storedResidualIndex = 0; storedResidualIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; storedResidualIndex++) { if (bMaintainResidual[storedResidualIndex]) { xStoreCrossComponentPredictionResult(resiLuma[storedResidualIndex], resiLumaPU[storedResidualIndex], tuRecurseWithPU, xOffset, yOffset, MAX_CU_SIZE, MAX_CU_SIZE ); } } } const UInt uiQPartNum = tuRecurseWithPU.GetAbsPartIdxNumParts(); ::memcpy( m_puhQTTempTrIdx, pcCU->getTransformIdx() + uiPartOffset, uiQPartNum * sizeof( UChar ) ); for (UInt component = 0; component < numberValidComponents; component++) { const ComponentID compID = ComponentID(component); ::memcpy( m_puhQTTempCbf[compID], pcCU->getCbf( compID ) + uiPartOffset, uiQPartNum * sizeof( UChar ) ); ::memcpy( m_puhQTTempTransformSkipFlag[compID], pcCU->getTransformSkip(compID) + uiPartOffset, uiQPartNum * sizeof( UChar ) ); } } } // Mode loop #endif DEBUG_STRING_APPEND(sDebug, sPU) //--- update overall distortion --- uiOverallDistY += uiBestPUDistY; uiOverallDistC += uiBestPUDistC; //--- update transform index and cbf --- const UInt uiQPartNum = tuRecurseWithPU.GetAbsPartIdxNumParts(); ::memcpy( pcCU->getTransformIdx() + uiPartOffset, m_puhQTTempTrIdx, uiQPartNum * sizeof( UChar ) ); for (UInt component = 0; component < numberValidComponents; component++) { const ComponentID compID = ComponentID(component); ::memcpy( pcCU->getCbf( compID ) + uiPartOffset, m_puhQTTempCbf[compID], uiQPartNum * sizeof( UChar ) ); ::memcpy( pcCU->getTransformSkip( compID ) + uiPartOffset, m_puhQTTempTransformSkipFlag[compID ], uiQPartNum * sizeof( UChar ) ); } //--- set reconstruction for next intra prediction blocks ---
// 变换量化/反变换反量化都已经处理完成了，那么设置重建块 if( !tuRecurseWithPU.IsLastSection() ) { const Bool bSkipChroma = tuRecurseWithPU.ProcessChannelSection(CHANNEL_TYPE_CHROMA); const UInt numChannelToProcess = (bLumaOnly || bSkipChroma) ? 1 : getNumberValidComponents(pcCU->getPic()->getChromaFormat()); for (UInt ch=0; ch<numChannelToProcess; ch++) { const ComponentID compID = ComponentID(ch); const TComRectangle &puRect=tuRecurseWithPU.getRect(compID); const UInt uiCompWidth = puRect.width; const UInt uiCompHeight = puRect.height; const UInt uiZOrder = pcCU->getZorderIdxInCU() + uiPartOffset; Pel* piDes = pcCU->getPic()->getPicYuvRec()->getAddr( compID, pcCU->getAddr(), uiZOrder ); const UInt uiDesStride = pcCU->getPic()->getPicYuvRec()->getStride( compID); const Pel* piSrc = pcRecoYuv->getAddr( compID, uiPartOffset ); const UInt uiSrcStride = pcRecoYuv->getStride( compID); for( UInt uiY = 0; uiY < uiCompHeight; uiY++, piSrc += uiSrcStride, piDes += uiDesStride ) { for( UInt uiX = 0; uiX < uiCompWidth; uiX++ ) { piDes[ uiX ] = piSrc[ uiX ]; } } } } //=== update PU data ==== pcCU->setIntraDirSubParts ( CHANNEL_TYPE_LUMA, uiBestPUMode, uiPartOffset, uiDepth + uiInitTrDepth ); if (!bLumaOnly && getChromasCorrespondingPULumaIdx(uiPartOffset, chFmt)==uiPartOffset) { UInt chromaDir=pcCU->getIntraDir(CHANNEL_TYPE_CHROMA, getChromasCorrespondingPULumaIdx(uiPartOffset, chFmt)); if (chromaDir == uiBestPUMode && tuRecurseWithPU.ProcessChannelSection(CHANNEL_TYPE_CHROMA)) { pcCU->setIntraDirSubParts ( CHANNEL_TYPE_CHROMA, DM_CHROMA_IDX, getChromasCorrespondingPULumaIdx(uiPartOffset, chFmt), uiDepth + uiInitTrDepthC ); } } //pcCU->copyToPic ( uiDepth, uiPU, uiInitTrDepth ); // Unnecessary copy? } while (tuRecurseWithPU.nextSection(tuRecurseCU)); if( uiNumPU > 1 ) { // set Cbf for all blocks UInt uiCombCbfY = 0; UInt uiCombCbfU = 0; UInt uiCombCbfV = 0; UInt uiPartIdx = 0; for( UInt uiPart = 0; uiPart < 4; uiPart++, uiPartIdx += uiQNumParts ) { uiCombCbfY |= pcCU->getCbf( uiPartIdx, COMPONENT_Y, 1 ); uiCombCbfU |= pcCU->getCbf( uiPartIdx, COMPONENT_Cb, 1 ); uiCombCbfV |= pcCU->getCbf( uiPartIdx, COMPONENT_Cr, 1 ); } for( UInt uiOffs = 0; uiOffs < 4 * uiQNumParts; uiOffs++ ) { pcCU->getCbf( COMPONENT_Y )[ uiOffs ] |= uiCombCbfY; pcCU->getCbf( COMPONENT_Cb )[ uiOffs ] |= uiCombCbfU; pcCU->getCbf( COMPONENT_Cr )[ uiOffs ] |= uiCombCbfV; } } //===== reset context models ===== m_pcRDGoOnSbacCoder->load(m_pppcRDSbacCoder[uiDepth][CI_CURR_BEST]); //===== set distortion (rate and r-d costs are determined later) ===== ruiDistC = uiOverallDistC; pcCU->getTotalDistortion() = uiOverallDistY + uiOverallDistC; }
Void TEncSearch::xRecurIntraCodingQT(Bool bLumaOnly, TComYuv* pcOrgYuv, TComYuv* pcPredYuv, TComYuv* pcResiYuv, Pel resiLuma[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE], Distortion& ruiDistY, Distortion& ruiDistC, #if HHI_RQT_INTRA_SPEEDUP Bool bCheckFirst, #endif Double& dRDCost, TComTU& rTu DEBUG_STRING_FN_DECLARE(sDebug)) { TComDataCU *pcCU = rTu.getCU(); const UInt uiAbsPartIdx = rTu.GetAbsPartIdxTU(); const UInt uiFullDepth = rTu.GetTransformDepthTotal(); const UInt uiTrDepth = rTu.GetTransformDepthRel(); const UInt uiLog2TrSize = rTu.GetLog2LumaTrSize(); Bool bCheckFull = ( uiLog2TrSize <= pcCU->getSlice()->getSPS()->getQuadtreeTULog2MaxSize() ); Bool bCheckSplit = ( uiLog2TrSize > pcCU->getQuadtreeTULog2MinSizeInCU(uiAbsPartIdx) ); const UInt numValidComp = (bLumaOnly) ? 1 : pcOrgYuv->getNumberValidComponents(); Pel resiLumaSplit [NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE]; Pel resiLumaSingle[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE]; Bool bMaintainResidual[NUMBER_OF_STORED_RESIDUAL_TYPES]; for (UInt residualTypeIndex = 0; residualTypeIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; residualTypeIndex++) { bMaintainResidual[residualTypeIndex] = true; //assume true unless specified otherwise } bMaintainResidual[RESIDUAL_ENCODER_SIDE] = !(m_pcEncCfg->getUseReconBasedCrossCPredictionEstimate()); #if HHI_RQT_INTRA_SPEEDUP Int maxTuSize = pcCU->getSlice()->getSPS()->getQuadtreeTULog2MaxSize(); Int isIntraSlice = (pcCU->getSlice()->getSliceType() == I_SLICE); // don't check split if TU size is less or equal to max TU size Bool noSplitIntraMaxTuSize = bCheckFull; if(m_pcEncCfg->getRDpenalty() && ! isIntraSlice) { // in addition don't check split if TU size is less or equal to 16x16 TU size for non-intra slice noSplitIntraMaxTuSize = ( uiLog2TrSize <= min(maxTuSize,4) ); // if maximum RD-penalty don't check TU size 32x32 if(m_pcEncCfg->getRDpenalty()==2) { bCheckFull = ( uiLog2TrSize <= min(maxTuSize,4)); } } if( bCheckFirst && noSplitIntraMaxTuSize ) { bCheckSplit = false; } #else Int maxTuSize = pcCU->getSlice()->getSPS()->getQuadtreeTULog2MaxSize(); Int isIntraSlice = (pcCU->getSlice()->getSliceType() == I_SLICE); // if maximum RD-penalty don't check TU size 32x32 if((m_pcEncCfg->getRDpenalty()==2) && !isIntraSlice) { bCheckFull = ( uiLog2TrSize <= min(maxTuSize,4)); } #endif Double dSingleCost = MAX_DOUBLE; Distortion uiSingleDist[MAX_NUM_CHANNEL_TYPE] = {0,0}; UInt uiSingleCbf[MAX_NUM_COMPONENT] = {0,0,0}; Bool checkTransformSkip = pcCU->getSlice()->getPPS()->getUseTransformSkip(); Int bestModeId[MAX_NUM_COMPONENT] = { 0, 0, 0}; checkTransformSkip &= TUCompRectHasAssociatedTransformSkipFlag(rTu.getRect(COMPONENT_Y), pcCU->getSlice()->getPPS()->getTransformSkipLog2MaxSize()); checkTransformSkip &= (!pcCU->getCUTransquantBypass(0)); if ( m_pcEncCfg->getUseTransformSkipFast() ) { checkTransformSkip &= (pcCU->getPartitionSize(uiAbsPartIdx)==SIZE_NxN); } if( bCheckFull ) {
// TransformSkip模式为true，表示将会跳过变换步骤 if(checkTransformSkip == true) { //----- store original entropy coding status ----- m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); Distortion singleDistTmp[MAX_NUM_CHANNEL_TYPE] = { 0, 0 }; UInt singleCbfTmp[MAX_NUM_COMPONENT] = { 0, 0, 0 }; Double singleCostTmp = 0; Int firstCheckId = 0;
// 遍历两次是为了选取最优模式，modeId能够决定xIntraCodingTUBlock的最后一个参数，该参数控制了预测像素如何生成 for(Int modeId = firstCheckId; modeId < 2; modeId ++) { DEBUG_STRING_NEW(sModeString) Int default0Save1Load2 = 0; singleDistTmp[0]=singleDistTmp[1]=0; if(modeId == firstCheckId) { default0Save1Load2 = 1; } else { default0Save1Load2 = 2; } for(UInt ch=COMPONENT_Y; ch<numValidComp; ch++) { const ComponentID compID = ComponentID(ch); if (rTu.ProcessComponentSection(compID)) { const UInt totalAdjustedDepthChan = rTu.GetTransformDepthTotalAdj(compID); pcCU->setTransformSkipSubParts ( modeId, compID, uiAbsPartIdx, totalAdjustedDepthChan );
// 亮度块的预测和量化 xIntraCodingTUBlock( pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaSingle, false, singleDistTmp[toChannelType(compID)], compID, rTu DEBUG_STRING_PASS_INTO(sModeString), default0Save1Load2 ); } singleCbfTmp[compID] = pcCU->getCbf( uiAbsPartIdx, compID, uiTrDepth ); } //----- determine rate and r-d cost ----- if(modeId == 1 && singleCbfTmp[COMPONENT_Y] == 0) { //In order not to code TS flag when cbf is zero, the case for TS with cbf being zero is forbidden. singleCostTmp = MAX_DOUBLE; } else { UInt uiSingleBits = xGetIntraBitsQT( rTu, true, !bLumaOnly, false ); singleCostTmp = m_pcRdCost->calcRdCost( uiSingleBits, singleDistTmp[CHANNEL_TYPE_LUMA] + singleDistTmp[CHANNEL_TYPE_CHROMA] ); }
// 代价更新 if(singleCostTmp < dSingleCost) { DEBUG_STRING_SWAP(sDebug, sModeString) dSingleCost = singleCostTmp; uiSingleDist[CHANNEL_TYPE_LUMA] = singleDistTmp[CHANNEL_TYPE_LUMA]; uiSingleDist[CHANNEL_TYPE_CHROMA] = singleDistTmp[CHANNEL_TYPE_CHROMA]; for (UInt ch=0; ch<MAX_NUM_COMPONENT; ch++) uiSingleCbf[ch] = singleCbfTmp[ch]; bestModeId[COMPONENT_Y] = modeId; if(bestModeId[COMPONENT_Y] == firstCheckId) { xStoreIntraResultQT(COMPONENT_Y, bLumaOnly?COMPONENT_Y:COMPONENT_Cr, rTu ); m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[ uiFullDepth ][ CI_TEMP_BEST ] ); } if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { const Int xOffset = rTu.getRect( COMPONENT_Y ).x0; const Int yOffset = rTu.getRect( COMPONENT_Y ).y0; for (UInt storedResidualIndex = 0; storedResidualIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; storedResidualIndex++) { if (bMaintainResidual[storedResidualIndex]) { xStoreCrossComponentPredictionResult(resiLuma[storedResidualIndex], resiLumaSingle[storedResidualIndex], rTu, xOffset, yOffset, MAX_CU_SIZE, MAX_CU_SIZE); } } } } if (modeId == firstCheckId) { m_pcRDGoOnSbacCoder->load ( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); } } for(UInt ch=COMPONENT_Y; ch<numValidComp; ch++) { const ComponentID compID=ComponentID(ch); if (rTu.ProcessComponentSection(compID)) { const UInt totalAdjustedDepthChan = rTu.GetTransformDepthTotalAdj(compID); pcCU ->setTransformSkipSubParts ( bestModeId[COMPONENT_Y], compID, uiAbsPartIdx, totalAdjustedDepthChan ); } } if(bestModeId[COMPONENT_Y] == firstCheckId) { xLoadIntraResultQT(COMPONENT_Y, bLumaOnly?COMPONENT_Y:COMPONENT_Cr, rTu ); for(UInt ch=COMPONENT_Y; ch< numValidComp; ch++) { const ComponentID compID=ComponentID(ch); if (rTu.ProcessComponentSection(compID)) pcCU->setCbfSubParts ( uiSingleCbf[compID] << uiTrDepth, compID, uiAbsPartIdx, rTu.GetTransformDepthTotalAdj(compID) ); } m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[ uiFullDepth ][ CI_TEMP_BEST ] ); } if( !bLumaOnly ) { bestModeId[COMPONENT_Cb] = bestModeId[COMPONENT_Cr] = bestModeId[COMPONENT_Y]; if (rTu.ProcessComponentSection(COMPONENT_Cb) && bestModeId[COMPONENT_Y] == 1) { //In order not to code TS flag when cbf is zero, the case for TS with cbf being zero is forbidden. for (UInt ch=COMPONENT_Cb; ch<numValidComp; ch++) { if (uiSingleCbf[ch] == 0) { const ComponentID compID=ComponentID(ch); const UInt totalAdjustedDepthChan = rTu.GetTransformDepthTotalAdj(compID); pcCU ->setTransformSkipSubParts ( 0, compID, uiAbsPartIdx, totalAdjustedDepthChan); bestModeId[ch] = 0; } } } } }
// TransformSkip模式为false，表示不会跳过变换步骤 else { //----- store original entropy coding status ----- if( bCheckSplit ) { m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); } //----- code luma/chroma block with given intra prediction mode and store Cbf----- dSingleCost = 0.0; for (UInt ch=COMPONENT_Y; ch<numValidComp; ch++) { const ComponentID compID = ComponentID(ch); if (rTu.ProcessComponentSection(compID)) { const UInt totalAdjustedDepthChan = rTu.GetTransformDepthTotalAdj(compID); pcCU ->setTransformSkipSubParts ( 0, compID, uiAbsPartIdx, totalAdjustedDepthChan ); }
// 亮度块的预测、变换和量化 xIntraCodingTUBlock( pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaSingle, false, uiSingleDist[toChannelType(compID)], compID, rTu DEBUG_STRING_PASS_INTO(sDebug)); if( bCheckSplit ) { uiSingleCbf[compID] = pcCU->getCbf( uiAbsPartIdx, compID, uiTrDepth ); } } //----- determine rate and r-d cost ----- UInt uiSingleBits = xGetIntraBitsQT( rTu, true, !bLumaOnly, false ); if(m_pcEncCfg->getRDpenalty() && (uiLog2TrSize==5) && !isIntraSlice) { uiSingleBits=uiSingleBits*4; } dSingleCost = m_pcRdCost->calcRdCost( uiSingleBits, uiSingleDist[CHANNEL_TYPE_LUMA] + uiSingleDist[CHANNEL_TYPE_CHROMA] ); if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { const Int xOffset = rTu.getRect( COMPONENT_Y ).x0; const Int yOffset = rTu.getRect( COMPONENT_Y ).y0; for (UInt storedResidualIndex = 0; storedResidualIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; storedResidualIndex++) { if (bMaintainResidual[storedResidualIndex]) { xStoreCrossComponentPredictionResult(resiLuma[storedResidualIndex], resiLumaSingle[storedResidualIndex], rTu, xOffset, yOffset, MAX_CU_SIZE, MAX_CU_SIZE); } } } } }
// 当前块是否向下继续划分为4个子块，如果是，那么就递归处理 if( bCheckSplit ) { //----- store full entropy coding status, load original entropy coding status ----- if( bCheckFull ) { m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_TEST ] ); m_pcRDGoOnSbacCoder->load ( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); } else { m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); } //----- code splitted block ----- Double dSplitCost = 0.0; Distortion uiSplitDist[MAX_NUM_CHANNEL_TYPE] = {0,0}; UInt uiSplitCbf[MAX_NUM_COMPONENT] = {0,0,0}; TComTURecurse tuRecurseChild(rTu, false); DEBUG_STRING_NEW(sSplit) do { DEBUG_STRING_NEW(sChild) #if HHI_RQT_INTRA_SPEEDUP
// 递归调用 xRecurIntraCodingQT( bLumaOnly, pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaSplit, uiSplitDist[0], uiSplitDist[1], bCheckFirst, dSplitCost, tuRecurseChild DEBUG_STRING_PASS_INTO(sChild) ); #else xRecurIntraCodingQT( bLumaOnly, pcOrgYuv, pcPredYuv, pcResiYuv, resiLumaSplit, uiSplitDist[0], uiSplitDist[1], dSplitCost, tuRecurseChild DEBUG_STRING_PASS_INTO(sChild) ); #endif DEBUG_STRING_APPEND(sSplit, sChild) for(UInt ch=0; ch<numValidComp; ch++) { uiSplitCbf[ch] |= pcCU->getCbf( tuRecurseChild.GetAbsPartIdxTU(), ComponentID(ch), tuRecurseChild.GetTransformDepthRel() ); } } while (tuRecurseChild.nextSection(rTu) ); UInt uiPartsDiv = rTu.GetAbsPartIdxNumParts(); for(UInt ch=COMPONENT_Y; ch<numValidComp; ch++) { if (uiSplitCbf[ch]) { const UInt flag=1<<uiTrDepth; const ComponentID compID=ComponentID(ch); UChar *pBase=pcCU->getCbf( compID ); for( UInt uiOffs = 0; uiOffs < uiPartsDiv; uiOffs++ ) { pBase[ uiAbsPartIdx + uiOffs ] |= flag; } } } //----- restore context states ----- m_pcRDGoOnSbacCoder->load ( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); //----- determine rate and r-d cost ----- UInt uiSplitBits = xGetIntraBitsQT( rTu, true, !bLumaOnly, false ); dSplitCost = m_pcRdCost->calcRdCost( uiSplitBits, uiSplitDist[CHANNEL_TYPE_LUMA] + uiSplitDist[CHANNEL_TYPE_CHROMA] ); //===== compare and set best ===== if( dSplitCost < dSingleCost ) { //--- update cost --- DEBUG_STRING_SWAP(sSplit, sDebug) ruiDistY += uiSplitDist[CHANNEL_TYPE_LUMA]; ruiDistC += uiSplitDist[CHANNEL_TYPE_CHROMA]; dRDCost += dSplitCost; if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { const Int xOffset = rTu.getRect( COMPONENT_Y ).x0; const Int yOffset = rTu.getRect( COMPONENT_Y ).y0; for (UInt storedResidualIndex = 0; storedResidualIndex < NUMBER_OF_STORED_RESIDUAL_TYPES; storedResidualIndex++) { if (bMaintainResidual[storedResidualIndex]) { xStoreCrossComponentPredictionResult(resiLuma[storedResidualIndex], resiLumaSplit[storedResidualIndex], rTu, xOffset, yOffset, MAX_CU_SIZE, MAX_CU_SIZE); } } } return; } //----- set entropy coding status ----- m_pcRDGoOnSbacCoder->load ( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_TEST ] ); //--- set transform index and Cbf values --- pcCU->setTrIdxSubParts( uiTrDepth, uiAbsPartIdx, uiFullDepth ); for(UInt ch=0; ch<numValidComp; ch++) { const ComponentID compID=ComponentID(ch); const TComRectangle &tuRect=rTu.getRect(compID); const UInt totalAdjustedDepthChan = rTu.GetTransformDepthTotalAdj(compID); pcCU->setCbfSubParts ( uiSingleCbf[compID] << uiTrDepth, compID, uiAbsPartIdx, totalAdjustedDepthChan ); pcCU ->setTransformSkipSubParts ( bestModeId[compID], compID, uiAbsPartIdx, totalAdjustedDepthChan ); //--- set reconstruction for next intra prediction blocks ---
// 执行像素块的重建操作 const UInt uiQTLayer = pcCU->getSlice()->getSPS()->getQuadtreeTULog2MaxSize() - uiLog2TrSize; const UInt uiZOrder = pcCU->getZorderIdxInCU() + uiAbsPartIdx; const UInt uiWidth = tuRect.width; const UInt uiHeight = tuRect.height; Pel* piSrc = m_pcQTTempTComYuv[ uiQTLayer ].getAddr( compID, uiAbsPartIdx ); UInt uiSrcStride = m_pcQTTempTComYuv[ uiQTLayer ].getStride ( compID ); Pel* piDes = pcCU->getPic()->getPicYuvRec()->getAddr( compID, pcCU->getAddr(), uiZOrder ); UInt uiDesStride = pcCU->getPic()->getPicYuvRec()->getStride ( compID ); for( UInt uiY = 0; uiY < uiHeight; uiY++, piSrc += uiSrcStride, piDes += uiDesStride ) { for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { piDes[ uiX ] = piSrc[ uiX ]; } } } } ruiDistY += uiSingleDist[CHANNEL_TYPE_LUMA]; ruiDistC += uiSingleDist[CHANNEL_TYPE_CHROMA]; dRDCost += dSingleCost; }
// 帧内预测的一整套流程：预测+变换量化+反变换反量化+重建像素
Void TEncSearch::xIntraCodingTUBlock( TComYuv* pcOrgYuv, TComYuv* pcPredYuv, TComYuv* pcResiYuv, Pel resiLuma[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE], const Bool checkCrossCPrediction, Distortion& ruiDist, const ComponentID compID, TComTU& rTu DEBUG_STRING_FN_DECLARE(sDebug) ,Int default0Save1Load2 ) { if (!rTu.ProcessComponentSection(compID)) return; const Bool bIsLuma = isLuma(compID); const TComRectangle &rect= rTu.getRect(compID); TComDataCU *pcCU=rTu.getCU(); const UInt uiAbsPartIdx=rTu.GetAbsPartIdxTU(); const UInt uiTrDepth=rTu.GetTransformDepthRelAdj(compID);
const UInt uiFullDepth = rTu.GetTransformDepthTotal(); // 获取深度 const UInt uiLog2TrSize = rTu.GetLog2LumaTrSize(); const ChromaFormat chFmt = pcOrgYuv->getChromaFormat(); const ChannelType chType = toChannelType(compID); const UInt uiWidth = rect.width; // 获取宽度 const UInt uiHeight = rect.height; // 获取高度 const UInt uiStride = pcOrgYuv ->getStride (compID); // 获取偏移 Pel* piOrg = pcOrgYuv ->getAddr( compID, uiAbsPartIdx ); // 原始的像素地址 Pel* piPred = pcPredYuv->getAddr( compID, uiAbsPartIdx ); // 预测的像素地址 Pel* piResi = pcResiYuv->getAddr( compID, uiAbsPartIdx ); // 残差的像素地址 Pel* piReco = pcPredYuv->getAddr( compID, uiAbsPartIdx ); // 重建的像素地址 const UInt uiQTLayer = pcCU->getSlice()->getSPS()->getQuadtreeTULog2MaxSize() - uiLog2TrSize; Pel* piRecQt = m_pcQTTempTComYuv[ uiQTLayer ].getAddr( compID, uiAbsPartIdx ); const UInt uiRecQtStride = m_pcQTTempTComYuv[ uiQTLayer ].getStride(compID); const UInt uiZOrder = pcCU->getZorderIdxInCU() + uiAbsPartIdx; // Z扫描的顺序 Pel* piRecIPred = pcCU->getPic()->getPicYuvRec()->getAddr( compID, pcCU->getAddr(), uiZOrder ); UInt uiRecIPredStride = pcCU->getPic()->getPicYuvRec()->getStride ( compID ); TCoeff* pcCoeff = m_ppcQTTempCoeff[compID][uiQTLayer] + rTu.getCoefficientOffset(compID); // 系数 Bool useTransformSkip = pcCU->getTransformSkip(uiAbsPartIdx, compID); #if ADAPTIVE_QP_SELECTION TCoeff* pcArlCoeff = m_ppcQTTempArlCoeff[compID][ uiQTLayer ] + rTu.getCoefficientOffset(compID); #endif const UInt uiChPredMode = pcCU->getIntraDir( chType, uiAbsPartIdx ); // 获取预测模式 const UInt uiChCodedMode = (uiChPredMode==DM_CHROMA_IDX && !bIsLuma) ? pcCU->getIntraDir(CHANNEL_TYPE_LUMA, getChromasCorrespondingPULumaIdx(uiAbsPartIdx, chFmt)) : uiChPredMode; const UInt uiChFinalMode = ((chFmt == CHROMA_422) && !bIsLuma) ? g_chroma422IntraAngleMappingTable[uiChCodedMode] : uiChCodedMode; const Int blkX = g_auiRasterToPelX[ g_auiZscanToRaster[ uiAbsPartIdx ] ]; const Int blkY = g_auiRasterToPelY[ g_auiZscanToRaster[ uiAbsPartIdx ] ]; const Int bufferOffset = blkX + (blkY * MAX_CU_SIZE); Pel *const encoderLumaResidual = resiLuma[RESIDUAL_ENCODER_SIDE ] + bufferOffset; Pel *const reconstructedLumaResidual = resiLuma[RESDIUAL_RECONSTRUCTED] + bufferOffset; const Bool bUseCrossCPrediction = isChroma(compID) && (uiChPredMode == DM_CHROMA_IDX) && checkCrossCPrediction; const Bool bUseReconstructedResidualForEstimate = m_pcEncCfg->getUseReconBasedCrossCPredictionEstimate(); Pel *const lumaResidualForEstimate = bUseReconstructedResidualForEstimate ? reconstructedLumaResidual : encoderLumaResidual; #ifdef DEBUG_STRING const Int debugPredModeMask=DebugStringGetPredModeMask(MODE_INTRA); #endif //===== init availability pattern ===== Bool bAboveAvail = false; // 上方是否有效 Bool bLeftAvail = false; // 左侧是否有效 DEBUG_STRING_NEW(sTemp) #ifndef DEBUG_STRING
// default0Save1Load2参数控制了预测像素的生成方式 if( default0Save1Load2 != 2 ) #endif { const Bool bUseFilteredPredictions=TComPrediction::filteringIntraReferenceSamples(compID, uiChFinalMode, uiWidth, uiHeight, chFmt, pcCU->getSlice()->getSPS()->getDisableIntraReferenceSmoothing()); initAdiPatternChType( rTu, bAboveAvail, bLeftAvail, compID, bUseFilteredPredictions DEBUG_STRING_PASS_INTO(sDebug) ); //===== get prediction signal =====
// 预测操作 predIntraAng( compID, uiChFinalMode, piOrg, uiStride, piPred, uiStride, rTu, bAboveAvail, bLeftAvail, bUseFilteredPredictions ); // save prediction
// 保存预测信息 if( default0Save1Load2 == 1 ) { Pel* pPred = piPred; Pel* pPredBuf = m_pSharedPredTransformSkip[compID]; Int k = 0; for( UInt uiY = 0; uiY < uiHeight; uiY++ ) { for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { pPredBuf[ k ++ ] = pPred[ uiX ]; } pPred += uiStride; } } } #ifndef DEBUG_STRING else { // load prediction
// 直接计算预测值 Pel* pPred = piPred; Pel* pPredBuf = m_pSharedPredTransformSkip[compID]; Int k = 0; for( UInt uiY = 0; uiY < uiHeight; uiY++ ) { for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { pPred[ uiX ] = pPredBuf[ k ++ ]; } pPred += uiStride; } } #endif //===== get residual signal ===== { // get residual
// 计算残差 Pel* pOrg = piOrg; Pel* pPred = piPred; Pel* pResi = piResi; for( UInt uiY = 0; uiY < uiHeight; uiY++ ) { for( UInt uiX = 0; uiX < uiWidth; uiX++ ) {
// 此处计算残差数据 pResi[ uiX ] = pOrg[ uiX ] - pPred[ uiX ]; } pOrg += uiStride; pResi += uiStride; pPred += uiStride; } } if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { if (bUseCrossCPrediction) { if (xCalcCrossComponentPredictionAlpha( rTu, compID, lumaResidualForEstimate, piResi, uiWidth, uiHeight, MAX_CU_SIZE, uiStride ) == 0) return; TComTrQuant::crossComponentPrediction ( rTu, compID, reconstructedLumaResidual, piResi, piResi, uiWidth, uiHeight, MAX_CU_SIZE, uiStride, uiStride, false ); } else if (isLuma(compID) && !bUseReconstructedResidualForEstimate) { xStoreCrossComponentPredictionResult( encoderLumaResidual, piResi, rTu, 0, 0, MAX_CU_SIZE, uiStride ); } } //===== transform and quantization =====
// 变换和量化 //--- init rate estimation arrays for RDOQ ---
// 是否跳过变换操作 if( useTransformSkip ? m_pcEncCfg->getUseRDOQTS() : m_pcEncCfg->getUseRDOQ() ) {
// 比特数估计 m_pcEntropyCoder->estimateBit( m_pcTrQuant->m_pcEstBitsSbac, uiWidth, uiHeight, chType ); } //--- transform and quantization --- TCoeff uiAbsSum = 0; if (bIsLuma) { pcCU ->setTrIdxSubParts ( uiTrDepth, uiAbsPartIdx, uiFullDepth ); } const QpParam cQP(*pcCU, compID); #if RDOQ_CHROMA_LAMBDA m_pcTrQuant->selectLambda (compID); #endif
// 变换（连同量化一起） m_pcTrQuant->transformNxN ( rTu, compID, piResi, uiStride, pcCoeff, #if ADAPTIVE_QP_SELECTION pcArlCoeff, #endif uiAbsSum, cQP ); //--- inverse transform --- // uiAbsSum表示变换系数的绝对值之和 #ifdef DEBUG_STRING if ( (uiAbsSum > 0) || (DebugOptionList::DebugString_InvTran.getInt()&debugPredModeMask) ) #else if ( uiAbsSum > 0 ) #endif {
// 反变换 m_pcTrQuant->invTransformNxN ( rTu, compID, piResi, uiStride, pcCoeff, cQP DEBUG_STRING_PASS_INTO_OPTIONAL(&sDebug, (DebugOptionList::DebugString_InvTran.getInt()&debugPredModeMask)) ); } else { Pel* pResi = piResi; memset( pcCoeff, 0, sizeof( TCoeff ) * uiWidth * uiHeight ); for( UInt uiY = 0; uiY < uiHeight; uiY++ ) { memset( pResi, 0, sizeof( Pel ) * uiWidth ); pResi += uiStride; } } //===== reconstruction =====
// 图像重建 { Pel* pPred = piPred; Pel* pResi = piResi; Pel* pReco = piReco; Pel* pRecQt = piRecQt; Pel* pRecIPred = piRecIPred; const UInt clipbd=g_bitDepth[chType]; if (pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction()) { if (bUseCrossCPrediction) { TComTrQuant::crossComponentPrediction( rTu, compID, reconstructedLumaResidual, piResi, piResi, uiWidth, uiHeight, MAX_CU_SIZE, uiStride, uiStride, true ); } else if (isLuma(compID)) { xStoreCrossComponentPredictionResult( reconstructedLumaResidual, piResi, rTu, 0, 0, MAX_CU_SIZE, uiStride ); } } #ifdef DEBUG_STRING std::stringstream ss(stringstream::out); const Bool bDebugPred=((DebugOptionList::DebugString_Pred.getInt()&debugPredModeMask) && DEBUG_STRING_CHANNEL_CONDITION(compID)); const Bool bDebugResi=((DebugOptionList::DebugString_Resi.getInt()&debugPredModeMask) && DEBUG_STRING_CHANNEL_CONDITION(compID)); const Bool bDebugReco=((DebugOptionList::DebugString_Reco.getInt()&debugPredModeMask) && DEBUG_STRING_CHANNEL_CONDITION(compID)); if (bDebugPred || bDebugResi || bDebugReco) { ss << "###: " << "CompID: " << compID << " pred mode (ch/fin): " << uiChPredMode << "/" << uiChFinalMode << " absPartIdx: " << rTu.GetAbsPartIdxTU() << "\n"; for( UInt uiY = 0; uiY < uiHeight; uiY++ ) { ss << "###: "; if (bDebugPred) { ss << " - pred: "; for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { ss << pPred[ uiX ] << ", "; } } if (bDebugResi) ss << " - resi: "; for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { if (bDebugResi) ss << pResi[ uiX ] << ", "; pReco [ uiX ] = Pel(ClipBD<Int>( Int(pPred[uiX]) + Int(pResi[uiX]), clipbd )); pRecQt [ uiX ] = pReco[ uiX ]; pRecIPred[ uiX ] = pReco[ uiX ]; } if (bDebugReco) { ss << " - reco: "; for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { ss << pReco[ uiX ] << ", "; } } pPred += uiStride; pResi += uiStride; pReco += uiStride; pRecQt += uiRecQtStride; pRecIPred += uiRecIPredStride; ss << "\n"; } DEBUG_STRING_APPEND(sDebug, ss.str()) } else #endif { for( UInt uiY = 0; uiY < uiHeight; uiY++ ) { for( UInt uiX = 0; uiX < uiWidth; uiX++ ) { pReco [ uiX ] = Pel(ClipBD<Int>( Int(pPred[uiX]) + Int(pResi[uiX]), clipbd )); pRecQt [ uiX ] = pReco[ uiX ]; pRecIPred[ uiX ] = pReco[ uiX ]; } pPred += uiStride; pResi += uiStride; pReco += uiStride; pRecQt += uiRecQtStride; pRecIPred += uiRecIPredStride; } } } //===== update distortion =====
// 失真代价更新 ruiDist += m_pcRdCost->getDistPart( g_bitDepth[chType], piReco, uiStride, piOrg, uiStride, uiWidth, uiHeight, compID ); }
Void TComPrediction::predIntraAng( const ComponentID compID, UInt uiDirMode, Pel* piOrg /* Will be null for decoding */, UInt uiOrgStride, Pel* piPred, UInt uiStride, TComTU &rTu, Bool bAbove, Bool bLeft, const Bool bUseFilteredPredSamples, const Bool bUseLosslessDPCM ) { const ChromaFormat format = rTu.GetChromaFormat(); // 获得图片格式，一般为YUV420 const ChannelType channelType = toChannelType(compID); const TComRectangle &rect = rTu.getRect(isLuma(compID) ? COMPONENT_Y : COMPONENT_Cb); const Int iWidth = rect.width; // TU的宽 const Int iHeight = rect.height; // TU的高 assert( g_aucConvertToBit[ iWidth ] >= 0 ); // 4x 4 assert( g_aucConvertToBit[ iWidth ] <= 5 ); // 128x128 //assert( iWidth == iHeight ); Pel *pDst = piPred; // 预测值的首地址 // get starting pixel in block
// 获取块中的开始像素 const Int sw = (2 * iWidth + 1);
// 如果预测方式为垂直或水平，则bUseLosslessDPCM为true if ( bUseLosslessDPCM ) {
// 得到参考块的左上方地址（不在参考块内） const Pel *ptrSrc = getPredictorPtr( compID, false ); // Sample Adaptive intra-Prediction (SAP)
// 水平模式预测 if (uiDirMode==HOR_IDX) { // left column filled with reference samples // remaining columns filled with piOrg data (if available). for(Int y=0; y<iHeight; y++) { piPred[y*uiStride+0] = ptrSrc[(y+1)*sw]; } if (piOrg!=0) { piPred+=1; // miss off first column for(Int y=0; y<iHeight; y++, piPred+=uiStride, piOrg+=uiOrgStride) { memcpy(piPred, piOrg, (iWidth-1)*sizeof(Pel)); } } }
// 垂直模式预测 else // VER_IDX { // top row filled with reference samples // remaining rows filled with piOrd data (if available) for(Int x=0; x<iWidth; x++) { piPred[x] = ptrSrc[x+1]; } if (piOrg!=0) { piPred+=uiStride; // miss off the first row for(Int y=1; y<iHeight; y++, piPred+=uiStride, piOrg+=uiOrgStride) { memcpy(piPred, piOrg, iWidth*sizeof(Pel)); } } } } else { const Pel *ptrSrc = getPredictorPtr( compID, bUseFilteredPredSamples ); if ( uiDirMode == PLANAR_IDX ) {
// Planar模式 xPredIntraPlanar( ptrSrc+sw+1, sw, pDst, uiStride, iWidth, iHeight, channelType, format ); } else { // Create the prediction TComDataCU *const pcCU = rTu.getCU(); const UInt uiAbsPartIdx = rTu.GetAbsPartIdxTU(); const Bool enableEdgeFilters = !(pcCU->isRDPCMEnabled(uiAbsPartIdx) && pcCU->getCUTransquantBypass(uiAbsPartIdx)); #if RExt__O0043_BEST_EFFORT_DECODING
// 角度模式 xPredIntraAng( g_bitDepthInStream[channelType], ptrSrc+sw+1, sw, pDst, uiStride, iWidth, iHeight, channelType, format, uiDirMode, bAbove, bLeft, enableEdgeFilters ); #else xPredIntraAng( g_bitDepth[channelType], ptrSrc+sw+1, sw, pDst, uiStride, iWidth, iHeight, channelType, format, uiDirMode, bAbove, bLeft, enableEdgeFilters ); #endif if(( uiDirMode == DC_IDX ) && bAbove && bLeft ) {
// DC模式 xDCPredFiltering( ptrSrc+sw+1, sw, pDst, uiStride, iWidth, iHeight, channelType ); } } } }
3、estIntraPredChromaQT（色度块的帧内预测）：
Void TEncSearch::estIntraPredChromaQT(TComDataCU* pcCU, TComYuv* pcOrgYuv, TComYuv* pcPredYuv, TComYuv* pcResiYuv, TComYuv* pcRecoYuv, Pel resiLuma[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE], Distortion uiPreCalcDistC DEBUG_STRING_FN_DECLARE(sDebug)) { pcCU->getTotalDistortion () -= uiPreCalcDistC; //const UInt uiDepthCU = pcCU->getDepth(0); const UInt uiInitTrDepth = pcCU->getPartitionSize(0) != SIZE_2Nx2N && enable4ChromaPUsInIntraNxNCU(pcOrgYuv->getChromaFormat()) ? 1 : 0; // const UInt uiNumPU = 1<<(2*uiInitTrDepth); TComTURecurse tuRecurseCU(pcCU, 0); TComTURecurse tuRecurseWithPU(tuRecurseCU, false, (uiInitTrDepth==0)?TComTU::DONT_SPLIT : TComTU::QUAD_SPLIT); const UInt uiQNumParts = tuRecurseWithPU.GetAbsPartIdxNumParts(); const UInt uiDepthCU=tuRecurseWithPU.getCUDepth(); const UInt numberValidComponents = pcCU->getPic()->getNumberValidComponents(); do { UInt uiBestMode = 0; Distortion uiBestDist = 0; Double dBestCost = MAX_DOUBLE; //----- init mode list ----- if (tuRecurseWithPU.ProcessChannelSection(CHANNEL_TYPE_CHROMA)) { UInt uiModeList[FAST_UDI_MAX_RDMODE_NUM]; const UInt uiQPartNum = uiQNumParts; const UInt uiPartOffset = tuRecurseWithPU.GetAbsPartIdxTU(); { UInt uiMinMode = 0; UInt uiMaxMode = NUM_CHROMA_MODE; //----- check chroma modes ----- pcCU->getAllowedChromaDir( uiPartOffset, uiModeList ); #if RExt__ENVIRONMENT_VARIABLE_DEBUG_AND_TEST if (DebugOptionList::ForceChromaMode.isSet()) { uiMinMode=DebugOptionList::ForceChromaMode.getInt(); if (uiModeList[uiMinMode]==34) uiMinMode=5; // if the fixed mode has been renumbered because DM_CHROMA covers it, use DM_CHROMA. uiMaxMode=uiMinMode+1; } #endif DEBUG_STRING_NEW(sPU) for( UInt uiMode = uiMinMode; uiMode < uiMaxMode; uiMode++ ) { //----- restore context models ----- m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[uiDepthCU][CI_CURR_BEST] ); DEBUG_STRING_NEW(sMode) //----- chroma coding ----- Distortion uiDist = 0; pcCU->setIntraDirSubParts ( CHANNEL_TYPE_CHROMA, uiModeList[uiMode], uiPartOffset, uiDepthCU+uiInitTrDepth ); xRecurIntraChromaCodingQT ( pcOrgYuv, pcPredYuv, pcResiYuv, resiLuma, uiDist, tuRecurseWithPU DEBUG_STRING_PASS_INTO(sMode) ); if( pcCU->getSlice()->getPPS()->getUseTransformSkip() ) { m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[uiDepthCU][CI_CURR_BEST] ); } UInt uiBits = xGetIntraBitsQT( tuRecurseWithPU, false, true, false ); Double dCost = m_pcRdCost->calcRdCost( uiBits, uiDist ); //----- compare ----- if( dCost < dBestCost ) { DEBUG_STRING_SWAP(sPU, sMode); dBestCost = dCost; uiBestDist = uiDist; uiBestMode = uiModeList[uiMode]; xSetIntraResultChromaQT( pcRecoYuv, tuRecurseWithPU ); for (UInt componentIndex = COMPONENT_Cb; componentIndex < numberValidComponents; componentIndex++) { const ComponentID compID = ComponentID(componentIndex); ::memcpy( m_puhQTTempCbf[compID], pcCU->getCbf( compID )+uiPartOffset, uiQPartNum * sizeof( UChar ) ); ::memcpy( m_puhQTTempTransformSkipFlag[compID], pcCU->getTransformSkip( compID )+uiPartOffset, uiQPartNum * sizeof( UChar ) ); ::memcpy( m_phQTTempCrossComponentPredictionAlpha[compID], pcCU->getCrossComponentPredictionAlpha(compID)+uiPartOffset, uiQPartNum * sizeof( Char ) ); } } } DEBUG_STRING_APPEND(sDebug, sPU) //----- set data ----- for (UInt componentIndex = COMPONENT_Cb; componentIndex < numberValidComponents; componentIndex++) { const ComponentID compID = ComponentID(componentIndex); ::memcpy( pcCU->getCbf( compID )+uiPartOffset, m_puhQTTempCbf[compID], uiQPartNum * sizeof( UChar ) ); ::memcpy( pcCU->getTransformSkip( compID )+uiPartOffset, m_puhQTTempTransformSkipFlag[compID], uiQPartNum * sizeof( UChar ) ); ::memcpy( pcCU->getCrossComponentPredictionAlpha(compID)+uiPartOffset, m_phQTTempCrossComponentPredictionAlpha[compID], uiQPartNum * sizeof( Char ) ); } } if( ! tuRecurseWithPU.IsLastSection() ) { for (UInt ch=COMPONENT_Cb; ch<numberValidComponents; ch++) { const ComponentID compID = ComponentID(ch); const TComRectangle &tuRect = tuRecurseWithPU.getRect(compID); const UInt uiCompWidth = tuRect.width; const UInt uiCompHeight = tuRect.height; const UInt uiZOrder = pcCU->getZorderIdxInCU() + tuRecurseWithPU.GetAbsPartIdxTU(); Pel* piDes = pcCU->getPic()->getPicYuvRec()->getAddr( compID, pcCU->getAddr(), uiZOrder ); const UInt uiDesStride = pcCU->getPic()->getPicYuvRec()->getStride( compID); const Pel* piSrc = pcRecoYuv->getAddr( compID, uiPartOffset ); const UInt uiSrcStride = pcRecoYuv->getStride( compID); for( UInt uiY = 0; uiY < uiCompHeight; uiY++, piSrc += uiSrcStride, piDes += uiDesStride ) { for( UInt uiX = 0; uiX < uiCompWidth; uiX++ ) { piDes[ uiX ] = piSrc[ uiX ]; } } } } pcCU->setIntraDirSubParts( CHANNEL_TYPE_CHROMA, uiBestMode, uiPartOffset, uiDepthCU+uiInitTrDepth ); pcCU->getTotalDistortion () += uiBestDist; } } while (tuRecurseWithPU.nextSection(tuRecurseCU)); //----- restore context models ----- if( uiInitTrDepth != 0 ) { // set Cbf for all blocks UInt uiCombCbfU = 0; UInt uiCombCbfV = 0; UInt uiPartIdx = 0; for( UInt uiPart = 0; uiPart < 4; uiPart++, uiPartIdx += uiQNumParts ) { uiCombCbfU |= pcCU->getCbf( uiPartIdx, COMPONENT_Cb, 1 ); uiCombCbfV |= pcCU->getCbf( uiPartIdx, COMPONENT_Cr, 1 ); } for( UInt uiOffs = 0; uiOffs < 4 * uiQNumParts; uiOffs++ ) { pcCU->getCbf( COMPONENT_Cb )[ uiOffs ] |= uiCombCbfU; pcCU->getCbf( COMPONENT_Cr )[ uiOffs ] |= uiCombCbfV; } } m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[uiDepthCU][CI_CURR_BEST] ); }
Void TEncSearch::xRecurIntraChromaCodingQT(TComYuv* pcOrgYuv, TComYuv* pcPredYuv, TComYuv* pcResiYuv, Pel resiLuma[NUMBER_OF_STORED_RESIDUAL_TYPES][MAX_CU_SIZE * MAX_CU_SIZE], Distortion& ruiDist, TComTU& rTu DEBUG_STRING_FN_DECLARE(sDebug)) { TComDataCU *pcCU = rTu.getCU(); const UInt uiTrDepth = rTu.GetTransformDepthRel(); const UInt uiAbsPartIdx = rTu.GetAbsPartIdxTU(); const ChromaFormat format = rTu.GetChromaFormat(); UInt uiTrMode = pcCU->getTransformIdx( uiAbsPartIdx ); const UInt numberValidComponents = getNumberValidComponents(format); if( uiTrMode == uiTrDepth ) { if (!rTu.ProcessChannelSection(CHANNEL_TYPE_CHROMA)) return; const UInt uiFullDepth = rTu.GetTransformDepthTotal(); Bool checkTransformSkip = pcCU->getSlice()->getPPS()->getUseTransformSkip(); checkTransformSkip &= TUCompRectHasAssociatedTransformSkipFlag(rTu.getRect(COMPONENT_Cb), pcCU->getSlice()->getPPS()->getTransformSkipLog2MaxSize()); if ( m_pcEncCfg->getUseTransformSkipFast() ) { checkTransformSkip &= TUCompRectHasAssociatedTransformSkipFlag(rTu.getRect(COMPONENT_Y), pcCU->getSlice()->getPPS()->getTransformSkipLog2MaxSize()); if (checkTransformSkip) { Int nbLumaSkip = 0; const UInt maxAbsPartIdxSub=uiAbsPartIdx + (rTu.ProcessingAllQuadrants(COMPONENT_Cb)?1:4); for(UInt absPartIdxSub = uiAbsPartIdx; absPartIdxSub < maxAbsPartIdxSub; absPartIdxSub ++) { nbLumaSkip += pcCU->getTransformSkip(absPartIdxSub, COMPONENT_Y); } checkTransformSkip &= (nbLumaSkip > 0); } } for (UInt ch=COMPONENT_Cb; ch<numberValidComponents; ch++) { const ComponentID compID = ComponentID(ch); DEBUG_STRING_NEW(sDebugBestMode) //use RDO to decide whether Cr/Cb takes TS m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[uiFullDepth][CI_QT_TRAFO_ROOT] ); const Bool splitIntoSubTUs = rTu.getRect(compID).width != rTu.getRect(compID).height; TComTURecurse TUIterator(rTu, false, (splitIntoSubTUs ? TComTU::VERTICAL_SPLIT : TComTU::DONT_SPLIT), true, compID); const UInt partIdxesPerSubTU = TUIterator.GetAbsPartIdxNumParts(compID); do { const UInt subTUAbsPartIdx = TUIterator.GetAbsPartIdxTU(compID); Double dSingleCost = MAX_DOUBLE; Int bestModeId = 0; Distortion singleDistC = 0; UInt singleCbfC = 0; Distortion singleDistCTmp = 0; Double singleCostTmp = 0; UInt singleCbfCTmp = 0; Char bestCrossCPredictionAlpha = 0; Int bestTransformSkipMode = 0; const Bool checkCrossComponentPrediction = (pcCU->getIntraDir(CHANNEL_TYPE_CHROMA, subTUAbsPartIdx) == DM_CHROMA_IDX) && pcCU->getSlice()->getPPS()->getUseCrossComponentPrediction() && (pcCU->getCbf(subTUAbsPartIdx, COMPONENT_Y, uiTrDepth) != 0); const Int crossCPredictionModesToTest = checkCrossComponentPrediction ? 2 : 1; const Int transformSkipModesToTest = checkTransformSkip ? 2 : 1; const Int totalModesToTest = crossCPredictionModesToTest * transformSkipModesToTest; Int currModeId = 0; Int default0Save1Load2 = 0; for(Int transformSkipModeId = 0; transformSkipModeId < transformSkipModesToTest; transformSkipModeId++) { for(Int crossCPredictionModeId = 0; crossCPredictionModeId < crossCPredictionModesToTest; crossCPredictionModeId++) { pcCU->setCrossComponentPredictionAlphaPartRange(0, compID, subTUAbsPartIdx, partIdxesPerSubTU); DEBUG_STRING_NEW(sDebugMode) pcCU->setTransformSkipPartRange( transformSkipModeId, compID, subTUAbsPartIdx, partIdxesPerSubTU ); currModeId++; const Bool isOneMode = (totalModesToTest == 1); const Bool isLastMode = (currModeId == totalModesToTest); //NOTE: RExt - currModeId is indexed from 1 if (isOneMode) { default0Save1Load2 = 0; } else if (!isOneMode && (transformSkipModeId == 0) && (crossCPredictionModeId == 0)) { default0Save1Load2 = 1; //save prediction on first mode } else { default0Save1Load2 = 2; //load it on subsequent modes } singleDistCTmp = 0; xIntraCodingTUBlock( pcOrgYuv, pcPredYuv, pcResiYuv, resiLuma, (crossCPredictionModeId != 0), singleDistCTmp, compID, TUIterator DEBUG_STRING_PASS_INTO(sDebugMode), default0Save1Load2); singleCbfCTmp = pcCU->getCbf( subTUAbsPartIdx, compID, uiTrDepth); if ( ((crossCPredictionModeId == 1) && (pcCU->getCrossComponentPredictionAlpha(subTUAbsPartIdx, compID) == 0)) || ((transformSkipModeId == 1) && (singleCbfCTmp == 0))) //In order not to code TS flag when cbf is zero, the case for TS with cbf being zero is forbidden. { singleCostTmp = MAX_DOUBLE; } else if (!isOneMode) { UInt bitsTmp = xGetIntraBitsQTChroma( TUIterator, compID, false ); singleCostTmp = m_pcRdCost->calcRdCost( bitsTmp, singleDistCTmp); } if(singleCostTmp < dSingleCost) { DEBUG_STRING_SWAP(sDebugBestMode, sDebugMode) dSingleCost = singleCostTmp; singleDistC = singleDistCTmp; bestCrossCPredictionAlpha = (crossCPredictionModeId != 0) ? pcCU->getCrossComponentPredictionAlpha(subTUAbsPartIdx, compID) : 0; bestTransformSkipMode = transformSkipModeId; bestModeId = currModeId; singleCbfC = singleCbfCTmp; if (!isOneMode && !isLastMode) { xStoreIntraResultQT(compID, compID, TUIterator); m_pcRDGoOnSbacCoder->store( m_pppcRDSbacCoder[ uiFullDepth ][ CI_TEMP_BEST ] ); } } if (!isOneMode && !isLastMode) { m_pcRDGoOnSbacCoder->load ( m_pppcRDSbacCoder[ uiFullDepth ][ CI_QT_TRAFO_ROOT ] ); } } } if(bestModeId < totalModesToTest) { xLoadIntraResultQT(compID, compID, TUIterator); pcCU->setCbfPartRange( singleCbfC << uiTrDepth, compID, subTUAbsPartIdx, partIdxesPerSubTU ); m_pcRDGoOnSbacCoder->load( m_pppcRDSbacCoder[ uiFullDepth ][ CI_TEMP_BEST ] ); } DEBUG_STRING_APPEND(sDebug, sDebugBestMode) pcCU ->setTransformSkipPartRange ( bestTransformSkipMode, compID, subTUAbsPartIdx, partIdxesPerSubTU ); pcCU ->setCrossComponentPredictionAlphaPartRange( bestCrossCPredictionAlpha, compID, subTUAbsPartIdx, partIdxesPerSubTU ); ruiDist += singleDistC; } while (TUIterator.nextSection(rTu)); if (splitIntoSubTUs) offsetSubTUCBFs(rTu, compID); } } else { UInt uiSplitCbf[MAX_NUM_COMPONENT] = {0,0,0}; TComTURecurse tuRecurseChild(rTu, false); const UInt uiTrDepthChild = tuRecurseChild.GetTransformDepthRel(); do { DEBUG_STRING_NEW(sChild) xRecurIntraChromaCodingQT( pcOrgYuv, pcPredYuv, pcResiYuv, resiLuma, ruiDist, tuRecurseChild DEBUG_STRING_PASS_INTO(sChild) ); DEBUG_STRING_APPEND(sDebug, sChild) const UInt uiAbsPartIdxSub=tuRecurseChild.GetAbsPartIdxTU(); for(UInt ch=COMPONENT_Cb; ch<numberValidComponents; ch++) { uiSplitCbf[ch] |= pcCU->getCbf( uiAbsPartIdxSub, ComponentID(ch), uiTrDepthChild ); } } while ( tuRecurseChild.nextSection(rTu) ); UInt uiPartsDiv = rTu.GetAbsPartIdxNumParts(); for(UInt ch=COMPONENT_Cb; ch<numberValidComponents; ch++) { if (uiSplitCbf[ch]) { const UInt flag=1<<uiTrDepth; ComponentID compID=ComponentID(ch); UChar *pBase=pcCU->getCbf( compID ); for( UInt uiOffs = 0; uiOffs < uiPartsDiv; uiOffs++ ) { pBase[ uiAbsPartIdx + uiOffs ] |= flag; } } } } }
