X265运动估计

考虑到帧间相关性，只需要编码视频内动态变化的信息(视频内对象的运动信息，MV，运动矢量)即可大幅度减少所需编码的比特数。
另一方面，准确分割运动对象是非常复杂的，视频编码标准的运动估计是基于像素块进行的(精确到像素级的话复杂度太大)。
X265中在Search::predInterSearch中实现运动估计，返回运动矢量和参考帧的序号。

运动估计流程

/* find the best inter prediction for each PU of specified mode */
void Search::predInterSearch(Mode& interMode, const CUGeom& cuGeom, bool bChromaMC, uint32_t refMasks[2])
{
   		......
        if (m_param->analysisReuseLevel == 10 && m_param->interRefine > 1)
        {
            interDataCTU = m_frame->m_analysisData.interData;
            if ((cu.m_predMode[pu.puAbsPartIdx] == interDataCTU->modes[cuIdx + pu.puAbsPartIdx])
                && (cu.m_partSize[pu.puAbsPartIdx] == interDataCTU->partSize[cuIdx + pu.puAbsPartIdx])
                && !(interDataCTU->mergeFlag[cuIdx + puIdx])
                && (cu.m_cuDepth[0] == interDataCTU->depth[cuIdx]))
                useAsMVP = true;
        }
        /* find best cost merge candidate. note: 2Nx2N merge and bidir are handled as separate modes */
        uint32_t mrgCost = numPart == 1 ? MAX_UINT : mergeEstimation(cu, cuGeom, pu, puIdx, merge);
        bestME[0].cost = MAX_UINT;
        bestME[1].cost = MAX_UINT;

        getBlkBits((PartSize)cu.m_partSize[0], slice->isInterP(), puIdx, lastMode, m_listSelBits);
        bool bDoUnidir = true;

        cu.getNeighbourMV(puIdx, pu.puAbsPartIdx, interMode.interNeighbours);
        /* Uni-directional prediction */
        if ((m_param->analysisLoad && m_param->analysisReuseLevel > 1 && m_param->analysisReuseLevel != 10)
            || (m_param->analysisMultiPassRefine && m_param->rc.bStatRead) || (m_param->bAnalysisType == AVC_INFO) || (useAsMVP))
        {
            for (int list = 0; list < numPredDir; list++)
            {

               ......
            }
        }
        else if (m_param->bDistributeMotionEstimation)
        {
           ......

            /* if no peer threads were bonded, fall back to doing unidirectional
             * searches ourselves without overhead of singleMotionEstimation() */
        }
        if (bDoUnidir)//默认会进这里执行单向帧间预测
        {
            interMode.bestME[puIdx][0].ref = interMode.bestME[puIdx][1].ref = -1;
            uint32_t refMask = refMasks[puIdx] ? refMasks[puIdx] : (uint32_t)-1;

            for (int list = 0; list < numPredDir; list++)
            {
                for (int ref = 0; ref < numRefIdx[list]; ref++)
                {
                    ProfileCounter(interMode.cu, totalMotionReferences[cuGeom.depth]);

                    if (!(refMask & (1 << ref)))
                    {
                        ProfileCounter(interMode.cu, skippedMotionReferences[cuGeom.depth]);
                        continue;
                    }

                    uint32_t bits = m_listSelBits[list] + MVP_IDX_BITS;
                    bits += getTUBits(ref, numRefIdx[list]);

                    int numMvc = cu.getPMV(interMode.interNeighbours, list, ref, interMode.amvpCand[list][ref], mvc);

                    const MV* amvp = interMode.amvpCand[list][ref];
                    int mvpIdx = selectMVP(cu, pu, amvp, list, ref);
                    MV mvmin, mvmax, outmv, mvp = amvp[mvpIdx], mvp_lowres;
                    bool bLowresMVP = false;

                    if (!m_param->analysisSave && !m_param->analysisLoad) /* Prevents load/save outputs from diverging when lowresMV is not available */
                    {
                        MV lmv = getLowresMV(cu, pu, list, ref);
                        if (lmv.notZero())
                            mvc[numMvc++] = lmv;
                        if (m_param->bEnableHME)
                            mvp_lowres = lmv;
                    }
                    if (m_param->searchMethod == X265_SEA)
                    {
                        int puX = puIdx & 1;
                        int puY = puIdx >> 1;
                        for (int planes = 0; planes < INTEGRAL_PLANE_NUM; planes++)
                            m_me.integral[planes] = interMode.fencYuv->m_integral[list][ref][planes] + puX * pu.width + puY * pu.height * m_slice->m_refFrameList[list][ref]->m_reconPic->m_stride;
                    }
                    setSearchRange(cu, mvp, m_param->searchRange, mvmin, mvmax);//估计searchRange计算MV的范围
                    int satdCost = m_me.motionEstimate(&slice->m_mref[list][ref], mvmin, mvmax, mvp, numMvc, mvc, m_param->searchRange, outmv, m_param->maxSlices, 
                      m_param->bSourceReferenceEstimation ? m_slice->m_refFrameList[list][ref]->m_fencPic->getLumaAddr(0) : 0);//执行实际运动估计

                    if (m_param->bEnableHME && mvp_lowres.notZero() && mvp_lowres != mvp)
                    {
                        MV outmv_lowres;
                        setSearchRange(cu, mvp_lowres, m_param->searchRange, mvmin, mvmax);
                        int lowresMvCost = m_me.motionEstimate(&slice->m_mref[list][ref], mvmin, mvmax, mvp_lowres, numMvc, mvc, m_param->searchRange, outmv_lowres, m_param->maxSlices,
                            m_param->bSourceReferenceEstimation ? m_slice->m_refFrameList[list][ref]->m_fencPic->getLumaAddr(0) : 0);
                        if (lowresMvCost < satdCost)
                        {
                            outmv = outmv_lowres;
                            satdCost = lowresMvCost;
                            bLowresMVP = true;
                        }
                    }

                    /* Get total cost of partition, but only include MV bit cost once */
                    bits += m_me.bitcost(outmv);
                    uint32_t mvCost = m_me.mvcost(outmv);
                    uint32_t cost = (satdCost - mvCost) + m_rdCost.getCost(bits);
                    /* Update LowresMVP to best AMVP cand*/
                    if (bLowresMVP)
                        updateMVP(amvp[mvpIdx], outmv, bits, cost, mvp_lowres);//更新最佳MVP

                    /* Refine MVP selection, updates: mvpIdx, bits, cost */
                    mvp = checkBestMVP(amvp, outmv, mvpIdx, bits, cost);

                    if (cost < bestME[list].cost)
                    {
                        bestME[list].mv      = outmv;
                        bestME[list].mvp     = mvp;
                        bestME[list].mvpIdx  = mvpIdx;
                        bestME[list].ref     = ref;
                        bestME[list].cost    = cost;
                        bestME[list].bits    = bits;
                        bestME[list].mvCost  = mvCost;
                    }
                }
                /* the second list ref bits start at bit 16 */
                refMask >>= 16;
            }
        }

        /* Bi-directional prediction */
        MotionData bidir[2];
        uint32_t bidirCost = MAX_UINT;
        int bidirBits = 0;

        if (slice->isInterB() && !cu.isBipredRestriction() &&  /* biprediction is possible for this PU */ B帧支持双向预测
            cu.m_partSize[pu.puAbsPartIdx] != SIZE_2Nx2N &&    /* 2Nx2N biprediction is handled elsewhere */
            bestME[0].cost != MAX_UINT && bestME[1].cost != MAX_UINT)
        {
            bidir[0] = bestME[0];
            bidir[1] = bestME[1];

            int satdCost;

            if (m_me.bChromaSATD)
            {
                ......
                motionCompensation(cu, pu, tmpPredYuv, true, true);

                satdCost = m_me.bufSATD(tmpPredYuv.getLumaAddr(pu.puAbsPartIdx), tmpPredYuv.m_size) +
                           m_me.bufChromaSATD(tmpPredYuv, pu.puAbsPartIdx);
            }
            else
            {
                PicYuv* refPic0 = slice->m_refReconPicList[0][bestME[0].ref];
                PicYuv* refPic1 = slice->m_refReconPicList[1][bestME[1].ref];
                Yuv* bidirYuv = m_rqt[cuGeom.depth].bidirPredYuv;

                /* Generate reference subpels */
                predInterLumaPixel(pu, bidirYuv[0], *refPic0, bestME[0].mv);
                predInterLumaPixel(pu, bidirYuv[1], *refPic1, bestME[1].mv);
                primitives.pu[m_me.partEnum].pixelavg_pp[(tmpPredYuv.m_size % 64 == 0) && (bidirYuv[0].m_size % 64 == 0) && (bidirYuv[1].m_size % 64 == 0)](tmpPredYuv.m_buf[0], tmpPredYuv.m_size, bidirYuv[0].getLumaAddr(pu.puAbsPartIdx), bidirYuv[0].m_size,
                                                                                                 bidirYuv[1].getLumaAddr(pu.puAbsPartIdx), bidirYuv[1].m_size, 32);
                satdCost = m_me.bufSATD(tmpPredYuv.m_buf[0], tmpPredYuv.m_size);
            }

            bidirBits = bestME[0].bits + bestME[1].bits + m_listSelBits[2] - (m_listSelBits[0] + m_listSelBits[1]);
            bidirCost = satdCost + m_rdCost.getCost(bidirBits);

            bool bTryZero = bestME[0].mv.notZero() || bestME[1].mv.notZero();
            if (bTryZero)
            {
                /* Do not try zero MV if unidir motion predictors are beyond
                 * valid search area */
               ......
            }
            if (bTryZero)
            {
                /* coincident blocks of the two reference pictures */
               ......
            }
        }

        /* select best option and store into CU */
        if (mrgCost < bidirCost && mrgCost < bestME[0].cost && mrgCost < bestME[1].cost)
        {
            ......
            totalmebits += merge.bits;
        }
        else if (bidirCost < bestME[0].cost && bidirCost < bestME[1].cost)
        {
            lastMode = 2;
           ......
            totalmebits += bidirBits;
        }
        else if (bestME[0].cost <= bestME[1].cost)
        {
            lastMode = 0;
            ......
            totalmebits += bestME[0].bits;
        }
        else
        {
            lastMode = 1;
           ......
            totalmebits += bestME[1].bits;
        }

        motionCompensation(cu, pu, *predYuv, true, bChromaMC);//运动估计结束完之后做运动补偿
    }
    interMode.sa8dBits += totalmebits;
}

运动估计函数

实际的运动估计是在MotionEstimate::motionEstimate中实现的

int MotionEstimate::motionEstimate(ReferencePlanes *ref,
                                   const MV &       mvmin,
                                   const MV &       mvmax,
                                   const MV &       qmvp,
                                   int              numCandidates,
                                   const MV *       mvc,
                                   int              merange,
                                   MV &             outQMv,
                                   uint32_t         maxSlices,
                                   pixel *          srcReferencePlane)
{
   ......
    MV omv = bmv;  // current search origin or starting point

    int search = ref->isHMELowres ? (hme ? searchMethodL0 : searchMethodL1) : searchMethod;
    switch (search)
    {
    case X265_DIA_SEARCH:
    {
        /* diamond search, radius 1 */
        ......
    }
	//X265默认的搜索方式，先六边形搜索，再正方形搜索细化
    case X265_HEX_SEARCH:
    {
me_hex2:
        /* hexagon search, radius 2 */
        ......        
         COST_MV_X3_DIR(-2, 0, -1, 2,  1, 2, costs);//计算六边形左，左上，右上三个点的cost
        bcost <<= 3;左移3位，低3位存放位置信息
        if ((bmv.y >= mvmin.y) & (bmv.y <= mvmax.y))
            COPY1_IF_LT(bcost, (costs[0] << 3) + 2);
        if ((bmv.y + 2 >= mvmin.y) & (bmv.y + 2 <= mvmax.y))
        {
            COPY1_IF_LT(bcost, (costs[1] << 3) + 3);
            COPY1_IF_LT(bcost, (costs[2] << 3) + 4);
        }//不超过搜索范围的点比较并存储位置信息

        COST_MV_X3_DIR(2, 0,  1, -2, -1, -2, costs);//计算六边形右，右下，左下三个点的cost
        if ((bmv.y >= mvmin.y) & (bmv.y <= mvmax.y))
            COPY1_IF_LT(bcost, (costs[0] << 3) + 5);
        if ((bmv.y - 2 >= mvmin.y) & (bmv.y - 2 <= mvmax.y))
        {
            COPY1_IF_LT(bcost, (costs[1] << 3) + 6);
            COPY1_IF_LT(bcost, (costs[2] << 3) + 7);
        }不超过搜索范围的点比较并存储位置信息，和上面的过程对称

        if (bcost & 7)//取末尾3位，得到位置信息，判断是否存在cost更小的点
        {
            int dir = (bcost & 7) - 2;//获取相当于左偏移的方向

            if ((bmv.y + hex2[dir + 1].y >= mvmin.y) & (bmv.y + hex2[dir + 1].y <= mvmax.y))
            {
                bmv += hex2[dir + 1];//新的中心点位置

                /* half hexagon, not overlapping the previous iteration */
                for (int i = (merange >> 1) - 1; i > 0 && bmv.checkRange(mvmin, mvmax); i--)
                {
                    COST_MV_X3_DIR(hex2[dir + 0].x, hex2[dir + 0].y,
                        hex2[dir + 1].x, hex2[dir + 1].y,
                        hex2[dir + 2].x, hex2[dir + 2].y,
                        costs);//只计算一半，因为新的六边形有3个点在上一次已经计算过了
                    bcost &= ~7;//清掉低位位置信息

                    if ((bmv.y + hex2[dir + 0].y >= mvmin.y) & (bmv.y + hex2[dir + 0].y <= mvmax.y))
                        COPY1_IF_LT(bcost, (costs[0] << 3) + 1);

                    if ((bmv.y + hex2[dir + 1].y >= mvmin.y) & (bmv.y + hex2[dir + 1].y <= mvmax.y))
                        COPY1_IF_LT(bcost, (costs[1] << 3) + 2);

                    if ((bmv.y + hex2[dir + 2].y >= mvmin.y) & (bmv.y + hex2[dir + 2].y <= mvmax.y))
                        COPY1_IF_LT(bcost, (costs[2] << 3) + 3);

                    if (!(bcost & 7))
                        break;

                    dir += (bcost & 7) - 2;
                    dir = mod6m1[dir + 1];
                    bmv += hex2[dir + 1];
                }
            } 
        }
        bcost >>= 3;
		/* square refine */
        int dir = 0;
        COST_MV_X4_DIR(0, -1,  0, 1, -1, 0, 1, 0, costs);
        if ((bmv.y - 1 >= mvmin.y) & (bmv.y - 1 <= mvmax.y))
            COPY2_IF_LT(bcost, costs[0], dir, 1);
        if ((bmv.y + 1 >= mvmin.y) & (bmv.y + 1 <= mvmax.y))
            COPY2_IF_LT(bcost, costs[1], dir, 2);
        COPY2_IF_LT(bcost, costs[2], dir, 3);
        COPY2_IF_LT(bcost, costs[3], dir, 4);
        COST_MV_X4_DIR(-1, -1, -1, 1, 1, -1, 1, 1, costs);
        if ((bmv.y - 1 >= mvmin.y) & (bmv.y - 1 <= mvmax.y))
            COPY2_IF_LT(bcost, costs[0], dir, 5);
        if ((bmv.y + 1 >= mvmin.y) & (bmv.y + 1 <= mvmax.y))
            COPY2_IF_LT(bcost, costs[1], dir, 6);
        if ((bmv.y - 1 >= mvmin.y) & (bmv.y - 1 <= mvmax.y))
            COPY2_IF_LT(bcost, costs[2], dir, 7);
        if ((bmv.y + 1 >= mvmin.y) & (bmv.y + 1 <= mvmax.y))
            COPY2_IF_LT(bcost, costs[3], dir, 8);
        bmv += square1[dir];//在此基础上做进一步正方形搜索
        break;
    }

    case X265_UMH_SEARCH:
    {
        .......
    }

    case X265_STAR_SEARCH: // Adapted from HM ME
    {
        ......
    }

    case X265_SEA:
    {
        // Successive Elimination Algorithm
       .......
    }

    case X265_FULL_SEARCH:
    {
        // dead slow exhaustive search, but at least it uses sad_x4()
       ......
        }

        break;
    }

    default:
        X265_CHECK(0, "invalid motion estimate mode\n");
        break;
    }

    if (bprecost < bcost)
    {
        bmv = bestpre;
        bcost = bprecost;
    }
    else
        bmv = bmv.toQPel(); // promote search bmv to qpel

    const SubpelWorkload& wl = workload[this->subpelRefine];

    // check mv range for slice bound
    if ((maxSlices > 1) & ((bmv.y < qmvmin.y) | (bmv.y > qmvmax.y)))
    {
        bmv.y = x265_min(x265_max(bmv.y, qmvmin.y), qmvmax.y);
        bcost = subpelCompare(ref, bmv, satd) + mvcost(bmv);
    }

    if (!bcost)
    {
        /* if there was zero residual at the clipped MVP, we can skip subpel
         * refine, but we do need to include the mvcost in the returned cost */
        bcost = mvcost(bmv);
    }
    else if (ref->isLowres)
    {
        int bdir = 0;
        for (int i = 1; i <= wl.hpel_dirs; i++)
        {
            MV qmv = bmv + square1[i] * 2;

            /* skip invalid range */
            if ((qmv.y < qmvmin.y) | (qmv.y > qmvmax.y))
                continue;

            int cost = ref->lowresQPelCost(fenc, blockOffset, qmv, sad, hme) + mvcost(qmv);
            COPY2_IF_LT(bcost, cost, bdir, i);
        }

        bmv += square1[bdir] * 2;
        bcost = ref->lowresQPelCost(fenc, blockOffset, bmv, satd, hme) + mvcost(bmv);

        bdir = 0;
        for (int i = 1; i <= wl.qpel_dirs; i++)
        {
            MV qmv = bmv + square1[i];

            /* skip invalid range */
            if ((qmv.y < qmvmin.y) | (qmv.y > qmvmax.y))
                continue;

            int cost = ref->lowresQPelCost(fenc, blockOffset, qmv, satd, hme) + mvcost(qmv);
            COPY2_IF_LT(bcost, cost, bdir, i);
        }

        bmv += square1[bdir];
    }
    else
    {
        pixelcmp_t hpelcomp;

        if (wl.hpel_satd)
        {
            bcost = subpelCompare(ref, bmv, satd) + mvcost(bmv);
            hpelcomp = satd;
        }
        else
            hpelcomp = sad;

        for (int iter = 0; iter < wl.hpel_iters; iter++)
        {
            int bdir = 0;
            for (int i = 1; i <= wl.hpel_dirs; i++)
            {
                MV qmv = bmv + square1[i] * 2;

                // check mv range for slice bound
                if ((qmv.y < qmvmin.y) | (qmv.y > qmvmax.y))
                    continue;

                int cost = subpelCompare(ref, qmv, hpelcomp) + mvcost(qmv);
                COPY2_IF_LT(bcost, cost, bdir, i);
            }

            if (bdir)
                bmv += square1[bdir] * 2;
            else
                break;
        }

        /* if HPEL search used SAD, remeasure with SATD before QPEL */
        if (!wl.hpel_satd)
            bcost = subpelCompare(ref, bmv, satd) + mvcost(bmv);

        for (int iter = 0; iter < wl.qpel_iters; iter++)
        {
            int bdir = 0;
            for (int i = 1; i <= wl.qpel_dirs; i++)
            {
                MV qmv = bmv + square1[i];

                // check mv range for slice bound
                if ((qmv.y < qmvmin.y) | (qmv.y > qmvmax.y))
                    continue;

                int cost = subpelCompare(ref, qmv, satd) + mvcost(qmv);
                COPY2_IF_LT(bcost, cost, bdir, i);
            }

            if (bdir)
                bmv += square1[bdir];
            else
                break;
        }
    }

    // check mv range for slice bound
    X265_CHECK(((bmv.y >= qmvmin.y) & (bmv.y <= qmvmax.y)), "mv beyond range!");

    x265_emms();
    outQMv = bmv;
    return bcost;
}

运动补偿函数

将运动估计得到的MV加到参考块上得到预测块，用于后续计算残差和编码

void Predict::motionCompensation(const CUData& cu, const PredictionUnit& pu, Yuv& predYuv, bool bLuma, bool bChroma)
{
    int refIdx0 = cu.m_refIdx[0][pu.puAbsPartIdx];
    int refIdx1 = cu.m_refIdx[1][pu.puAbsPartIdx];

    if (cu.m_slice->isInterP())
    {
        /* P Slice */
        WeightValues wv0[3];

        X265_CHECK(refIdx0 >= 0, "invalid P refidx\n");
        X265_CHECK(refIdx0 < cu.m_slice->m_numRefIdx[0], "P refidx out of range\n");
        const WeightParam *wp0 = cu.m_slice->m_weightPredTable[0][refIdx0];

        MV mv0 = cu.m_mv[0][pu.puAbsPartIdx];
        cu.clipMv(mv0);

        if (cu.m_slice->m_pps->bUseWeightPred && wp0->wtPresent)//加权预测
        {
            for (int plane = 0; plane < (bChroma ? 3 : 1); plane++)
            {
                wv0[plane].w      = wp0[plane].inputWeight;
                wv0[plane].offset = wp0[plane].inputOffset * (1 << (X265_DEPTH - 8));
                wv0[plane].shift  = wp0[plane].log2WeightDenom;
                wv0[plane].round  = wp0[plane].log2WeightDenom >= 1 ? 1 << (wp0[plane].log2WeightDenom - 1) : 0;
            }

            ShortYuv& shortYuv = m_predShortYuv[0];//为了避免浮点运算，需要放大加权系数，因此需要使用short型变量(16位)

            if (bLuma)
                predInterLumaShort(pu, shortYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
            if (bChroma)
                predInterChromaShort(pu, shortYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);

            addWeightUni(pu, predYuv, shortYuv, wv0, bLuma, bChroma);
        }
        else
        {
            if (bLuma)
                predInterLumaPixel(pu, predYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
            if (bChroma)
                predInterChromaPixel(pu, predYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
        }
    }
    else
    {
        /* B Slice */

        WeightValues wv0[3], wv1[3];
        const WeightParam *pwp0, *pwp1;

        X265_CHECK(refIdx0 < cu.m_slice->m_numRefIdx[0], "bidir refidx0 out of range\n");
        X265_CHECK(refIdx1 < cu.m_slice->m_numRefIdx[1], "bidir refidx1 out of range\n");

        if (cu.m_slice->m_pps->bUseWeightedBiPred)
        {
            pwp0 = refIdx0 >= 0 ? cu.m_slice->m_weightPredTable[0][refIdx0] : NULL;
            pwp1 = refIdx1 >= 0 ? cu.m_slice->m_weightPredTable[1][refIdx1] : NULL;

            if (pwp0 && pwp1 && (pwp0->wtPresent || pwp1->wtPresent))
            {//B帧有可能有两个参考图像队列，这两个队列是否使用加权预测需要单独处理
                /* biprediction weighting */
                for (int plane = 0; plane < (bChroma ? 3 : 1); plane++)
                {
                    wv0[plane].w = pwp0[plane].inputWeight;
                    wv0[plane].o = pwp0[plane].inputOffset * (1 << (X265_DEPTH - 8));
                    wv0[plane].shift = pwp0[plane].log2WeightDenom;
                    wv0[plane].round = 1 << pwp0[plane].log2WeightDenom;

                    wv1[plane].w = pwp1[plane].inputWeight;
                    wv1[plane].o = pwp1[plane].inputOffset * (1 << (X265_DEPTH - 8));
                    wv1[plane].shift = wv0[plane].shift;
                    wv1[plane].round = wv0[plane].round;
                }
            }
            else
            {
                /* uniprediction weighting, always outputs to wv0 */
                const WeightParam* pwp = (refIdx0 >= 0) ? pwp0 : pwp1;
                for (int plane = 0; plane < (bChroma ? 3 : 1); plane++)
                {
                    wv0[plane].w = pwp[plane].inputWeight;
                    wv0[plane].offset = pwp[plane].inputOffset * (1 << (X265_DEPTH - 8));
                    wv0[plane].shift = pwp[plane].log2WeightDenom;
                    wv0[plane].round = pwp[plane].log2WeightDenom >= 1 ? 1 << (pwp[plane].log2WeightDenom - 1) : 0;
                }
            }
        }
        else
            pwp0 = pwp1 = NULL;

        if (refIdx0 >= 0 && refIdx1 >= 0)
        {
            MV mv0 = cu.m_mv[0][pu.puAbsPartIdx];
            MV mv1 = cu.m_mv[1][pu.puAbsPartIdx];
            cu.clipMv(mv0);
            cu.clipMv(mv1);

            if (bLuma)
            {
                predInterLumaShort(pu, m_predShortYuv[0], *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
                predInterLumaShort(pu, m_predShortYuv[1], *cu.m_slice->m_refReconPicList[1][refIdx1], mv1);
            }
            if (bChroma)
            {
                predInterChromaShort(pu, m_predShortYuv[0], *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
                predInterChromaShort(pu, m_predShortYuv[1], *cu.m_slice->m_refReconPicList[1][refIdx1], mv1);
            }

            if (pwp0 && pwp1 && (pwp0->wtPresent || pwp1->wtPresent))
                addWeightBi(pu, predYuv, m_predShortYuv[0], m_predShortYuv[1], wv0, wv1, bLuma, bChroma);
            else
                predYuv.addAvg(m_predShortYuv[0], m_predShortYuv[1], pu.puAbsPartIdx, pu.width, pu.height, bLuma, bChroma);
        }
        else if (refIdx0 >= 0)
        {
            MV mv0 = cu.m_mv[0][pu.puAbsPartIdx];
            cu.clipMv(mv0);

            if (pwp0 && pwp0->wtPresent)
            {
                ShortYuv& shortYuv = m_predShortYuv[0];

                if (bLuma)
                    predInterLumaShort(pu, shortYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
                if (bChroma)
                    predInterChromaShort(pu, shortYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);

                addWeightUni(pu, predYuv, shortYuv, wv0, bLuma, bChroma);
            }
            else
            {
                if (bLuma)
                    predInterLumaPixel(pu, predYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
                if (bChroma)
                    predInterChromaPixel(pu, predYuv, *cu.m_slice->m_refReconPicList[0][refIdx0], mv0);
            }
        }
        else
        {
            MV mv1 = cu.m_mv[1][pu.puAbsPartIdx];
            cu.clipMv(mv1);

            /* uniprediction to L1 */
            X265_CHECK(refIdx1 >= 0, "refidx1 was not positive\n");

            if (pwp1 && pwp1->wtPresent)
            {
                ShortYuv& shortYuv = m_predShortYuv[0];

                if (bLuma)
                    predInterLumaShort(pu, shortYuv, *cu.m_slice->m_refReconPicList[1][refIdx1], mv1);
                if (bChroma)
                    predInterChromaShort(pu, shortYuv, *cu.m_slice->m_refReconPicList[1][refIdx1], mv1);

                addWeightUni(pu, predYuv, shortYuv, wv0, bLuma, bChroma);
            }
            else
            {
                if (bLuma)
                    predInterLumaPixel(pu, predYuv, *cu.m_slice->m_refReconPicList[1][refIdx1], mv1);
                if (bChroma)
                    predInterChromaPixel(pu, predYuv, *cu.m_slice->m_refReconPicList[1][refIdx1], mv1);
            }
        }
    }
}