h263_enc_frame.cpp

audio-video-codecs.rar语音编解码器

}

static void h263e_ME_FullSearch_SAD_16x16(Ipp8u *pCur, Ipp8u *pRef, Ipp32s step, Ipp32s xL, Ipp32s xR,
                                          Ipp32s yT, Ipp32s yB, Ipp32s *xPos, Ipp32s *yPos, Ipp32s *bestSAD, Ipp32s thrSAD_16x16)
{
    Ipp32s  cSAD, bSAD, i, j, jPos, iPos;

    jPos = *xPos;
    iPos = *yPos;
    bSAD = *bestSAD;
#ifdef ME_FULLSEARCH_RECT
    pRef += yT * step;
    for (i = yT; i <= yB; i ++) {
        for (j = xL; j <= xR; j ++) {
            ippiSAD16x16_8u32s(pCur, step, pRef+j, step, &cSAD, IPPVC_MC_APX_FF);
            if (cSAD < bSAD) {
                bSAD = cSAD;
                iPos = i;
                jPos = j;
#ifdef ME_USE_THRESHOLD
                if (cSAD <= thrSAD_16x16) goto L_01;
#endif
            }
        }
        pRef += step;
    }
#else   // involute search
    Ipp32s  nLoop, cLoop, lPos;
    i = iPos;
    j = jPos;
    pRef += iPos * step + jPos;
    nLoop = IPP_MAX(IPP_MAX(jPos - xL, xR - jPos), IPP_MAX(iPos - yT, yB - iPos));
    for (cLoop = 1; cLoop <= nLoop; cLoop ++) {
        i --;
        j --;
        pRef -= step + 1;
        for (lPos = 0; lPos < (cLoop << 3); lPos ++) {
            if (j >= xL && j <= xR && i >= yT && i <= yB) {
                ippiSAD16x16_8u32s(pCur, step, pRef, step, &cSAD, IPPVC_MC_APX_FF);
                if (cSAD < bSAD) {
                    bSAD = cSAD;
                    iPos = i;
                    jPos = j;
#ifdef ME_USE_THRESHOLD
                    if (cSAD <= thrSAD_16x16) goto L_01;
#endif
                }
            }
            if (lPos < (cLoop * 2)) {
                j ++;  pRef ++;
            } else if (lPos < (cLoop * 4)) {
                i ++;  pRef += step;
            } else if (lPos < (cLoop * 6)) {
                j --;  pRef --;
            } else {
                i --;  pRef -= step;
            }
        }
    }
#endif
#ifdef ME_USE_THRESHOLD
L_01:
#endif
    *bestSAD = bSAD;
    *xPos = jPos;
    *yPos = iPos;
}

static const Ipp32s bdJ[9] = {0, -2, -1, 0, 1, 2, 1, 0, -1},
                    bdI[9] = {0, 0, -1, -2, -1, 0, 1, 2, 1},
                    sdJ[5] = {0, -1, 0, 1, 0},
                    sdI[5] = {0, 0, -1, 0, 1};

static void h263e_ME_DiamondSearch_SAD_16x16(Ipp8u *pCur, Ipp8u *pRef, Ipp32s step, Ipp32s xL, Ipp32s xR,
                                             Ipp32s yT, Ipp32s yB, Ipp32s *xPos, Ipp32s *yPos, Ipp32s *bestSAD,
                                             Ipp32s *sadBuf, Ipp32s thrSAD_16x16)
{
    Ipp32s  cSAD, bSAD, i, j, k, l, iPos, jPos, swWidth, swHeight;

    jPos = *xPos;
    iPos = *yPos;
    bSAD = *bestSAD;
    swWidth = (xR - xL + 1);
    swHeight = (yB - yT + 1);
    ippsSet_8u((Ipp8u)-1, (Ipp8u*)sadBuf, swWidth * swHeight * 4);
    sadBuf[(iPos-yT)*swWidth+jPos-xL] = bSAD;
    for (;;) {
        // find SAD at big diamond
        l = 0;
        for (k = 1; k < 9; k ++) {
            j = jPos + bdJ[k];
            i = iPos + bdI[k];
            if (j >= xL && j <= xR && i >= yT && i <= yB) {
                if (sadBuf[(i-yT)*swWidth+j-xL] == -1) {
                    ippiSAD16x16_8u32s(pCur, step, pRef+i*step+j, step, &cSAD, IPPVC_MC_APX_FF);
#ifdef ME_USE_THRESHOLD
                    if (cSAD <= thrSAD_16x16) {
                        *xPos = j;
                        *yPos = i;
                        *bestSAD = cSAD;
                        return;
                    }
#endif
                    sadBuf[(i-yT)*swWidth+j-xL] = cSAD;
                    if (cSAD < bSAD) {
                        l = k;
                        bSAD = cSAD;
                    }
                }
            }
        }
        if (l == 0) {
            // find SAD at small diamond
            for (k = 1; k <= 4; k ++) {
                j = jPos + sdJ[k];
                i = iPos + sdI[k];
                if (j >= xL && j <= xR && i >= yT && i <= yB) {
                    if (sadBuf[(i-yT)*swWidth+j-xL] == -1) {
                        ippiSAD16x16_8u32s(pCur, step, pRef+i*step+j, step, &cSAD, IPPVC_MC_APX_FF);
#ifdef ME_USE_THRESHOLD
                        if (cSAD <= thrSAD_16x16) {
                            *xPos = j;
                            *yPos = i;
                            *bestSAD = cSAD;
                            return;
                        }
#endif
                        sadBuf[(i-yT)*swWidth+j-xL]= cSAD;
                        if (cSAD < bSAD) {
                            l = k;
                            bSAD = cSAD;
                        }
                    }
                }
            }
            *xPos = jPos + sdJ[l];
            *yPos = iPos + sdI[l];
            *bestSAD = bSAD;
            return;
        }
        iPos += bdI[l];
        jPos += bdJ[l];
    }
}

static void h263e_ME_HalfPel_SAD_16x16(Ipp8u *pCur, Ipp8u *pRef, Ipp32s step, Ipp32s xL, Ipp32s xR,
                                       Ipp32s yT, Ipp32s yB, Ipp32s xPos, Ipp32s yPos, Ipp32s *bestSAD,
                                       IppMotionVector *mv, Ipp32s thrSAD_16x16)
{
    Ipp32s  cSAD, bSAD, i, j;

    i = 0;  j = 0;  bSAD = *bestSAD;
    if (xPos > xL && yPos > yT) {
        ippiSAD16x16_8u32s(pCur, step, pRef+(yPos-1)*step+xPos-1, step, &cSAD, IPPVC_MC_APX_HH);
        if (cSAD < bSAD) {
            i = -1;  j = -1;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (yPos > yT) {
        ippiSAD16x16_8u32s(pCur, step, pRef+(yPos-1)*step+xPos, step, &cSAD, IPPVC_MC_APX_FH);
        if (cSAD < bSAD) {
            i = -1;  j =  0;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (xPos < xR && yPos > yT) {
        ippiSAD16x16_8u32s(pCur, step, pRef+(yPos-1)*step+xPos, step, &cSAD, IPPVC_MC_APX_HH);
        if (cSAD < bSAD) {
            i = -1;  j =  1;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (xPos > xL) {
        ippiSAD16x16_8u32s(pCur, step, pRef+yPos*step+xPos-1, step, &cSAD, IPPVC_MC_APX_HF);
        if (cSAD < bSAD) {
            i =  0;  j = -1;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (xPos < xR) {
        ippiSAD16x16_8u32s(pCur, step, pRef+yPos*step+xPos, step, &cSAD, IPPVC_MC_APX_HF);
        if (cSAD < bSAD) {
            i =  0;  j =  1;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (xPos > xL && yPos < yB) {
        ippiSAD16x16_8u32s(pCur, step, pRef+yPos*step+xPos-1, step, &cSAD, IPPVC_MC_APX_HH);
        if (cSAD < bSAD) {
            i =  1;  j = -1;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (yPos < yB) {
        ippiSAD16x16_8u32s(pCur, step, pRef+yPos*step+xPos, step, &cSAD, IPPVC_MC_APX_FH);
        if (cSAD < bSAD) {
            i =  1;  j =  0;  bSAD = cSAD;
#ifdef ME_USE_THRESHOLD
            if (bSAD <= thrSAD_16x16) goto L_01;
#endif
        }
    }
    if (xPos < xR && yPos < yB) {
        ippiSAD16x16_8u32s(pCur, step, pRef+yPos*step+xPos, step, &cSAD, IPPVC_MC_APX_HH);
        if (cSAD < bSAD) {
            i =  1;  j =  1;  bSAD = cSAD;
        }
    }
#ifdef ME_USE_THRESHOLD
L_01:
#endif
    mv->dx = (Ipp16s)((xPos << 1) + j);
    mv->dy = (Ipp16s)((yPos << 1) + i);
    *bestSAD = bSAD;
}

static void h263e_ME_HalfPel_SAD_Fast_16x16(Ipp8u *pCur, Ipp8u *pRef, Ipp32s step, Ipp32s xL, Ipp32s xR,
                                            Ipp32s yT, Ipp32s yB, Ipp32s xPos, Ipp32s yPos, Ipp32s *bestSAD,
                                            IppMotionVector *mv, Ipp32s *sadBuff)
{
    Ipp32s  i, j, swWidth, swHeight;
    Ipp32s  m0, m1, m2, m3, m4;

    m0 = *bestSAD;
    swWidth = (xR - xL + 1);
    swHeight = (yB - yT + 1);
    j = xPos;
    i = yPos - 1;
    if (j >= xL && j <= xR && i >= yT && i <= yB) {
        m1 = sadBuff[(i-yT)*swWidth+j-xL];
        if (m1 == -1)
            ippiSAD16x16_8u32s(pCur, step, pRef+i*step+j, step, &m1, IPPVC_MC_APX_FF);
    } else
        m1 = 1000000;
    j = xPos;
    i = yPos + 1;
    if (j >= xL && j <= xR && i >= yT && i <= yB) {
        m2 = sadBuff[(i-yT)*swWidth+j-xL];
        if (m2 == -1)
            ippiSAD16x16_8u32s(pCur, step, pRef+i*step+j, step, &m2, IPPVC_MC_APX_FF);
    } else
        m2 = 1000000;
    j = xPos - 1;
    i = yPos;
    if (j >= xL && j <= xR && i >= yT && i <= yB) {
        m3 = sadBuff[(i-yT)*swWidth+j-xL];
        if (m3 == -1)
            ippiSAD16x16_8u32s(pCur, step, pRef+i*step+j, step, &m3, IPPVC_MC_APX_FF);
    } else
        m3 = 1000000;
    j = xPos + 1;
    i = yPos;
    if (j >= xL && j <= xR && i >= yT && i <= yB) {
        m4 = sadBuff[(i-yT)*swWidth+j-xL];
        if (m4 == -1)
            ippiSAD16x16_8u32s(pCur, step, pRef+i*step+j, step, &m4, IPPVC_MC_APX_FF);
    } else
        m4 = 1000000;
    if (2*(m3-m0) < (m4-m0)) {
        j = -1;
    } else if ((m3-m0) > 2*(m4-m0)) {
        j = +1;
    } else
        j = 0;
    if (2*(m1-m0) < (m2-m0)) {
        i = -1;
    } else if ((m1-m0) > 2*(m2-m0)) {
        i = +1;
    } else
        i = 0;
    mv->dx = (Ipp16s)((xPos << 1) + j);
    mv->dy = (Ipp16s)((yPos << 1) + i);
    if (j != 0 || i != 0) {
        ippiSAD16x16_8u32s(pCur, step, pRef+h263e_MC_offs(mv, step), step, bestSAD, h263e_MC_type(mv));
        if (*bestSAD > m0) {
            // false choosing
            mv->dx = (Ipp16s)(xPos << 1);
            mv->dy = (Ipp16s)(yPos << 1);
            *bestSAD = m0;
        }
    }
}

void ippVideoEncoderH263::ME_SAD_16x16(Ipp8u *pCur, Ipp8u *pRef, Ipp32s step, Ipp32s xL, Ipp32s xR,
                                       Ipp32s yT, Ipp32s yB, Ipp32s *bestSAD, IppMotionVector *mv)
{
    Ipp32s  xPos, yPos, sadL;

    // mv - Predicted MV
    // bestSAD is the SAD at (0,0)
#ifdef ME_USE_THRESHOLD
    if (*bestSAD <= mMEthrSAD16x16) {
        mv->dx = mv->dy = 0;
        return;
    }
#endif
    xPos = h263e_Div2(mv->dx);
    yPos = h263e_Div2(mv->dy);
    // check init mv is in rect
    if (xPos < xL || xPos > xR || yPos < yT || yPos > yB) {
        // reset to (0,0)
        xPos = yPos = 0;
    } else {
        // calc SAD at Predict (intpel acc)
        if (xPos != 0 || yPos != 0) {
            ippiSAD16x16_8u32s(pCur, mStepLuma, pRef+yPos*mStepLuma+xPos, mStepLuma, &sadL, IPPVC_MC_APX_FF);
            if (sadL - SAD_FAVOR_PRED < *bestSAD) {
                // choose Predict as origin for ME
                *bestSAD = sadL;
            } else {
                // choose (0,0) as origin for ME
                xPos = yPos = 0;
            }
        }
    }

    xL = IPP_MAX(xL, xPos - mPPicsearchHor);
    xR = IPP_MIN(xR, xPos + mPPicsearchHor);
    yT = IPP_MAX(yT, yPos - mPPicsearchVer);
    yB = IPP_MIN(yB, yPos + mPPicsearchVer);

    *bestSAD -= SAD_FAVOR_ZERO;
    sadL = *bestSAD;
#ifdef ME_USE_THRESHOLD
    if (*bestSAD > mMEthrSAD16x16) {
#endif
        if (mMEalgorithm == 0)
            h263e_ME_FullSearch_SAD_16x16(pCur, pRef, step, xL, xR, yT, yB, &xPos, &yPos, bestSAD, mMEthrSAD16x16);
        else
            h263e_ME_DiamondSearch_SAD_16x16(pCur, pRef, step, xL, xR, yT, yB, &xPos, &yPos, bestSAD, mMEfastSAD, mMEthrSAD16x16);
#ifdef ME_USE_THRESHOLD
    }
    if (mMEaccuracy == 1 || *bestSAD <= mMEthrSAD16x16) {
        mv->dx = (Ipp16s)(xPos << 1);
        mv->dy = (Ipp16s)(yPos << 1);
        if (mMEaccuracy > 2) {
            // quarter pel
            mv->dx = (Ipp16s)(mv->dx << 1);
            mv->dy = (Ipp16s)(mv->dy << 1);
        }
#else
    if (mMEaccuracy == 1) {
        mv->dx = (Ipp16s)(xPos << 1);
        mv->dy = (Ipp16s)(yPos << 1);
#endif
    } else {
        if (mMEalgorithm == 1 && mMEfastHP)
            h263e_ME_HalfPel_SAD_Fast_16x16(pCur, pRef, step, xL, xR, yT, yB, xPos, yPos, bestSAD, mv, mMEfastSAD);
        else
            h263e_ME_HalfPel_SAD_16x16(pCur, pRef, step, xL, xR, yT, yB, xPos, yPos, bestSAD, mv, mMEthrSAD16x16);
    }
    if (*bestSAD == sadL)
        *bestSAD += SAD_FAVOR_ZERO;
}

static void h263e_ME_HalfPel_SAD_8x8(Ipp8u *pCur, Ipp8u *pRef, Ipp32s step, Ipp32s xL, Ipp32s xR,
                                     Ipp32s yT, Ipp32s yB, Ipp32s xPos, Ipp32s yPos, Ipp32s *bestSAD,
                                     IppMotionVector *mv, Ipp32s thrSAD_8x8)
{
    Ipp32s  cSAD, bSAD, i, j;

    i = 0;  j = 0;  bSAD = *bestSAD;
    if (xPos > xL && yPos > yT) {
        ippiSAD8x8_8u32s_C1R(pCur, step, pRef+(yPos-1)*step+xPos-1, step, &cSAD, IPPVC_MC_APX_HH);
        if (cSAD < bSAD) {
            i = -1;  j = -1;  bSAD = cSAD;