h264dsputil.c

君正早期ucos系统(只有早期的才不没有打包成库),MPLAYER,文件系统,图片解码,浏览,电子书,录音,想学ucos,识货的人就下吧 russblock fmradio explore set

      Q16ADD_SS_WW (xr1,  xr0, xr2,  xr0);    // xr1  = -xr2     -delta
      Q16ADD_SS_WW (xr11, xr0, xr12, xr0);    // xr11 = -xr12    -delta
      Q8ACCE_AA (xr2, xr0, xr10, xr12);    // (p0 + i_delta)
      Q8ACCE_AA (xr1, xr0, xr9,  xr11);    // (q0 - i_delta)
//
      Q16SAT    (xr2, xr2, xr12);
      Q16SAT    (xr1, xr1, xr11);      
      p0 = S32M2I (xr2);
      q0 = S32M2I (xr1);
//
      pix[-1] = p0;
      pix[0]  = q0;
      pix += ystride;
      pix[-1] = (p0 >> 8);
      pix[0]  = (q0 >> 8);
      pix += ystride;
      pix[-1] = (p0 >> 16);
      pix[0]  = (q0 >> 16);
      pix += ystride;
      pix[-1] = (p0 >> 24);
      pix[0]  = (q0 >> 24);
      pix += ystride;
 }
}

#else
static inline void h264_loop_filter_chroma_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta, int8_t *tc0)
{
    int i, d;
    for( i = 0; i < 4; i++ ) {
        const int tc = tc0[i];
        if( tc <= 0 ) {
            pix += 2*ystride;
            continue;
        }
        for( d = 0; d < 2; d++ ) {
            const int p0 = pix[-1*xstride];
            const int p1 = pix[-2*xstride];
            const int q0 = pix[0];
            const int q1 = pix[1*xstride];

            if( FFABS( p0 - q0 ) < alpha &&
                FFABS( p1 - p0 ) < beta &&
                FFABS( q1 - q0 ) < beta ) {

                int delta = av_clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );

                pix[-xstride] = av_clip_uint8( p0 + delta );    /* p0' */
                pix[0]        = av_clip_uint8( q0 - delta );    /* q0' */
            }
            pix += ystride;
        }
    }
}
static void h264_v_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
{
    h264_loop_filter_chroma_c(pix, stride, 1, alpha, beta, tc0);
}
static void h264_h_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
{
    h264_loop_filter_chroma_c(pix, 1, stride, alpha, beta, tc0);
}
#endif

#ifdef JZ4740_MXU_OPT
static void h264_v_loop_filter_chroma_intra_mxu(uint8_t *pix, int xstride, int alpha, int beta)
{
  int d;
  uint8_t *tpix;

  S32I2M (xr14, beta);
  S32SFL (xr0,  xr14, xr14, xr14, ptn0);
  S32SFL (xr0,  xr14, xr14, xr14, ptn3);   // xr14: beta
  S32I2M (xr13, alpha);
  S32SFL (xr0,  xr13, xr13, xr13, ptn0);
  S32SFL (xr0,  xr13, xr13, xr13, ptn3);   // xr13: alpha
  S32I2M (xr8, 0x01010101);

  for (d = 0; d < 2; d++)
  {
    tpix = pix - 2*xstride;
    S32LDD (xr11, tpix, 0);             // p1
    S32LDIV(xr10, tpix, xstride, 0);    // p0
    S32LDIV(xr9,  tpix, xstride, 0);    // q0
    S32LDIV(xr1,  tpix, xstride, 0);    // q1
    Q8ABD(xr3, xr10, xr9 );   // FFABS (p0 - q0)
    Q8ABD(xr4, xr11, xr10);   // FFABS (p1 - p0)
    Q8ABD(xr5, xr1,  xr9 );   // FFABS (q1 - q0)
// FFABS(p0 - q0) - alpha, FFABS(p1 - p0) - beta
    Q8ADDE_SS (xr6, xr3, xr13,  xr7);  // FFABS(p0 - q0) - alpha
    Q8ADDE_SS (xr3, xr4, xr14,  xr4);  // FFABS(p1 - p0) - beta
    Q16SLR (xr6, xr6, xr7, xr7, 15);   // 1: < 0 (FFABS( p0 - q0 ) < alpha)
    Q16SLR (xr3, xr3, xr4, xr4, 15);   // 1: < 0 (FFABS( p1 - p0 ) < beta)
    Q16SAT (xr4, xr3, xr4);   // xr4: 1: < 0 (FFABS( p1 - p0 ) < beta)
    Q16SAT (xr3, xr6, xr7);   // xr3: 1: < 0 (FFABS( p0 - q0 ) < alpha)
// FFABS(q1 - q0) - beta
    Q8ADDE_SS (xr6, xr5, xr14,  xr7);  // FFABS(q1 - q0) - beta
    Q16SLR (xr6, xr6, xr7, xr7, 15);   // 1: < 0 (FFABS( q1 - q0 ) < beta)
    Q16SAT (xr5, xr6, xr7);   // xr5: 1: < 0 (FFABS( q1 - q0 ) < beta)

// if( FFABS( p0 - q0 ) < alpha && ..)
    Q8MADL_AA (xr0, xr3, xr4, xr3);        // if( FFABS( p0 - q0 ) < alpha && ..)
//
    Q8ADDE_AA (xr12, xr11, xr8, xr7);      // (xr12, xr7): p1 + 1
    Q16SLL (xr12, xr12, xr7, xr7, 1);      // (xr12, xr7): 2*p1 + 2
    Q8ACCE_AA (xr12, xr10, xr1, xr7);      // (xr12, xr7): 2*p1 + p0 + q1 + 2
//
    Q8MADL_AA (xr0, xr3, xr5, xr3);        // if( FFABS( p0 - q0 ) < alpha && ... && ...)
//
    Q8ADDE_AA (xr2, xr1, xr8, xr6);        // (xr2,  xr6): q1 + 1
    Q16SLL (xr2, xr2, xr6, xr6, 1);        // (xr2,  xr6): 2*q1 + 2
    Q8ACCE_AA (xr2, xr9, xr11, xr6);       // (xr2,  xr6): 2*q1 + q0 + p1 + 2
//
    Q16SAR (xr12, xr12, xr7, xr7, 2);      // (xr12, xr7): (2*p1 + p0 + q1 + 2) >> 2, new p0
    S32SFL (xr0, xr12, xr7, xr12, ptn1);   // xr12: new p0
    Q8MADL_AA (xr0, xr3, xr12, xr12);      // xr12: new p0 mask
    Q16SAR (xr2,  xr2, xr6, xr6, 2);       // (xr2, xr6): (2*q1 + q0 + p1 + 2) >> 2, new  q0
    S32SFL (xr0, xr2, xr6, xr2, ptn1);   // xr12: new q0
    Q8MADL_AA (xr0, xr3, xr2, xr2);      // xr2: new q0 mask
//
    Q8ADD_SS  (xr4, xr8, xr3);             // xr4: ~(xr3)
    Q8MADL_AA (xr0, xr4, xr10, xr10);      // xr10: old p0 mask
    Q8MADL_AA (xr0, xr4, xr9,  xr9);       //  xr9:  old q0 mask
// store
    tpix = pix - xstride;
    Q8ADD_AA  (xr10, xr10, xr12);         // xr10: final p0'
    S32STD    (xr10, tpix, 0);            // pix[-xstride] = ...p0'
    Q8ADD_AA  (xr9,  xr9, xr2);           // xr9:  final q0'
    S32STDV   (xr9,  tpix, xstride, 0);   // pix[0] = ...q0'
    pix += 4;
  }
}

static void h264_h_loop_filter_chroma_intra_mxu(uint8_t *pix, int ystride, int alpha, int beta)
{
    int d;
    uint8_t *tpix;
    unsigned int p0, q0;

    S32I2M (xr14, beta);
    S32SFL (xr0,  xr14, xr14, xr14, ptn0);
    S32SFL (xr0,  xr14, xr14, xr14, ptn3);   // xr14: beta
    S32I2M (xr13, alpha);
    S32SFL (xr0,  xr13, xr13, xr13, ptn0);
    S32SFL (xr0,  xr13, xr13, xr13, ptn3);   // xr13: alpha
    S32I2M (xr8,  0x01010101);
    
    for( d = 0; d < 2; d++ ) {
      tpix = pix;
      S32LDD  (xr2, tpix,  0);          // xr2: q3,q2,q1,q0
      S32LDD  (xr1, tpix, -4);          // xr1: p0,p1,p2,p3
      S32LDIV (xr4, tpix, ystride, 0);  // xr4: q3',q2',q1',q0'
      S32LDD  (xr3, tpix, -4);          // xr3: p0',p1',p2',p3'
//
      S32SFL  (xr6, xr4, xr2, xr12, ptn0); // xr6: q3'q3q2'q2, xr12:q1'q1q0'q0
      S32SFL  (xr5, xr3, xr1, xr7, ptn0); // xr5: p0'p0p1'p1, xr7:p2'p2p3'p3
//
      S32LDIV (xr2, tpix, ystride, 0);    // xr2: q3",q2",q1",q0"
      S32LDD  (xr1, tpix, -4);          // xr1: p0",p1",p2",p3"
      S32LDIV (xr4, tpix, ystride, 0);    // xr4: q3^,q2^,q1^,q0^
      S32LDD  (xr3, tpix, -4);          // xr3: p0^,p1^,p2^,p3^
//
      S32SFL  (xr4, xr4, xr2, xr2, ptn0); // xr4: q3^q3"q2^q2", xr2:q1^q1"q0^q0"
      S32SFL  (xr3, xr3, xr1, xr1, ptn0); // xr3: p0^p0"p1^p1", xr1:p2^p2"p3^p3"
//
      S32SFL  (xr10, xr3, xr5, xr11, ptn3); // xr10: p0^p0"p0'p0, xr11:p1^p1"p1'p1
      S32SFL  (xr1,  xr2, xr12, xr9,  ptn3); // xr1: q1^q1"q1'q1, xr9:q0^q0"q0'q0
//
      Q8ABD(xr3, xr10, xr9 );   // FFABS (p0 - q0)
      Q8ABD(xr4, xr11, xr10);   // FFABS (p1 - p0)
      Q8ABD(xr5, xr1,  xr9 );   // FFABS (q1 - q0)
// FFABS(p0 - q0) - alpha, FFABS(p1 - p0) - beta
      Q8ADDE_SS (xr6, xr3, xr13,  xr7);  // FFABS(p0 - q0) - alpha
      Q8ADDE_SS (xr3, xr4, xr14,  xr4);  // FFABS(p1 - p0) - beta
      Q16SLR (xr6, xr6, xr7, xr7, 15);   // 1: < 0 (FFABS( p0 - q0 ) < alpha)
      Q16SLR (xr3, xr3, xr4, xr4, 15);   // 1: < 0 (FFABS( p1 - p0 ) < beta)
      Q16SAT (xr4, xr3, xr4);   // xr4: 1: < 0 (FFABS( p1 - p0 ) < beta)
      Q16SAT (xr3, xr6, xr7);   // xr3: 1: < 0 (FFABS( p0 - q0 ) < alpha)
// FFABS(q1 - q0) - beta
      Q8ADDE_SS (xr6, xr5, xr14,  xr7);  // FFABS(q1 - q0) - beta
      Q16SLR (xr6, xr6, xr7, xr7, 15);   // 1: < 0 (FFABS( q1 - q0 ) < beta)
      Q16SAT (xr5, xr6, xr7);   // xr5: 1: < 0 (FFABS( q1 - q0 ) < beta)

// if( FFABS( p0 - q0 ) < alpha && ..)
      Q8MADL_AA (xr0, xr3, xr4, xr3);        // if( FFABS( p0 - q0 ) < alpha && ..)
//
      Q8ADDE_AA (xr12, xr11, xr8, xr7);      // (xr12, xr7): p1 + 1
      Q16SLL (xr12, xr12, xr7, xr7, 1);      // (xr12, xr7): 2*p1 + 2
      Q8ACCE_AA (xr12, xr10, xr1, xr7);      // (xr12, xr7): 2*p1 + p0 + q1 + 2
//
      Q8MADL_AA (xr0, xr3, xr5, xr3);        // if( FFABS( p0 - q0 ) < alpha && ... && ...)//
      Q8ADDE_AA (xr2, xr1, xr8, xr6);        // (xr2,  xr6): q1 + 1
      Q16SLL (xr2, xr2, xr6, xr6, 1);        // (xr2,  xr6): 2*q1 + 2
      Q8ACCE_AA (xr2, xr9, xr11, xr6);       // (xr2,  xr6): 2*q1 + q0 + p1 + 2
//
      Q16SAR (xr12, xr12, xr7, xr7, 2);      // (xr12, xr7): (2*p1 + p0 + q1 + 2) >> 2, new p0
      S32SFL (xr0, xr12, xr7, xr12, ptn1);   // xr12: new p0
      Q8MADL_AA (xr0, xr3, xr12, xr12);      // xr12: new p0 mask
      Q16SAR (xr2,  xr2, xr6, xr6, 2);       // (xr2, xr6): (2*q1 + q0 + p1 + 2) >> 2, new  q0
      S32SFL (xr0, xr2, xr6, xr2, ptn1);   // xr12: new q0
      Q8MADL_AA (xr0, xr3, xr2, xr2);      // xr2: new q0 mask
//
      Q8ADD_SS  (xr4, xr8, xr3);             // xr4: ~(xr3)
      Q8MADL_AA (xr0, xr4, xr10, xr10);      // xr10: old p0 mask
      Q8MADL_AA (xr0, xr4, xr9,  xr9);       //  xr9:  old q0 mask
//
      Q8ADD_AA  (xr10, xr10, xr12);         // xr10: final p0'
      Q8ADD_AA  (xr9,  xr9, xr2);           // xr9:  final q0'
//
      p0 = S32M2I (xr10);
      q0 = S32M2I (xr9);
//
      pix[-1] = p0;
      pix[0]  = q0;
      pix += ystride;
      pix[-1] = (p0 >> 8);
      pix[0]  = (q0 >> 8);
      pix += ystride;
      pix[-1] = (p0 >> 16);
      pix[0]  = (q0 >> 16);
      pix += ystride;
      pix[-1] = (p0 >> 24);
      pix[0]  = (q0 >> 24);
      pix += ystride;
    }
}

#else

static inline void h264_loop_filter_chroma_intra_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta)
{
    int d;
    for( d = 0; d < 8; d++ ) {
        const int p0 = pix[-1*xstride];
        const int p1 = pix[-2*xstride];
        const int q0 = pix[0];
        const int q1 = pix[1*xstride];

        if( FFABS( p0 - q0 ) < alpha &&
            FFABS( p1 - p0 ) < beta &&
            FFABS( q1 - q0 ) < beta ) {

            pix[-xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;   /* p0' */
            pix[0]        = ( 2*q1 + q0 + p1 + 2 ) >> 2;   /* q0' */
        }
        pix += ystride;
    }
}
static void h264_v_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
{
    h264_loop_filter_chroma_intra_c(pix, stride, 1, alpha, beta);
}
static void h264_h_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
{
    h264_loop_filter_chroma_intra_c(pix, 1, stride, alpha, beta);
}
#endif

#if 1
#define H264_LOWPASS(OPNAME, OP, OP2) \
static av_unused void OPNAME ## h264_qpel2_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    const int h=2;\
    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
    int i;\
    for(i=0; i<h; i++)\
    {\
        OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\
        OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\
        dst+=dstStride;\
        src+=srcStride;\
    }\
}\
\
static av_unused void OPNAME ## h264_qpel2_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    const int w=2;\
    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
    int i;\
    for(i=0; i<w; i++)\
    {\
        const int srcB= src[-2*srcStride];\
        const int srcA= src[-1*srcStride];\
        const int src0= src[0 *srcStride];\
        const int src1= src[1 *srcStride];\
        const int src2= src[2 *srcStride];\
        const int src3= src[3 *srcStride];\
        const int src4= src[4 *srcStride];\
        OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\
        OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\
        dst++;\
        src++;\
    }\
}\
\
static av_unused void OPNAME ## h264_qpel2_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
    const int h=2;\
    const int w=2;\
    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
    int i;\
    src -= 2*srcStride;\
    for(i=0; i<h+5; i++)\
    {\
        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]);\
        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]);\
        tmp+=tmpStride;\
        src+=srcStride;\
    }\
    tmp -= tmpStride*(h+5-2);\
    for(i=0; i<w; i++)\
    {\
        const int tmpB= tmp[-2*tmpStride];\
        const int tmpA= tmp[-1*tmpStride];\
        const int tmp0= tmp[0 *tmpStride];\
        const int tmp1= tmp[1 *tmpStride];\
        const int tmp2= tmp[2 *tmpStride];\
        const int tmp3= tmp[3 *tmpStride];\
        const int tmp4= tmp[4 *tmpStride];\
        OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
        OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
        dst++;\
        tmp++;\
    }\
}\
static void OPNAME ## h264_qpel4_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\