dsputil.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

        dst[j] = (2731*(3*src[j] + 2*src[j+1] + 4*src[j+stride] + 3*src[j+stride+1] + 6)) >> 15;
      }
      src += stride;
      dst += stride;
    }
}

static inline void put_tpel_pixels_mc02_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (683*(src[j] + 2*src[j+stride] + 1)) >> 11;
      }
      src += stride;
      dst += stride;
    }
}

static inline void put_tpel_pixels_mc21_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (2731*(3*src[j] + 4*src[j+1] + 2*src[j+stride] + 3*src[j+stride+1] + 6)) >> 15;
      }
      src += stride;
      dst += stride;
    }
}

static inline void put_tpel_pixels_mc22_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (2731*(2*src[j] + 3*src[j+1] + 3*src[j+stride] + 4*src[j+stride+1] + 6)) >> 15;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    switch(width){
    case 2: avg_pixels2_c (dst, src, stride, height); break;
    case 4: avg_pixels4_c (dst, src, stride, height); break;
    case 8: avg_pixels8_c (dst, src, stride, height); break;
    case 16:avg_pixels16_c(dst, src, stride, height); break;
    }
}

static inline void avg_tpel_pixels_mc10_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((683*(2*src[j] + src[j+1] + 1)) >> 11) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc20_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((683*(src[j] + 2*src[j+1] + 1)) >> 11) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc01_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((683*(2*src[j] + src[j+stride] + 1)) >> 11) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc11_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((2731*(4*src[j] + 3*src[j+1] + 3*src[j+stride] + 2*src[j+stride+1] + 6)) >> 15) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc12_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((2731*(3*src[j] + 2*src[j+1] + 4*src[j+stride] + 3*src[j+stride+1] + 6)) >> 15) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc02_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((683*(src[j] + 2*src[j+stride] + 1)) >> 11) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc21_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((2731*(3*src[j] + 4*src[j+1] + 2*src[j+stride] + 3*src[j+stride+1] + 6)) >> 15) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}

static inline void avg_tpel_pixels_mc22_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
    int i,j;
    for (i=0; i < height; i++) {
      for (j=0; j < width; j++) {
        dst[j] = (dst[j] + ((2731*(2*src[j] + 3*src[j+1] + 3*src[j+stride] + 4*src[j+stride+1] + 6)) >> 15) + 1) >> 1;
      }
      src += stride;
      dst += stride;
    }
}
#ifdef JZ4740_MXU_OPT
static void put_mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
  int i;
  uint32_t src_aln0,src_rs0;
  int mul_ct0,mul_ct1,mul_ct2,mul_ct3;
  int ct0;
  mul_ct0=0x14141414;   //20 20 20 20
  mul_ct1=0x06060606;   // 6  6  6  6
  mul_ct2=0x03030303;   // 3  3  3  3
  mul_ct3=0x01010101;   // 1  1  1  1
      ct0=0x00100010;   //   16    16
  src_aln0 = (((uint32_t)src) & 0xFFFFFFFC);
  src_rs0 = 4-(((uint32_t)src) & 3);
  S32I2M(xr15,mul_ct0); //xr15:20 20 20 20
  S32I2M(xr14,mul_ct1); //xr14:6  6  6  6
  S32I2M(xr13,mul_ct2); //xr13:3  3  3  3
  S32I2M(xr12,mul_ct3); //xr12:1  1  1  1
  S32I2M(xr11,    ct0); //xr11:  16    16
  for(i=0; i<h; i++)
  {
  S32LDD(xr1,src_aln0,0);
  S32LDD(xr2,src_aln0,4);
  S32LDD(xr3,src_aln0,8);
  S32ALN(xr1,xr2,xr1,src_rs0);  //xr1:src[3] src[2] src[1] src[0]
  S32ALN(xr2,xr3,xr2,src_rs0);  //xr2:src[7] src[6] src[5] src[4]  
  
  S32ALN(xr5,xr1,xr1,1); //xr5:src[2] src[1] src[0] src[3]
  S32ALN(xr6,xr1,xr1,2); //xr6:src[1] src[0] src[3] src[2]
  S32ALN(xr7,xr1,xr1,3); //xr7:src[0] src[3] src[2] src[1]
  S32ALN(xr8,xr1,xr7,1); //xr8:src[2] src[1] src[0] src[0]
  S32ALN(xr9,xr8,xr6,1); //xr9:src[1] src[0] src[0] src[1]
  S32ALN(xr10,xr9,xr8,1);//xr10: src[0] src[0] src[1] src[2]
  
  Q8MUL(xr5,xr1,xr15,xr4); //xr5:src[3]*20 src[2]*20 xr4:src[1]*20 src[0]*20
  Q8MAC_SS(xr5,xr8,xr14,xr4);//xr5:src[3]*20-src[2]*6 src[2]*20-src[1]*6 
                             //xr4:src[1]*20-src[0]*6 src[0]*20-src[0]*6
  Q8MAC_AA(xr5,xr9,xr13,xr4);//xr5:xr3+src[1]*3 xr3+src[0]*3 xr4:xr4+src[0]*3 xr4+src[1]*3
  Q8MAC_SS(xr5,xr10,xr12,xr4);//xr5:xr3-src[0]  xr3-src[0]  xr4:xr4-src[1] xr4-src[2]
  
  S32ALN(xr8,xr2,xr1,3);     //xr8:src[4] src[3] src[2] src[1]
  S32ALN(xr9,xr2,xr1,2);     //xr9:src[5] src[4] src[3] src[2]
  S32ALN(xr10,xr2,xr1,1);    //xr10:src[6]src[5] src[4] src[3]
  
  Q8MAC_AA(xr5,xr8,xr15,xr4);//xr5:xr3+src[4]*20 xr3+src[3]*20 xr4:xr4+src[2]*20 xr4+src[1]*20
  Q8MAC_SS(xr5,xr9,xr14,xr4);//xr5:xr3-src[5]*6  xr3-src[4]*6  xr4:xr4-src[3]*6  xr4-src[2]*6
  Q8MAC_AA(xr5,xr10,xr13,xr4);
  Q8MAC_SS(xr5,xr2,xr12,xr4);

  Q16ACC_AA(xr5,xr11,xr0,xr4); 
  Q16SAR(xr5,xr5,xr4,xr4,5);
  Q16SAT(xr5,xr5,xr4);        //xr5:dst[3] dst[2] dst[1] dst[0]
 
  ///////////////////////////////////////////
  Q8MUL(xr1,xr2,xr15,xr4);    //xr1:src[7]*20 src[6]*20 xr4:src[5]*20 src[4]*20
  Q8MAC_SS(xr1,xr10,xr14,xr4);
  Q8MAC_AA(xr1,xr9 ,xr13,xr4);
  Q8MAC_SS(xr1,xr8 ,xr12,xr4);
  
  S32LDD(xr6,src_aln0,0xc);   
  S32ALN(xr3,xr6,xr3,src_rs0); //xr3:src[11] src[10] src[9] src[8]

  S32ALN(xr7,xr3,xr2,1); //xr7:src[10] src[9]  src[8] src[7]
  S32ALN(xr8,xr3,xr2,2); //xr8:src[9]  src[8]  src[7] src[6]
  S32ALN(xr9,xr3,xr2,3); //xr9:src[8]  src[7]  src[6] src[5]
  Q8MAC_AA(xr1,xr9,xr15,xr4);
  Q8MAC_SS(xr1,xr8,xr14,xr4);
  Q8MAC_AA(xr1,xr7,xr13,xr4);
  Q8MAC_SS(xr1,xr3,xr12,xr4);
  Q16ACC_AA(xr1,xr11,xr0,xr4);
  Q16SAR(xr1,xr1,xr4,xr4,5);
  Q16SAT(xr1,xr1,xr4);        //xr1:dst[7] dst[6] dst[5] dst[4]
  ///////////////////////////////////////////  
  Q8MUL(xr4,xr3,xr15,xr2);
  Q8MAC_SS(xr4,xr7,xr14,xr2);
  Q8MAC_AA(xr4,xr8,xr13,xr2);
  Q8MAC_SS(xr4,xr9,xr12,xr2);
  S32LDD(xr10,src_aln0,0x10); 
  S32ALN(xr6,xr10,xr6,src_rs0); //xr6:src[15] src[14] src[13] src[12]
  
  S32ALN(xr7,xr6,xr3,1);//xr7 :src[14] src[13] src[12] src[11]
  S32ALN(xr8,xr6,xr3,2);//xr8 :src[13] src[12] src[11] src[10]
  S32ALN(xr9,xr6,xr3,3);//xr9 :src[12] src[11] src[10] src[9]
  Q8MAC_AA(xr4,xr7,xr13,xr2);
  Q8MAC_SS(xr4,xr8,xr14,xr2);
  Q8MAC_SS(xr4,xr6,xr12,xr2);
  Q8MAC_AA(xr4,xr9,xr15,xr2);

  Q16ACC_AA(xr4,xr11,xr0,xr2);
  Q16SAR(xr4,xr4,xr2,xr2,5);
  Q16SAT(xr4,xr4,xr2);       //xr4:dst[11] dst[10] dst[9] dst[8]
  //////////////////////////////////////////////////
  Q8MUL(xr2,xr6,xr15,xr3);  //xr2:src[15]*20 src[14]*20 xr3:src[13]*20 src[12]*20
  Q8MAC_SS(xr2,xr7,xr14,xr3);
  Q8MAC_AA(xr2,xr8,xr13,xr3);
  Q8MAC_SS(xr2,xr9,xr12,xr3);
  S32LDD(xr7,src_aln0,0x14); 
  S32ALN(xr10,xr7,xr10,src_rs0);//xr10:src[19] src[18] src[17] src[16]
  S32ALN(xr7,xr10,xr6,3); //xr7:src[16]src[15] src[14] src[13]
  S32ALN(xr8,xr10,xr7, 3); //xr8:src[16]src[16]src[15] src[14]
  S32ALN(xr9,xr6,xr6, 1); //xr9:src[14] src[13] src[12] src[15]
  Q8MAC_AA(xr2,xr7,xr15,xr3);
  Q8MAC_SS(xr2,xr8,xr14,xr3);
  S32ALN(xr7,xr9,xr8,3);     //xr7:src[15] src[16] src[16] src[15] 
  S32ALN(xr8,xr8,xr7,3);     //xr8:src[14] src[15] src[16] src[16]
  Q8MAC_AA(xr2,xr7,xr13,xr3);
  Q8MAC_SS(xr2,xr8,xr12,xr3);
  Q16ACC_AA(xr2,xr11,xr0,xr3);
  Q16SAR(xr2,xr2,xr3,xr3,5);
  Q16SAT(xr2,xr2,xr3);      //xr2:dst[15] dst[14] dst[13] dst[12]  
  S32STD(xr5,dst,0x0);
  S32STD(xr1,dst,0x4);
  S32STD(xr4,dst,0x8);
  S32STD(xr2,dst,0xc);
  dst += dstStride;
  src_aln0 +=srcStride;
}
}
#else
static void put_mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
  uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
  int i;
  for(i=0; i<h; i++)
   {
   dst[ 0]=cm[(((src[ 0]+src[ 1])*20 - (src[ 0]+src[ 2])*6 + (src[ 1]+src[ 3])*3 - (src[ 2]+src[ 4]))+16)>>5];
   dst[ 1]=cm[(((src[ 1]+src[ 2])*20 - (src[ 0]+src[ 3])*6 + (src[ 0]+src[ 4])*3 - (src[ 1]+src[ 5]))+16)>>5];
   dst[ 2]=cm[(((src[ 2]+src[ 3])*20 - (src[ 1]+src[ 4])*6 + (src[ 0]+src[ 5])*3 - (src[ 0]+src[ 6]))+16)>>5];
   dst[ 3]=cm[(((src[ 3]+src[ 4])*20 - (src[ 2]+src[ 5])*6 + (src[ 1]+src[ 6])*3 - (src[ 0]+src[ 7]))+16)>>5];
   dst[ 4]=cm[(((src[ 4]+src[ 5])*20 - (src[ 3]+src[ 6])*6 + (src[ 2]+src[ 7])*3 - (src[ 1]+src[ 8]))+16)>>5];
   dst[ 5]=cm[(((src[ 5]+src[ 6])*20 - (src[ 4]+src[ 7])*6 + (src[ 3]+src[ 8])*3 - (src[ 2]+src[ 9]))+16)>>5];
   dst[ 6]=cm[(((src[ 6]+src[ 7])*20 - (src[ 5]+src[ 8])*6 + (src[ 4]+src[ 9])*3 - (src[ 3]+src[10]))+16)>>5];
   dst[ 7]=cm[(((src[ 7]+src[ 8])*20 - (src[ 6]+src[ 9])*6 + (src[ 5]+src[10])*3 - (src[ 4]+src[11]))+16)>>5];
   dst[ 8]=cm[(((src[ 8]+src[ 9])*20 - (src[ 7]+src[10])*6 + (src[ 6]+src[11])*3 - (src[ 5]+src[12]))+16)>>5];
   dst[ 9]=cm[(((src[ 9]+src[10])*20 - (src[ 8]+src[11])*6 + (src[ 7]+src[12])*3 - (src[ 6]+src[13]))+16)>>5];
   dst[10]=cm[(((src[10]+src[11])*20 - (src[ 9]+src[12])*6 + (src[ 8]+src[13])*3 - (src[ 7]+src[14]))+16)>>5];
   dst[11]=cm[(((src[11]+src[12])*20 - (src[10]+src[13])*6 + (src[ 9]+src[14])*3 - (src[ 8]+src[15]))+16)>>5];
   dst[12]=cm[(((src[12]+src[13])*20 - (src[11]+src[14])*6 + (src[10]+src[15])*3 - (src[ 9]+src[16]))+16)>>5];
   dst[13]=cm[(((src[13]+src[14])*20 - (src[12]+src[15])*6 + (src[11]+src[16])*3 - (src[10]+src[16]))+16)>>5];
   dst[14]=cm[(((src[14]+src[15])*20 - (src[13]+src[16])*6 + (src[12]+src[16])*3 - (src[11]+src[15]))+16)>>5];
   dst[15]=cm[(((src[15]+src[16])*20 - (src[14]+src[16])*6 + (src[13]+src[15])*3 - (src[12]+src[14]))+16)>>5];
   dst+=dstStride;
   src+=srcStride;
   }
}
#endif
static void put_no_rnd_mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
   uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
   int i;    
   for(i=0; i<h; i++)
   {
   dst[ 0]=cm[(((src[ 0]+src[ 1])*20 - (src[ 0]+src[ 2])*6 + (src[ 1]+src[ 3])*3 - (src[ 2]+src[ 4]))+15)>>5];
   dst[ 1]=cm[(((src[ 1]+src[ 2])*20 - (src[ 0]+src[ 3])*6 + (src[ 0]+src[ 4])*3 - (src[ 1]+src[ 5]))+15)>>5];
   dst[ 2]=cm[(((src[ 2]+src[ 3])*20 - (src[ 1]+src[ 4])*6 + (src[ 0]+src[ 5])*3 - (src[ 0]+src[ 6]))+15)>>5];
   dst[ 3]=cm[(((src[ 3]+src[ 4])*20 - (src[ 2]+src[ 5])*6 + (src[ 1]+src[ 6])*3 - (src[ 0]+src[ 7]))+15)>>5];
   dst[ 4]=cm[(((src[ 4]+src[ 5])*20 - (src[ 3]+src[ 6])*6 + (src[ 2]+src[ 7])*3 - (src[ 1]+src[ 8]))+15)>>5];
   dst[ 5]=cm[(((src[ 5]+src[ 6])*20 - (src[ 4]+src[ 7])*6 + (src[ 3]+src[ 8])*3 - (src[ 2]+src[ 9]))+15)>>5];
   dst[ 6]=cm[(((src[ 6]+src[ 7])*20 - (src[ 5]+src[ 8])*6 + (src[ 4]+src[ 9])*3 - (src[ 3]+src[10]))+15)>>5];
   dst[ 7]=cm[(((src[ 7]+src[ 8])*20 - (src[ 6]+src[ 9])*6 + (src[ 5]+src[10])*3 - (src[ 4]+src[11]))+15)>>5];
   dst[ 8]=cm[(((src[ 8]+src[ 9])*20 - (src[ 7]+src[10])*6 + (src[ 6]+src[11])*3 - (src[ 5]+src[12]))+15)>>5];
   dst[ 9]=cm[(((src[ 9]+src[10])*20 - (src[ 8]+src[11])*6 + (src[ 7]+src[12])*3 - (src[ 6]+src[13]))+15)>>5];
   dst[10]=cm[(((src[10]+src[11])*20 - (src[ 9]+src[12])*6 + (src[ 8]+src[13])*3 - (src[ 7]+src[14]))+15)>>5];
   dst[11]=cm[(((src[11]+src[12])*20 - (src[10]+src[13])*6 + (src[ 9]+src[14])*3 - (src[ 8]+src[15]))+15)>>5];
   dst[12]=cm[(((src[12]+src[13])*20 - (src[11]+src[14])*6 + (src[10]+src[15])*3 - (src[ 9]+src[16]))+15)>>5];
   dst[13]=cm[(((src[13]+src[14])*20 - (src[12]+src[15])*6 + (src[11]+src[16])*3 - (src[10]+src[16]))+15)>>5];
   dst[14]=cm[(((src[14]+src[15])*20 - (src[13]+src[16])*6 + (src[12]+src[16])*3 - (src[11]+src[15]))+15)>>5];
   dst[15]=cm[(((src[15]+src[16])*20 - (src[14]+src[16])*6 + (src[13]+src[15])*3 - (src[12]+src[14]))+15)>>5];
   dst+=dstStride;
   src+=srcStride;
   }
}

static void avg_mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
{
   uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
   int i;
   for(i=0; i<h; i++)
   {
   dst[ 0]=(dst[0]+cm[(((src[ 0]+src[ 1])*20 - (src[ 0]+src[ 2])*6 + (src[ 1]+src[ 3])*3 - (src[ 2]+src[ 4]))+16)>>5]+1) >>1;
   dst[ 1]=(dst[1]+cm[(((src[ 1]+src[ 2])*20 - (src[ 0]+src[ 3])*6 + (src[ 0]+src[ 4])*3 - (src[ 1]+src[ 5]))+16)>>5]+1) >>1;
   dst[ 2]=(dst[2]+cm[(((src[ 2]+src[ 3])*20 - (src[ 1]+src[ 4])*6 + (src[ 0]+src[ 5])*3 - (src[ 0]+src[ 6]))+16)>>5]+1) >>1;
   dst[ 3]=(dst[3]+cm[(((src[ 3]+src[ 4])*20 - (src[ 2]+src[ 5])*6 + (src[ 1]+src[ 6])*3 - (src[ 0]+src[ 7]))+16)>>5]+1) >>1;
   dst[ 4]=(dst[4]+cm[(((src[ 4]+src[ 5])*20 - (src[ 3]+src[ 6])*6 + (src[ 2]+src[ 7])*3 - (src[ 1]+src[ 8]))+16)>>5]+1) >>1;
   dst[ 5]=(dst[5]+cm[(((src[ 5]+src[ 6])*20 - (src[ 4]+src[ 7])*6 + (src[ 3]+src[ 8])*3 - (src[ 2]+src[ 9]))+16)>>5]+1) >>1;
   dst[ 6]=(dst[6]+cm[(((src[ 6]+src[ 7])*20 - (src[ 5]+src[ 8])*6 + (src[ 4]+src[ 9])*3 - (src[ 3]+src[10]))+16)>>5]+1) >>1;
   dst[ 7]=(dst[7]+cm[(((src[ 7]+src[ 8])*20 - (src[ 6]+src[ 9])*6 + (src[ 5]+src[10])*3 - (src[ 4]+src[11]))+16)>>5]+1) >>1;
   dst[ 8]=(dst[8]+cm[(((src[ 8]+src[ 9])*20 - (src[ 7]+src[10])*6 + (src[ 6]+src[11])*3 - (src[ 5]+src[12]))+16)>>5]+1) >>1;
   dst[ 9]=(dst[9]+cm[(((src[ 9]+src[10])*20 - (src[ 8]+src[11])*6 + (src[ 7]+src[12])*3 - (src[ 6]+src[13]))+16)>>5]+1) >>1;
   dst[10]=(dst[10]+cm[(((src[10]+src[11])*20 - (src[ 9]+src[12])*6 + (src[ 8]+src[13])*3 - (src[ 7]+src[14]))+16)>>5]+1) >>1;
   dst[11]=(dst[11]+cm[(((src[11]+src[12])*20 - (src[10]+src[13])*6 + (src[ 9]+src[14])*3 - (src[ 8]+src[15]))+16)>>5]+1) >>1;
   dst[12]=(dst[12]+cm[(((src[12]+src[13])*20 - (src[11]+src[14])*6 + (src[10]+src[15])*3 - (src[ 9]+src[16]))+16)>>5]+1) >>1;
   dst[13]=(dst[13]+cm[(((src[13]+src[14])*20 - (src[12]+src[15])*6 + (src[11]+src[16])*3 - (src[10]+src[16]))+16)>>5]+1) >>1;
   dst[14]=(dst[14]+cm[(((src[14]+src[15])*20 - (src[13]+src[16])*6 + (src[12]+src[16])*3 - (src[11]+src[15]))+16)>>5]+1) >>1;
   dst[15]=(dst[15]+cm[(((src[15]+src[16])*20 - (src[14]+src[16])*6 + (src[13]+src[15])*3 - (src[12]+src[14]))+16)>>5]+1) >>1;
   dst+=dstStride;
   src+=srcStride;
   }
}
#ifdef JZ4740_MXU_OPT
static void put_mpeg4_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){
  int i,j;
  uint32_t src_aln0,src_rs0,src_aln,dst_aln,dst_aln1;
  int mul_ct0,mul_ct1,mul_ct2,mul_ct3;
  int ct0;
  mul_ct0=0x14141414;   //20 20 20 20
  mul_ct1=0x06060606;   // 6  6  6  6
  mul_ct2=0x03030303;   // 3  3  3  3
  mul_ct3=0x01010101;   // 1  1  1  1
      ct0=0x00100010;   //   16    16

  src_aln0 = (((uint32_t)src) & 0xFFFFFFFC);
  src_aln  = src_aln0;
  src_rs0 = 4-(((uint32_t)src) & 3); 
  dst_aln = (uint32_t)dst;
  dst_aln1 = dst_aln;
  S32I2M(xr15,mul_ct0); //xr15:20 20 20 20
  S32I2M(xr14,mul_ct1); //xr14:6  6  6  6
  S32I2M(xr13,mul_ct2); //xr13:3  3  3  3
  S32I2M(xr12,mul_ct3); //xr12:1  1  1  1
  S32I2M(xr11,    ct0); //xr11:  16    16

 for(i=0; i<4; i++)
 { 
  S32LDD(xr1,src_aln0,0