h264.c

手机端的H264源码

    src[0+1*stride]=
    src[1+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
    src[1+1*stride]=
    src[2+3*stride]=(lt + 2*t0 + t1 + 2)>>2;
    src[2+1*stride]=
    src[3+3*stride]=(t0 + 2*t1 + t2 + 2)>>2;
    src[3+1*stride]=(t1 + 2*t2 + t3 + 2)>>2;
    src[0+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
    src[0+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
}

static void pred4x4_vertical_left_c(uint8_t *src, uint8_t *topright, int stride){
    LOAD_TOP_EDGE    
    LOAD_TOP_RIGHT_EDGE    
    //const __attribute__((unused)) int unu= t7;

    src[0+0*stride]=(t0 + t1 + 1)>>1;
    src[1+0*stride]=
    src[0+2*stride]=(t1 + t2 + 1)>>1;
    src[2+0*stride]=
    src[1+2*stride]=(t2 + t3 + 1)>>1;
    src[3+0*stride]=
    src[2+2*stride]=(t3 + t4+ 1)>>1;
    src[3+2*stride]=(t4 + t5+ 1)>>1;
    src[0+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
    src[1+1*stride]=
    src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
    src[2+1*stride]=
    src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
    src[3+1*stride]=
    src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
    src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
}

static void pred4x4_horizontal_up_c(uint8_t *src, uint8_t *topright, int stride){
    LOAD_LEFT_EDGE    

    src[0+0*stride]=(l0 + l1 + 1)>>1;
    src[1+0*stride]=(l0 + 2*l1 + l2 + 2)>>2;
    src[2+0*stride]=
    src[0+1*stride]=(l1 + l2 + 1)>>1;
    src[3+0*stride]=
    src[1+1*stride]=(l1 + 2*l2 + l3 + 2)>>2;
    src[2+1*stride]=
    src[0+2*stride]=(l2 + l3 + 1)>>1;
    src[3+1*stride]=
    src[1+2*stride]=(l2 + 2*l3 + l3 + 2)>>2;
    src[3+2*stride]=
    src[1+3*stride]=
    src[0+3*stride]=
    src[2+2*stride]=
    src[2+3*stride]=
    src[3+3*stride]=l3;
}
    
static void pred4x4_horizontal_down_c(uint8_t *src, uint8_t *topright, int stride){
    const int lt= src[-1-1*stride];
    LOAD_TOP_EDGE    
    LOAD_LEFT_EDGE    
    //const __attribute__((unused)) int unu= t3;

    src[0+0*stride]=
    src[2+1*stride]=(lt + l0 + 1)>>1;
    src[1+0*stride]=
    src[3+1*stride]=(l0 + 2*lt + t0 + 2)>>2;
    src[2+0*stride]=(lt + 2*t0 + t1 + 2)>>2;
    src[3+0*stride]=(t0 + 2*t1 + t2 + 2)>>2;
    src[0+1*stride]=
    src[2+2*stride]=(l0 + l1 + 1)>>1;
    src[1+1*stride]=
    src[3+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
    src[0+2*stride]=
    src[2+3*stride]=(l1 + l2+ 1)>>1;
    src[1+2*stride]=
    src[3+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
    src[0+3*stride]=(l2 + l3 + 1)>>1;
    src[1+3*stride]=(l1 + 2*l2 + l3 + 2)>>2;
}

static void pred16x16_vertical_c(uint8_t *src, int stride){
    int i;
    const uint32_t a= ((uint32_t*)(src-stride))[0];
    const uint32_t b= ((uint32_t*)(src-stride))[1];
    const uint32_t c= ((uint32_t*)(src-stride))[2];
    const uint32_t d= ((uint32_t*)(src-stride))[3];
    
    for(i=0; i<16; i++){
        ((uint32_t*)(src+i*stride))[0]= a;
        ((uint32_t*)(src+i*stride))[1]= b;
        ((uint32_t*)(src+i*stride))[2]= c;
        ((uint32_t*)(src+i*stride))[3]= d;
    }
}

static void pred16x16_horizontal_c(uint8_t *src, int stride){
    int i;

    for(i=0; i<16; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]=
        ((uint32_t*)(src+i*stride))[2]=
        ((uint32_t*)(src+i*stride))[3]= src[-1+i*stride]*0x01010101;
    }
}

static void pred16x16_dc_c(uint8_t *src, int stride){
    int i, dc=0;

    for(i=0;i<16; i++){
        dc+= src[-1+i*stride];
    }
    
    for(i=0;i<16; i++){
        dc+= src[i-stride];
    }

    dc= 0x01010101*((dc + 16)>>5);

    for(i=0; i<16; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]=
        ((uint32_t*)(src+i*stride))[2]=
        ((uint32_t*)(src+i*stride))[3]= dc;
    }
}

static void pred16x16_left_dc_c(uint8_t *src, int stride){
    int i, dc=0;

    for(i=0;i<16; i++){
        dc+= src[-1+i*stride];
    }
    
    dc= 0x01010101*((dc + 8)>>4);

    for(i=0; i<16; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]=
        ((uint32_t*)(src+i*stride))[2]=
        ((uint32_t*)(src+i*stride))[3]= dc;
    }
}

static void pred16x16_top_dc_c(uint8_t *src, int stride){
    int i, dc=0;

    for(i=0;i<16; i++){
        dc+= src[i-stride];
    }
    dc= 0x01010101*((dc + 8)>>4);

    for(i=0; i<16; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]=
        ((uint32_t*)(src+i*stride))[2]=
        ((uint32_t*)(src+i*stride))[3]= dc;
    }
}

static void pred16x16_128_dc_c(uint8_t *src, int stride){
    int i;

    for(i=0; i<16; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]=
        ((uint32_t*)(src+i*stride))[2]=
        ((uint32_t*)(src+i*stride))[3]= 0x01010101U*128U;
    }
}

static  void pred16x16_plane_compat_c(uint8_t *src, int stride, const int svq3){
  int i, j, k;
  int a;
  uint8_t *cm = cropTbl + MAX_NEG_CROP;
  const uint8_t * const src0 = src+7-stride;
  const uint8_t *src1 = src+8*stride-1;
  const uint8_t *src2 = src1-2*stride;      // == src+6*stride-1;
  int H = src0[1] - src0[-1];
  int V = src1[0] - src2[ 0];
  for(k=2; k<=8; ++k) {
    src1 += stride; src2 -= stride;
    H += k*(src0[k] - src0[-k]);
    V += k*(src1[0] - src2[ 0]);
  }
  if(svq3){
    H = ( 5*(H/4) ) / 16;
    V = ( 5*(V/4) ) / 16;

    /* required for 100% accuracy */
    i = H; H = V; V = i;
  }else{
    H = ( 5*H+32 ) >> 6;
    V = ( 5*V+32 ) >> 6;
  }

  a = 16*(src1[0] + src2[16] + 1) - 7*(V+H);
  for(j=16; j>0; --j) {
    int b = a;
    a += V;
    for(i=-16; i<0; i+=4) {
      src[16+i] = cm[ (b    ) >> 5 ];
      src[17+i] = cm[ (b+  H) >> 5 ];
      src[18+i] = cm[ (b+2*H) >> 5 ];
      src[19+i] = cm[ (b+3*H) >> 5 ];
      b += 4*H;
    }
    src += stride;
  }
}

static void pred16x16_plane_c(uint8_t *src, int stride){
    pred16x16_plane_compat_c(src, stride, 0);
}

static void pred8x8_vertical_c(uint8_t *src, int stride){
    int i;
    const uint32_t a= ((uint32_t*)(src-stride))[0];
    const uint32_t b= ((uint32_t*)(src-stride))[1];
    
    for(i=0; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]= a;
        ((uint32_t*)(src+i*stride))[1]= b;
    }
}

static void pred8x8_horizontal_c(uint8_t *src, int stride){
    int i;

    for(i=0; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]= src[-1+i*stride]*0x01010101;
    }
}

static void pred8x8_128_dc_c(uint8_t *src, int stride){
    int i;

    for(i=0; i<4; i++){
        ((uint32_t*)(src+i*stride))[0]= 
        ((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U;
    }
    for(i=4; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]= 
        ((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U;
    }
}

static void pred8x8_left_dc_c(uint8_t *src, int stride){
    int i;
    int dc0, dc2;

    dc0=dc2=0;
    for(i=0;i<4; i++){
        dc0+= src[-1+i*stride];
        dc2+= src[-1+(i+4)*stride];
    }
    dc0= 0x01010101*((dc0 + 2)>>2);
    dc2= 0x01010101*((dc2 + 2)>>2);

    for(i=0; i<4; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]= dc0;
    }
    for(i=4; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]= dc2;
    }
}

static void pred8x8_top_dc_c(uint8_t *src, int stride){
    int i;
    int dc0, dc1;

    dc0=dc1=0;
    for(i=0;i<4; i++){
        dc0+= src[i-stride];
        dc1+= src[4+i-stride];
    }
    dc0= 0x01010101*((dc0 + 2)>>2);
    dc1= 0x01010101*((dc1 + 2)>>2);

    for(i=0; i<4; i++){
        ((uint32_t*)(src+i*stride))[0]= dc0;
        ((uint32_t*)(src+i*stride))[1]= dc1;
    }
    for(i=4; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]= dc0;
        ((uint32_t*)(src+i*stride))[1]= dc1;
    }
}


static void pred8x8_dc_c(uint8_t *src, int stride){
    int i;
    int dc0, dc1, dc2, dc3;

    dc0=dc1=dc2=0;
    for(i=0;i<4; i++){
        dc0+= src[-1+i*stride] + src[i-stride];
        dc1+= src[4+i-stride];
        dc2+= src[-1+(i+4)*stride];
    }
    dc3= 0x01010101*((dc1 + dc2 + 4)>>3);
    dc0= 0x01010101*((dc0 + 4)>>3);
    dc1= 0x01010101*((dc1 + 2)>>2);
    dc2= 0x01010101*((dc2 + 2)>>2);

    for(i=0; i<4; i++){
        ((uint32_t*)(src+i*stride))[0]= dc0;
        ((uint32_t*)(src+i*stride))[1]= dc1;
    }
    for(i=4; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]= dc2;
        ((uint32_t*)(src+i*stride))[1]= dc3;
    }
}

static void pred8x8_plane_c(uint8_t *src, int stride){
  int j, k;
  int a;
  uint8_t *cm = cropTbl + MAX_NEG_CROP;
  const uint8_t * const src0 = src+3-stride;
  const uint8_t *src1 = src+4*stride-1;
  const uint8_t *src2 = src1-2*stride;      // == src+2*stride-1;
  int H = src0[1] - src0[-1];
  int V = src1[0] - src2[ 0];
  for(k=2; k<=4; ++k) {
    src1 += stride; src2 -= stride;
    H += k*(src0[k] - src0[-k]);
    V += k*(src1[0] - src2[ 0]);
  }
  H = ( 17*H+16 ) >> 5;
  V = ( 17*V+16 ) >> 5;

  a = 16*(src1[0] + src2[8]+1) - 3*(V+H);
  for(j=8; j>0; --j) {
    int b = a;
    a += V;
    src[0] = cm[ (b    ) >> 5 ];
    src[1] = cm[ (b+  H) >> 5 ];
    src[2] = cm[ (b+2*H) >> 5 ];
    src[3] = cm[ (b+3*H) >> 5 ];
    src[4] = cm[ (b+4*H) >> 5 ];
    src[5] = cm[ (b+5*H) >> 5 ];
    src[6] = cm[ (b+6*H) >> 5 ];
    src[7] = cm[ (b+7*H) >> 5 ];
    src += stride;
  }
}

static  void mc_dir_part(H264Context *h, Picture *pic, int n, int square, int chroma_height, int delta, int list,
                           uint8_t *dest_y, uint8_t *dest_cb, uint8_t *dest_cr,
                           int src_x_offset, int src_y_offset,
                           qpel_mc_func *qpix_op, h264_chroma_mc_func chroma_op){
    MpegEncContext * const s = &h->s;
    const int mx= h->mv_cache[list][ scan8[n] ][0] + src_x_offset*8;
    const int my= h->mv_cache[list][ scan8[n] ][1] + src_y_offset*8;
    const int luma_xy= (mx&3) + ((my&3)<<2);
    uint8_t * src_y = pic->data[0] + (mx