comparison dsputil.c @ 2045:9447bbd8a7e9 libavcodec

rewrite h261 loop filter no malloc(64) memcpy free stuff no filter1 A->B then filter2 A->B (yes not B->A) no incorrect rouding after the 1d filter

2044:b6f2add2511e

        src[y*stride-2] = p0 - d2;
        src[y*stride+1] = p3 + d2;
    }
}

static void h261_v_loop_filter_c(uint8_t *dest,uint8_t *src, int stride){
    int i,j,xy,yz;
    int res;
    for(i=0; i<8; i++){
        for(j=1; j<7; j++){
            xy = j * stride + i;
            yz = j * 8 + i;
            res = (int)src[yz-1*8] + ((int)(src[yz+0*8]) * 2) + (int)src[yz+1*8];
            res +=2;
            res >>=2;
            dest[xy] = (uint8_t)res;
        }
    }
}

static void h261_h_loop_filter_c(uint8_t *dest,uint8_t *src, int stride){
    int i,j,xy,yz;
    int res;    
    for(i=1; i<7; i++){
        for(j=0; j<8; j++){
            xy = j * stride + i;
            yz = j * 8 + i;
            res = (int)src[yz-1] + ((int)(src[yz]) *2) + (int)src[yz+1];
            res+=2;
            res>>=2;
            dest[xy] = (uint8_t)res;
        }
    }
}

static inline int pix_abs16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)

    c->bswap_buf= bswap_buf;
    
    c->h263_h_loop_filter= h263_h_loop_filter_c;
    c->h263_v_loop_filter= h263_v_loop_filter_c;
    
    c->h261_h_loop_filter= h261_h_loop_filter_c;
    c->h261_v_loop_filter= h261_v_loop_filter_c;
    
    c->try_8x8basis= try_8x8basis_c;
    c->add_8x8basis= add_8x8basis_c;

#ifdef HAVE_MMX

2045:9447bbd8a7e9

        src[y*stride-2] = p0 - d2;
        src[y*stride+1] = p3 + d2;
    }
}

static void h261_loop_filter_c(uint8_t *src, int stride){
    int x,y,xy,yz;
    int temp[64];

    for(x=0; x<8; x++){
        temp[x      ] = 4*src[x           ];
        temp[x + 7*8] = 4*src[x + 7*stride];
    }
    for(y=1; y<7; y++){
        for(x=0; x<8; x++){
            xy = y * stride + x;
            yz = y * 8 + x;
            temp[yz] = src[xy - stride] + 2*src[xy] + src[xy + stride];
        }
    }
        
    for(y=0; y<8; y++){
        src[  y*stride] = (temp[  y*8] + 2)>>2;
        src[7+y*stride] = (temp[7+y*8] + 2)>>2;
        for(x=1; x<7; x++){
            xy = y * stride + x;
            yz = y * 8 + x;
            src[xy] = (temp[yz-1] + 2*temp[yz] + temp[yz+1] + 8)>>4;
        }
    }
}

static inline int pix_abs16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)

    c->bswap_buf= bswap_buf;
    
    c->h263_h_loop_filter= h263_h_loop_filter_c;
    c->h263_v_loop_filter= h263_v_loop_filter_c;
    
    c->h261_loop_filter= h261_loop_filter_c;
    
    c->try_8x8basis= try_8x8basis_c;
    c->add_8x8basis= add_8x8basis_c;

#ifdef HAVE_MMX