macroblock-dz.c

H.264 source codes

    for( ch = 0; ch < 2; ch++ )
    {
        int16_t chroma[4][4][4];
        int16_t dct2x2[2][2];
        int16_t dct4x4[4][4][4];

        p_src = ctx->p_img[1+ch];
        i_src = ctx->i_img[1+ch];
        p_dst = ctx->p_fdec[1+ch];
        i_dst = ctx->i_fdec[1+ch];

        /* calculate the diff */
        h->pixf.sub8x8( chroma, p_src, i_src, p_dst, i_dst );

        /* calculate dct coeffs */
        for( i = 0; i < 4; i++ )
        {
            h->dctf.dct4x4( dct4x4[i], chroma[i] );

            /* copy dc coeff */
            dct2x2[block_idx_y[i]][block_idx_x[i]] = dct4x4[i][0][0];

            quant_4x4( dct4x4[i], i_qscale, b_inter ? 0 : 1 );
            scan_zigzag_4x4( mb->block[16+i+ch*4].residual_ac, dct4x4[i] );

            i_decimate_score += x264_mb_decimate_score( mb->block[16+i+ch*4].residual_ac, 15 );
        }

        h->dctf.dct2x2dc( dct2x2, dct2x2 );
        quant_2x2_dc( dct2x2, i_qscale, b_inter ? 0 : 1  );
        scan_zigzag_2x2_dc( mb->chroma_dc[ch], dct2x2 );

        if( i_decimate_score < 7 && b_inter )
        {
            /* Near null chroma 8x8 block so make it null (bits saving) */
            for( i = 0; i < 4; i++ )
            {
                int x, y;
                for( x = 0; x < 15; x++ )
                {
                    mb->block[16+i+ch*4].residual_ac[x] = 0;
                }
                for( x = 0; x < 4; x++ )
                {
                    for( y = 0; y < 4; y++ )
                    {
                        dct4x4[i][x][y] = 0;
                    }
                }
            }
        }

        /* output samples to fdec */
        h->dctf.idct2x2dc( dct2x2, dct2x2 );
        x264_mb_dequant_2x2_dc( dct2x2, i_qscale );  /* XXX not inversed */

        /* calculate dct coeffs */
        for( i = 0; i < 4; i++ )
        {
            x264_mb_dequant_4x4( dct4x4[i], i_qscale );

            /* copy dc coeff */
            dct4x4[i][0][0] = dct2x2[block_idx_y[i]][block_idx_x[i]];

            h->dctf.idct4x4( chroma[i], dct4x4[i] );
        }
        h->pixf.add8x8( p_dst, i_dst, chroma );
    }
}

/*****************************************************************************
 * x264_macroblock_encode:
 *****************************************************************************/
void x264_macroblock_encode( x264_t *h, x264_macroblock_t *mb )
{
    x264_mb_context_t *ctx = mb->context;
    int i;

    int     i_qscale;

    /* quantification scale */
    i_qscale = mb->i_qp;

    if( mb->i_type == I_16x16 )
    {
        /* do the right prediction */
        h->predict_16x16[mb->i_intra16x16_pred_mode]( ctx->p_fdec[0], ctx->i_fdec[0] );

        /* encode the 16x16 macroblock */
        x264_mb_encode_i16x16( h, mb, i_qscale );

        /* fix the pred mode value */
        mb->i_intra16x16_pred_mode = x264_mb_pred_mode16x16_fix[mb->i_intra16x16_pred_mode];
    }
    else if( mb->i_type == I_4x4 )
    {
        for( i = 0; i < 16; i++ )
        {
            uint8_t *p_dst_by;

            /* Do the right prediction */
            p_dst_by = ctx->p_fdec[0] + 4 * block_idx_x[i] + 4 * block_idx_y[i] * ctx->i_fdec[0];
            h->predict_4x4[mb->block[i].i_intra4x4_pred_mode]( p_dst_by, ctx->i_fdec[0] );

            /* encode one 4x4 block */
            x264_mb_encode_i4x4( h, mb, i, i_qscale );

            /* fix the pred mode value */
            mb->block[i].i_intra4x4_pred_mode = x264_mb_pred_mode4x4_fix[mb->block[i].i_intra4x4_pred_mode];
        }
    }
    else    /* Inter MB */
    {
        int16_t dct4x4[16][4][4];

        int i8x8, i4x4, idx;
        int i_decimate_mb = 0;

        /* Motion compensation */
        x264_mb_mc( h, mb );

        for( i8x8 = 0; i8x8 < 4; i8x8++ )
        {
            int16_t luma[4][4];
            int i_decimate_8x8;

            /* encode one 4x4 block */
            i_decimate_8x8 = 0;
            for( i4x4 = 0; i4x4 < 4; i4x4++ )
            {
                uint8_t *p_src, *p_dst;

                idx = i8x8 * 4 + i4x4;

                p_src = ctx->p_img[0] + 4 * block_idx_x[idx] + 4 * block_idx_y[idx] * ctx->i_img[0];
                p_dst = ctx->p_fdec[0] + 4 * block_idx_x[idx] + 4 * block_idx_y[idx] * ctx->i_fdec[0];

                /* we calculate diff */
                h->pixf.sub4x4( luma, p_src, ctx->i_img[0],p_dst, ctx->i_fdec[0] );

                /* calculate dct coeffs */
                h->dctf.dct4x4( dct4x4[idx], luma );
                quant_4x4( dct4x4[idx], i_qscale, 0 );

                scan_zigzag_4x4full( mb->block[idx].luma4x4, dct4x4[idx] );
                i_decimate_8x8 += x264_mb_decimate_score( mb->block[idx].luma4x4, 16 );
            }

            /* decimate this 8x8 block */
            i_decimate_mb += i_decimate_8x8;
            if( i_decimate_8x8 < 4 )
            {
                for( i4x4 = 0; i4x4 < 4; i4x4++ )
                {
                    int x, y;
                    idx = i8x8 * 4 + i4x4;
                    for( i = 0; i < 16; i++ )
                    {
                        mb->block[idx].luma4x4[i] = 0;
                    }
                    for( x = 0; x < 4; x++ )
                    {
                        for( y = 0; y < 4; y++ )
                        {
                            dct4x4[idx][x][y] = 0;
                        }
                    }
                }
            }
        }

        if( i_decimate_mb < 6 )
        {
            for( i8x8 = 0; i8x8 < 4; i8x8++ )
            {
                for( i4x4 = 0; i4x4 < 4; i4x4++ )
                {
                    for( i = 0; i < 16; i++ )
                    {
                        mb->block[i8x8 * 4 + i4x4].luma4x4[i] = 0;
                    }
                }
            }
        }
        else
        {
            for( i8x8 = 0; i8x8 < 4; i8x8++ )
            {
                int16_t luma[4][4];
                /* TODO we could avoid it if we had decimate this 8x8 block */
                /* output samples to fdec */
                for( i4x4 = 0; i4x4 < 4; i4x4++ )
                {
                    uint8_t *p_dst;

                    idx = i8x8 * 4 + i4x4;

                    x264_mb_dequant_4x4( dct4x4[idx], i_qscale );
                    h->dctf.idct4x4( luma, dct4x4[idx] );

                    /* put pixel to fdec */
                    p_dst = ctx->p_fdec[0] + 4 * block_idx_x[idx] + 4 * block_idx_y[idx] * ctx->i_fdec[0];
                    h->pixf.add4x4( p_dst, ctx->i_fdec[0], luma );
                }
            }
        }
    }

    /* encode chroma */
    i_qscale = i_chroma_qp_table[x264_clip3( i_qscale + h->pps->i_chroma_qp_index_offset, 0, 51 )];
    if( IS_INTRA( mb->i_type ) )
    {
        /* do the right prediction */
        h->predict_8x8[mb->i_chroma_pred_mode]( ctx->p_fdec[1], ctx->i_fdec[1] );
        h->predict_8x8[mb->i_chroma_pred_mode]( ctx->p_fdec[2], ctx->i_fdec[2] );
    }

    /* encode the 8x8 blocks */
    x264_mb_encode_8x8( h, mb, !IS_INTRA( mb->i_type ), i_qscale );

    /* fix the pred mode value */
    if( IS_INTRA( mb->i_type ) )
    {
        mb->i_chroma_pred_mode = x264_mb_pred_mode8x8_fix[mb->i_chroma_pred_mode];
    }

    /* Calculate the Luma/Chroma patern and non_zero_count */
    if( mb->i_type == I_16x16 )
    {
        mb->i_cbp_luma = 0x00;
        for( i = 0; i < 16; i++ )
        {
            mb->block[i].i_non_zero_count = array_non_zero_count( mb->block[i].residual_ac, 15 );
            if( mb->block[i].i_non_zero_count > 0 )
            {
                mb->i_cbp_luma = 0x0f;
            }
        }
    }
    else
    {
        mb->i_cbp_luma = 0x00;
        for( i = 0; i < 16; i++ )
        {
            mb->block[i].i_non_zero_count = array_non_zero_count( mb->block[i].luma4x4, 16 );
            if( mb->block[i].i_non_zero_count > 0 )
            {
                mb->i_cbp_luma |= 1 << (i/4);
            }
        }
    }

    /* Calculate the chroma patern */
    mb->i_cbp_chroma = 0x00;
    for( i = 0; i < 8; i++ )
    {
        mb->block[16+i].i_non_zero_count = array_non_zero_count( mb->block[16+i].residual_ac, 15 );
        if( mb->block[16+i].i_non_zero_count > 0 )
        {
            mb->i_cbp_chroma = 0x02;    /* dc+ac (we can't do only ac) */
        }
    }
    if( mb->i_cbp_chroma == 0x00 &&
        ( array_non_zero_count( mb->chroma_dc[0], 4 ) > 0 || array_non_zero_count( mb->chroma_dc[1], 4 ) ) > 0 )
    {
        mb->i_cbp_chroma = 0x01;    /* dc only */
    }

    /* Check for P_SKIP
     * XXX: in the me perhaps we should take x264_mb_predict_mv_pskip into account
     *      (if multiple mv give same result)*/
    if( mb->i_type == P_L0 && mb->i_partition == D_16x16 &&
        mb->i_cbp_luma == 0x00 && mb->i_cbp_chroma == 0x00 )
    {

        int i_ref;
        int mvx, mvy;
        x264_mb_partition_get( mb, 0, 0, 0, &i_ref, &mvx, &mvy );

        if( i_ref == 0 )
        {
            int mvp[2];

            x264_mb_predict_mv_pskip( mb, mvp );
            if( mvp[0] == mvx && mvp[1] == mvy )
            {
                mb->i_type = P_SKIP;
            }
        }
    }
}


#define BLOCK_INDEX_CHROMA_DC   (-1)
#define BLOCK_INDEX_LUMA_DC     (-2)

static inline void bs_write_vlc( bs_t *s, vlc_t v )
{
    bs_write( s, v.i_size, v.i_bits );
}

/****************************************************************************
 * block_residual_write_cavlc:
 ****************************************************************************/
static void block_residual_write_cavlc( x264_t *h, bs_t *s, x264_macroblock_t *mb, int i_idx, int *l, int i_count )
{
    int level[16], run[16];
    int i_total, i_trailing;
    int i_total_zero;
    int i_last;
    unsigned int i_sign;

    int i;
    int i_zero_left;
    int i_suffix_length;

    /* first find i_last */
    i_last = i_count - 1;
    while( i_last >= 0 && l[i_last] == 0 )
    {
        i_last--;
    }

    i_sign = 0;
    i_total = 0;
    i_trailing = 0;
    i_total_zero = 0;

    if( i_last >= 0 )
    {
        int b_trailing = 1;
        int idx = 0;

        /* level and run and total */
        while( i_last >= 0 )
        {
            level[idx] = l[i_last--];

            run[idx] = 0;
            while( i_last >= 0 && l[i_last] == 0 )
            {
                run[idx]++;
                i_last--;
            }

            i_total++;
            i_total_zero += run[idx];

            if( b_trailing && abs( level[idx] ) == 1 && i_trailing < 3 )
            {
                i_sign <<= 1;
                if( level[idx] < 0 )
                {
                    i_sign |= 0x01;
                }

                i_trailing++;
            }
            else
            {
                b_trailing = 0;
            }

            idx++;
        }
    }

    /* total/trailing */
    if( i_idx == BLOCK_INDEX_CHROMA_DC )
    {
        bs_write_vlc( s, x264_coeff_token[4][i_total*4+i_trailing] );
    }
    else
    {
        /* x264_mb_predict_non_zero_code return 0 <-> (16+16+1)>>1 = 16 */
        static const int ct_index[17] = {0,0,1,1,2,2,2,2,3,3,3,3,3,3,3,3,3 };
        int nC;

        if( i_idx == BLOCK_INDEX_LUMA_DC )
        {
            nC = x264_mb_predict_non_zero_code( h, mb, 0 );
        }
        else
        {
            nC = x264_mb_predict_non_zero_code( h, mb, i_idx );
        }

        bs_write_vlc( s, x264_coeff_token[ct_index[nC]][i_total*4+i_trailing] );
    }

    if( i_total <= 0 )
    {
        return;
    }

    i_suffix_length = i_total > 10 && i_trailing < 3 ? 1 : 0;
    if( i_trailing > 0 )
    {
        bs_write( s, i_trailing, i_sign );
    }
    for( i = i_trailing; i < i_total; i++ )
    {
        int i_level_code;

        /* calculate level code */
        if( level[i] < 0 )
        {
            i_level_code = -2*level[i] - 1;
        }
        else /* if( level[i] > 0 ) */
        {
            i_level_code = 2 * level[i] - 2;
        }
        if( i == i_trailing && i_trailing < 3 )
        {
            i_level_code -=2; /* as level[i] can't be 1 for the first one if i_trailing < 3 */
        }

        if( ( i_level_code >> i_suffix_length ) < 14 )
        {
            bs_write_vlc( s, x264_level_prefix[i_level_code >> i_suffix_length] );
            if( i_suffix_length > 0 )
            {
                bs_write( s, i_suffix_length, i_level_code );
            }
        }
        else if( i_suffix_length == 0 && i_level_code < 30 )
        {
            bs_write_vlc( s, x264_level_prefix[14] );
            bs_write( s, 4, i_level_code - 14 );
        }
        else if( i_suffix_length > 0 && ( i_level_code >> i_suffix_length ) == 14 )
        {
            bs_write_vlc( s, x264_level_prefix[14] );
            bs_write( s, i_suffix_length, i_level_code );
        }
        else
        {
            bs_write_vlc( s, x264_level_prefix[15] );
            i_level_code -= 15 << i_suffix_length;
            if( i_suffix_length == 0 )
            {
                i_level_code -= 15;
            }

            if( i_level_code >= ( 1 << 12 ) || i_level_code < 0 )
            {
                fprintf( stderr, "OVERFLOW levelcode=%d\n", i_level_code );
            }

            bs_write( s, 12, i_level_code );    /* check overflow ?? */