edge-detec.c.svn-base

H.264 source codes

/*****************************************************************************
 * macroblock.c: h264 encoder library
 *****************************************************************************
 * Copyright (C) 2003 Laurent Aimar
 * $Id: edge-detec.c,v 1.1 2004/06/03 19:27:08 fenrir Exp $
 *
 * Authors: Laurent Aimar <[e-mail]>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111, USA.
 *****************************************************************************/

#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <stdint.h>
#include <math.h>

#include "common.h"
#include "me.h"
#include "vlc.h"

static inline int x264_median( int a, int b, int c )
{
    int min = a, max =a;
    if( b < min )
    {
        min = b;
    }
    else
    {
        max = b;    /* no need to do 'b > max' (more consuming than always doing affectation) */
    }
    if( c < min )
    {
        min = c;
    }
    else if( c > max )
    {
        max = c;
    }

    return a + b + c - min - max;
}

static const uint8_t intra4x4_cbp_to_golomb[48]=
{
  3, 29, 30, 17, 31, 18, 37,  8, 32, 38, 19,  9, 20, 10, 11,  2,
 16, 33, 34, 21, 35, 22, 39,  4, 36, 40, 23,  5, 24,  6,  7,  1,
 41, 42, 43, 25, 44, 26, 46, 12, 45, 47, 27, 13, 28, 14, 15,  0
};
static const uint8_t inter_cbp_to_golomb[48]=
{
  0,  2,  3,  7,  4,  8, 17, 13,  5, 18,  9, 14, 10, 15, 16, 11,
  1, 32, 33, 36, 34, 37, 44, 40, 35, 45, 38, 41, 39, 42, 43, 19,
  6, 24, 25, 20, 26, 21, 46, 28, 27, 47, 22, 29, 23, 30, 31, 12
};

static const uint8_t block_idx_x[16] =
{
    0, 1, 0, 1, 2, 3, 2, 3, 0, 1, 0, 1, 2, 3, 2, 3
};
static const uint8_t block_idx_y[16] =
{
    0, 0, 1, 1, 0, 0, 1, 1, 2, 2, 3, 3, 2, 2, 3, 3
};
static const uint8_t block_idx_xy[4][4] =
{
    { 0, 2, 8,  10},
    { 1, 3, 9,  11},
    { 4, 6, 12, 14},
    { 5, 7, 13, 15}
};

static const int quant_mf[6][4][4] =
{
    {  { 13107, 8066, 13107, 8066}, {  8066, 5243,  8066, 5243},
       { 13107, 8066, 13107, 8066}, {  8066, 5243,  8066, 5243}  },
    {  { 11916, 7490, 11916, 7490}, {  7490, 4660,  7490, 4660},
       { 11916, 7490, 11916, 7490}, {  7490, 4660,  7490, 4660}  },
    {  { 10082, 6554, 10082, 6554}, {  6554, 4194,  6554, 4194},
       { 10082, 6554, 10082, 6554}, {  6554, 4194,  6554, 4194}  },
    {  {  9362, 5825,  9362, 5825}, {  5825, 3647,  5825, 3647},
       {  9362, 5825,  9362, 5825}, {  5825, 3647,  5825, 3647}  },
    {  {  8192, 5243,  8192, 5243}, {  5243, 3355,  5243, 3355},
       {  8192, 5243,  8192, 5243}, {  5243, 3355,  5243, 3355}  },
    {  {  7282, 4559,  7282, 4559}, {  4559, 2893,  4559, 2893},
       {  7282, 4559,  7282, 4559}, {  4559, 2893,  4559, 2893}  }
};

static const int dequant_mf[6][4][4] =
{
    { {10, 13, 10, 13}, {13, 16, 13, 16}, {10, 13, 10, 13}, {13, 16, 13, 16} },
    { {11, 14, 11, 14}, {14, 18, 14, 18}, {11, 14, 11, 14}, {14, 18, 14, 18} },
    { {13, 16, 13, 16}, {16, 20, 16, 20}, {13, 16, 13, 16}, {16, 20, 16, 20} },
    { {14, 18, 14, 18}, {18, 23, 18, 23}, {14, 18, 14, 18}, {18, 23, 18, 23} },
    { {16, 20, 16, 20}, {20, 25, 20, 25}, {16, 20, 16, 20}, {20, 25, 20, 25} },
    { {18, 23, 18, 23}, {23, 29, 23, 29}, {18, 23, 18, 23}, {23, 29, 23, 29} }
};


static int predict_pred_intra4x4_mode( x264_t *h, x264_macroblock_t *mb, int idx )
{
    x264_macroblock_t *mba = mb->context->block[idx].mba;
    x264_macroblock_t *mbb = mb->context->block[idx].mbb;

    int i_mode_a = I_PRED_4x4_DC;
    int i_mode_b = I_PRED_4x4_DC;

    if( !mba || !mbb )
    {
        return I_PRED_4x4_DC;
    }

    if( mba->i_type == I_4x4 )
    {
        i_mode_a = mb->context->block[idx].bka->i_intra4x4_pred_mode;
    }
    if( mbb->i_type == I_4x4 )
    {
        i_mode_b = mb->context->block[idx].bkb->i_intra4x4_pred_mode;
    }

    return X264_MIN( i_mode_a, i_mode_b );
}

static int predict_non_zero_code( x264_t *h, x264_macroblock_t *mb, int idx )
{
    x264_macroblock_t *mba = mb->context->block[idx].mba;
    x264_macroblock_t *mbb = mb->context->block[idx].mbb;

    int i_z_a = 0x80, i_z_b = 0x80;
    int i_ret;

    /* none avail -> 0, one avail -> this one, both -> (a+b+1)>>1 */
    if( mba )
    {
        i_z_a = mb->context->block[idx].bka->i_non_zero_count;
    }
    if( mbb )
    {
        i_z_b = mb->context->block[idx].bkb->i_non_zero_count;
    }

    i_ret = i_z_a+i_z_b;
    if( i_ret < 0x80 )
    {
        i_ret = ( i_ret + 1 ) >> 1;
    }
    return i_ret & 0x7f;
}


/*
 * Handle intra mb
 */
/* Max = 4 */
static void predict_16x16_mode_available( x264_macroblock_t *mb, int *mode, int *pi_count )
{
    if( ( mb->i_neighbour & (MB_LEFT|MB_TOP) ) == (MB_LEFT|MB_TOP) )
    {
        /* top and left avaible */
        *mode++ = I_PRED_16x16_DC;
        *mode++ = I_PRED_16x16_V;
        *mode++ = I_PRED_16x16_H;
        *mode++ = I_PRED_16x16_P;
        *pi_count = 4;
    }
    else if( ( mb->i_neighbour & MB_LEFT ) )
    {
        /* left available*/
        *mode++ = I_PRED_16x16_DC_LEFT;
        *mode++ = I_PRED_16x16_H;
        *pi_count = 2;
    }
    else if( ( mb->i_neighbour & MB_TOP ) )
    {
        /* top available*/
        *mode++ = I_PRED_16x16_DC_TOP;
        *mode++ = I_PRED_16x16_V;
        *pi_count = 2;
    }
    else
    {
        /* none avaible */
        *mode = I_PRED_16x16_DC_128;
        *pi_count = 1;
    }
}

/* Max = 4 */
static void predict_8x8_mode_available( x264_macroblock_t *mb, int *mode, int *pi_count )
{
    if( ( mb->i_neighbour & (MB_LEFT|MB_TOP) ) == (MB_LEFT|MB_TOP) )
    {
        /* top and left avaible */
        *mode++ = I_PRED_CHROMA_DC;
        *mode++ = I_PRED_CHROMA_V;
        *mode++ = I_PRED_CHROMA_H;
        *mode++ = I_PRED_CHROMA_P;
        *pi_count = 4;
    }
    else if( ( mb->i_neighbour & MB_LEFT ) )
    {
        /* left available*/
        *mode++ = I_PRED_CHROMA_DC_LEFT;
        *mode++ = I_PRED_CHROMA_H;
        *pi_count = 2;
    }
    else if( ( mb->i_neighbour & MB_TOP ) )
    {
        /* top available*/
        *mode++ = I_PRED_CHROMA_DC_TOP;
        *mode++ = I_PRED_CHROMA_V;
        *pi_count = 2;
    }
    else
    {
        /* none avaible */
        *mode = I_PRED_CHROMA_DC_128;
        *pi_count = 1;
    }
}

/* MAX = 8 */
static void predict_4x4_mode_available( x264_macroblock_t *mb, int idx, int *mode, int *pi_count )
{
    int b_a, b_b, b_c;
    static const int needmb[16] =
    {
        MB_LEFT|MB_TOP, MB_TOP,
        MB_LEFT,        MB_PRIVATE,
        MB_TOP,         MB_TOP|MB_TOPRIGHT,
        0,              MB_PRIVATE,
        MB_LEFT,        0,
        MB_LEFT,        MB_PRIVATE,
        0,              MB_PRIVATE,
        0,              MB_PRIVATE
    };

    /* FIXME even when b_c == 0 there is some case where missing pixels
     * are emulated and thus more mode are available TODO
     * analysis and encode should be fixed too */
    b_a = (needmb[idx]&mb->i_neighbour&MB_LEFT) == (needmb[idx]&MB_LEFT);
    b_b = (needmb[idx]&mb->i_neighbour&MB_TOP) == (needmb[idx]&MB_TOP);
    b_c = (needmb[idx]&mb->i_neighbour&(MB_TOPRIGHT|MB_PRIVATE)) == (needmb[idx]&(MB_TOPRIGHT|MB_PRIVATE));

    if( b_a && b_b )
    {
        *mode++ = I_PRED_4x4_DC;
        *mode++ = I_PRED_4x4_H;
        *mode++ = I_PRED_4x4_V;
        *mode++ = I_PRED_4x4_DDR;
        *mode++ = I_PRED_4x4_VR;
        *mode++ = I_PRED_4x4_HD;
        *mode++ = I_PRED_4x4_HU;

        *pi_count = 7;

        if( b_c )
        {
            *mode++ = I_PRED_4x4_DDL;
            *mode++ = I_PRED_4x4_VL;
            (*pi_count) += 2;
        }
    }
    else if( b_a && !b_b )
    {
        *mode++ = I_PRED_4x4_DC_LEFT;
        *mode++ = I_PRED_4x4_H;
        *pi_count = 2;
    }
    else if( !b_a && b_b )
    {
        *mode++ = I_PRED_4x4_DC_TOP;
        *mode++ = I_PRED_4x4_V;
        *pi_count = 2;
    }
    else
    {
        *mode++ = I_PRED_4x4_DC_128;
        *pi_count = 1;
    }
}

/****************************************************************************
 * Scan and Quant functions
 ****************************************************************************/
static const int scan_zigzag_x[16]={0, 1, 0, 0, 1, 2, 3, 2, 1, 0, 1, 2, 3, 3, 2, 3};
static const int scan_zigzag_y[16]={0, 0, 1, 2, 1, 0, 0, 1, 2, 3, 3, 2, 1, 2, 3, 3};

static inline void scan_zigzag_4x4full( int level[16], int16_t dct[4][4] )
{
    int i;

    for( i = 0; i < 16; i++ )
    {
        level[i] = dct[scan_zigzag_y[i]][scan_zigzag_x[i]];
    }
}
static inline void scan_zigzag_4x4( int level[15], int16_t dct[4][4] )
{
    int i;

    for( i = 1; i < 16; i++ )
    {
        level[i - 1] = dct[scan_zigzag_y[i]][scan_zigzag_x[i]];
    }
}

static inline void scan_zigzag_2x2_dc( int level[4], int16_t dct[2][2] )
{
    level[0] = dct[0][0];
    level[1] = dct[0][1];
    level[2] = dct[1][0];
    level[3] = dct[1][1];
}


static void quant_4x4( int16_t dct[4][4], int i_qscale, int b_intra )
{
    int i_qbits = 15 + i_qscale / 6;
    int i_mf = i_qscale % 6;
    int f = ( 1 << i_qbits ) / ( b_intra ? 3 : 6 );

    int x,y;
    for( y = 0; y < 4; y++ )
    {
        for( x = 0; x < 4; x++ )
        {
            if( dct[y][x] > 0 )
            {
                dct[y][x] =( f + (int64_t)dct[y][x]  * (int64_t)quant_mf[i_mf][y][x] ) >> i_qbits;
            }
            else
            {
                dct[y][x] = - ( ( f - (int64_t)dct[y][x]  * (int64_t)quant_mf[i_mf][y][x] ) >> i_qbits );
            }
        }
    }
}
static void quant_4x4_dc( int16_t dct[4][4], int i_qscale, int b_intra )
{
    int i_qbits = 15 + i_qscale / 6;
    int i_mf = i_qscale % 6;
    int f = ( 1 << i_qbits ) / ( b_intra ? 3 : 6 );

    int x,y;
    for( y = 0; y < 4; y++ )
    {
        for( x = 0; x < 4; x++ )
        {
            if( dct[y][x] > 0 )
            {
                dct[y][x] =( 2*f + (int64_t)dct[y][x]  * (int64_t)quant_mf[i_mf][0][0] ) >> ( 1 + i_qbits );
            }
            else
            {
                dct[y][x] = - ( ( 2*f - (int64_t)dct[y][x]  * (int64_t)quant_mf[i_mf][0][0] ) >> (1 + i_qbits ) );
            }
        }
    }
}
static void quant_2x2_dc( int16_t dct[2][2], int i_qscale, int b_intra )
{
    int i_qbits = 15 + i_qscale / 6;
    int i_mf = i_qscale % 6;
    int f = ( 1 << i_qbits ) / ( b_intra ? 3 : 6 );

    int x,y;
    for( y = 0; y < 2; y++ )
    {
        for( x = 0; x < 2; x++ )
        {
            /* XXX: is int64_t really needed ? */
            if( dct[y][x] > 0 )
            {
                dct[y][x] =( 2*f + (int64_t)dct[y][x]  * (int64_t)quant_mf[i_mf][0][0] ) >> ( 1 + i_qbits );
            }
            else
            {
                dct[y][x] = - ( ( 2*f - (int64_t)dct[y][x]  * (int64_t)quant_mf[i_mf][0][0] ) >> (1 + i_qbits ) );
            }
        }
    }
}

static void dequant_4x4_dc( int16_t dct[4][4], int i_qscale )
{
    int i_mf = i_qscale%6;
    int i_qbits = i_qscale/6;
    int f;
    int x,y;

    if( i_qbits <= 1 )
    {
        f = 1 << ( 1 - i_qbits );
    }
    else
    {
        f = 0;
    }

    for( y = 0; y < 4; y++ )
    {
        for( x = 0; x < 4; x++ )
        {
            if( i_qbits >= 2 )
            {
                dct[y][x] = ( dct[y][x] * dequant_mf[i_mf][0][0] ) << (i_qbits - 2);
            }
            else
            {
                dct[y][x] = ( dct[y][x] * dequant_mf[i_mf][0][0] + f ) >> ( 2 -i_qbits );
            }
        }
    }
}

static void dequant_2x2_dc( int16_t dct[2][2], int i_qscale )
{
    int i_mf = i_qscale%6;
    int i_qbits = i_qscale/6;
    int x,y;

    for( y = 0; y < 2; y++ )
    {
        for( x = 0; x < 2; x++ )
        {