流媒体/Mpeg4/MP4

开发平台：
Visual C++

quant.c：源码内容
							/*****************************************************************************
 * quant.c: h264 encoder library
 *****************************************************************************
 * Copyright (C) 2005-2008 x264 project
 *
 * Authors: Loren Merritt <lorenm@u.washington.edu>
 *          Christian Heine <sennindemokrit@gmx.net>
 *
 * 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., 51 Franklin Street, Fifth Floor, Boston, MA  02111, USA.
 *****************************************************************************/
#include "common.h"
#ifdef HAVE_MMX
#include "x86/quant.h"
#endif
#ifdef ARCH_PPC
#   include "ppc/quant.h"
#endif
#ifdef ARCH_ARM
#   include "arm/quant.h"
#endif
#define QUANT_ONE( coef, mf, f ) 
{ 
    if( (coef) > 0 ) 
        (coef) = (f + (coef)) * (mf) >> 16; 
    else 
        (coef) = - ((f - (coef)) * (mf) >> 16); 
    nz |= (coef); 
}
static int quant_8x8( int16_t dct[8][8], uint16_t mf[64], uint16_t bias[64] )
{
    int i, nz = 0;
    for( i = 0; i < 64; i++ )
        QUANT_ONE( dct[0][i], mf[i], bias[i] );
    return !!nz;
}
static int quant_4x4( int16_t dct[4][4], uint16_t mf[16], uint16_t bias[16] )
{
    int i, nz = 0;
    for( i = 0; i < 16; i++ )
        QUANT_ONE( dct[0][i], mf[i], bias[i] );
    return !!nz;
}
static int quant_4x4_dc( int16_t dct[4][4], int mf, int bias )
{
    int i, nz = 0;
    for( i = 0; i < 16; i++ )
        QUANT_ONE( dct[0][i], mf, bias );
    return !!nz;
}
static int quant_2x2_dc( int16_t dct[2][2], int mf, int bias )
{
    int nz = 0;
    QUANT_ONE( dct[0][0], mf, bias );
    QUANT_ONE( dct[0][1], mf, bias );
    QUANT_ONE( dct[0][2], mf, bias );
    QUANT_ONE( dct[0][3], mf, bias );
    return !!nz;
}
#define DEQUANT_SHL( x ) 
    dct[y][x] = ( dct[y][x] * dequant_mf[i_mf][y][x] ) << i_qbits
#define DEQUANT_SHR( x ) 
    dct[y][x] = ( dct[y][x] * dequant_mf[i_mf][y][x] + f ) >> (-i_qbits)
static void dequant_4x4( int16_t dct[4][4], int dequant_mf[6][4][4], int i_qp )
{
    const int i_mf = i_qp%6;
    const int i_qbits = i_qp/6 - 4;
    int y;
    if( i_qbits >= 0 )
    {
        for( y = 0; y < 4; y++ )
        {
            DEQUANT_SHL( 0 );
            DEQUANT_SHL( 1 );
            DEQUANT_SHL( 2 );
            DEQUANT_SHL( 3 );
        }
    }
    else
    {
        const int f = 1 << (-i_qbits-1);
        for( y = 0; y < 4; y++ )
        {
            DEQUANT_SHR( 0 );
            DEQUANT_SHR( 1 );
            DEQUANT_SHR( 2 );
            DEQUANT_SHR( 3 );
        }
    }
}
static void dequant_8x8( int16_t dct[8][8], int dequant_mf[6][8][8], int i_qp )
{
    const int i_mf = i_qp%6;
    const int i_qbits = i_qp/6 - 6;
    int y;
    if( i_qbits >= 0 )
    {
        for( y = 0; y < 8; y++ )
        {
            DEQUANT_SHL( 0 );
            DEQUANT_SHL( 1 );
            DEQUANT_SHL( 2 );
            DEQUANT_SHL( 3 );
            DEQUANT_SHL( 4 );
            DEQUANT_SHL( 5 );
            DEQUANT_SHL( 6 );
            DEQUANT_SHL( 7 );
        }
    }
    else
    {
        const int f = 1 << (-i_qbits-1);
        for( y = 0; y < 8; y++ )
        {
            DEQUANT_SHR( 0 );
            DEQUANT_SHR( 1 );
            DEQUANT_SHR( 2 );
            DEQUANT_SHR( 3 );
            DEQUANT_SHR( 4 );
            DEQUANT_SHR( 5 );
            DEQUANT_SHR( 6 );
            DEQUANT_SHR( 7 );
        }
    }
}
static void dequant_4x4_dc( int16_t dct[4][4], int dequant_mf[6][4][4], int i_qp )
{
    const int i_qbits = i_qp/6 - 6;
    int y;
    if( i_qbits >= 0 )
    {
        const int i_dmf = dequant_mf[i_qp%6][0][0] << i_qbits;
        for( y = 0; y < 4; y++ )
        {
            dct[y][0] *= i_dmf;
            dct[y][1] *= i_dmf;
            dct[y][2] *= i_dmf;
            dct[y][3] *= i_dmf;
        }
    }
    else
    {
        const int i_dmf = dequant_mf[i_qp%6][0][0];
        const int f = 1 << (-i_qbits-1);
        for( y = 0; y < 4; y++ )
        {
            dct[y][0] = ( dct[y][0] * i_dmf + f ) >> (-i_qbits);
            dct[y][1] = ( dct[y][1] * i_dmf + f ) >> (-i_qbits);
            dct[y][2] = ( dct[y][2] * i_dmf + f ) >> (-i_qbits);
            dct[y][3] = ( dct[y][3] * i_dmf + f ) >> (-i_qbits);
        }
    }
}
static void x264_denoise_dct( int16_t *dct, uint32_t *sum, uint16_t *offset, int size )
{
    int i;
    for( i=1; i<size; i++ )
    {
        int level = dct[i];
        int sign = level>>15;
        level = (level+sign)^sign;
        sum[i] += level;
        level -= offset[i];
        dct[i] = level<0 ? 0 : (level^sign)-sign;
    }
}
/* (ref: JVT-B118)
 * x264_mb_decimate_score: given dct coeffs it returns a score to see if we could empty this dct coeffs
 * to 0 (low score means set it to null)
 * Used in inter macroblock (luma and chroma)
 *  luma: for a 8x8 block: if score < 4 -> null
 *        for the complete mb: if score < 6 -> null
 *  chroma: for the complete mb: if score < 7 -> null
 */
const uint8_t x264_decimate_table4[16] = {
    3,2,2,1,1,1,0,0,0,0,0,0,0,0,0,0 };
const uint8_t x264_decimate_table8[64] = {
    3,3,3,3,2,2,2,2,2,2,2,2,1,1,1,1,
    1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,
    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 };
static int ALWAYS_INLINE x264_decimate_score_internal( int16_t *dct, int i_max )
{
    const uint8_t *ds_table = (i_max == 64) ? x264_decimate_table8 : x264_decimate_table4;
    int i_score = 0;
    int idx = i_max - 1;
    /* Yes, dct[idx-1] is guaranteed to be 32-bit aligned.  idx>=0 instead of 1 works correctly for the same reason */
    while( idx >= 0 && *(uint32_t*)&dct[idx-1] == 0 )
        idx -= 2;
    if( idx >= 0 && dct[idx] == 0 )
        idx--;
    while( idx >= 0 )
    {
        int i_run;
        if( (unsigned)(dct[idx--] + 1) > 2 )
            return 9;
        i_run = 0;
        while( idx >= 0 && dct[idx] == 0 )
        {
            idx--;
            i_run++;
        }
        i_score += ds_table[i_run];
    }
    return i_score;
}
static int x264_decimate_score15( int16_t *dct )
{
    return x264_decimate_score_internal( dct+1, 15 );
}
static int x264_decimate_score16( int16_t *dct )
{
    return x264_decimate_score_internal( dct, 16 );
}
static int x264_decimate_score64( int16_t *dct )
{
    return x264_decimate_score_internal( dct, 64 );
}
static int ALWAYS_INLINE x264_coeff_last_internal( int16_t *l, int i_count )
{
    int i_last;
    for( i_last = i_count-1; i_last >= 3; i_last -= 4 )
        if( *(uint64_t*)(l+i_last-3) )
            break;
    while( i_last >= 0 && l[i_last] == 0 )
        i_last--;
    return i_last;
}
static int x264_coeff_last4( int16_t *l )
{
    return x264_coeff_last_internal( l, 4 );
}
static int x264_coeff_last15( int16_t *l )
{
    return x264_coeff_last_internal( l, 15 );
}
static int x264_coeff_last16( int16_t *l )
{
    return x264_coeff_last_internal( l, 16 );
}
static int x264_coeff_last64( int16_t *l )
{
    return x264_coeff_last_internal( l, 64 );
}
#define level_run(num)
static int x264_coeff_level_run##num( int16_t *dct, x264_run_level_t *runlevel )
{
    int i_last = runlevel->last = x264_coeff_last##num(dct);
    int i_total = 0;
    do
    {
        int r = 0;
        runlevel->level[i_total] = dct[i_last];
        while( --i_last >= 0 && dct[i_last] == 0 )
            r++;
        runlevel->run[i_total++] = r;
    } while( i_last >= 0 );
    return i_total;
}
level_run(4)
level_run(15)
level_run(16)
void x264_quant_init( x264_t *h, int cpu, x264_quant_function_t *pf )
{
    pf->quant_8x8 = quant_8x8;
    pf->quant_4x4 = quant_4x4;
    pf->quant_4x4_dc = quant_4x4_dc;
    pf->quant_2x2_dc = quant_2x2_dc;
    pf->dequant_4x4 = dequant_4x4;
    pf->dequant_4x4_dc = dequant_4x4_dc;
    pf->dequant_8x8 = dequant_8x8;
    pf->denoise_dct = x264_denoise_dct;
    pf->decimate_score15 = x264_decimate_score15;
    pf->decimate_score16 = x264_decimate_score16;
    pf->decimate_score64 = x264_decimate_score64;
    pf->coeff_last[DCT_CHROMA_DC] = x264_coeff_last4;
    pf->coeff_last[  DCT_LUMA_AC] = x264_coeff_last15;
    pf->coeff_last[ DCT_LUMA_4x4] = x264_coeff_last16;
    pf->coeff_last[ DCT_LUMA_8x8] = x264_coeff_last64;
    pf->coeff_level_run[DCT_CHROMA_DC] = x264_coeff_level_run4;
    pf->coeff_level_run[  DCT_LUMA_AC] = x264_coeff_level_run15;
    pf->coeff_level_run[ DCT_LUMA_4x4] = x264_coeff_level_run16;
#ifdef HAVE_MMX
    if( cpu&X264_CPU_MMX )
    {
#ifdef ARCH_X86
        pf->quant_4x4 = x264_quant_4x4_mmx;
        pf->quant_8x8 = x264_quant_8x8_mmx;
        pf->dequant_4x4 = x264_dequant_4x4_mmx;
        pf->dequant_4x4_dc = x264_dequant_4x4dc_mmxext;
        pf->dequant_8x8 = x264_dequant_8x8_mmx;
        if( h->param.i_cqm_preset == X264_CQM_FLAT )
        {
            pf->dequant_4x4 = x264_dequant_4x4_flat16_mmx;
            pf->dequant_8x8 = x264_dequant_8x8_flat16_mmx;
        }
        pf->denoise_dct = x264_denoise_dct_mmx;
#endif
    }
    if( cpu&X264_CPU_MMXEXT )
    {
        pf->quant_2x2_dc = x264_quant_2x2_dc_mmxext;
#ifdef ARCH_X86
        pf->quant_4x4_dc = x264_quant_4x4_dc_mmxext;
        pf->decimate_score15 = x264_decimate_score15_mmxext;
        pf->decimate_score16 = x264_decimate_score16_mmxext;
        pf->decimate_score64 = x264_decimate_score64_mmxext;
        pf->coeff_last[  DCT_LUMA_AC] = x264_coeff_last15_mmxext;
        pf->coeff_last[ DCT_LUMA_4x4] = x264_coeff_last16_mmxext;
        pf->coeff_last[ DCT_LUMA_8x8] = x264_coeff_last64_mmxext;
        pf->coeff_level_run[  DCT_LUMA_AC] = x264_coeff_level_run15_mmxext;
        pf->coeff_level_run[ DCT_LUMA_4x4] = x264_coeff_level_run16_mmxext;
#endif
        pf->coeff_last[DCT_CHROMA_DC] = x264_coeff_last4_mmxext;
        pf->coeff_level_run[DCT_CHROMA_DC] = x264_coeff_level_run4_mmxext;
        if( cpu&X264_CPU_LZCNT )
        {
            pf->coeff_last[DCT_CHROMA_DC] = x264_coeff_last4_mmxext_lzcnt;
            pf->coeff_level_run[DCT_CHROMA_DC] = x264_coeff_level_run4_mmxext_lzcnt;
        }
    }
    if( cpu&X264_CPU_SSE2 )
    {
        pf->quant_4x4_dc = x264_quant_4x4_dc_sse2;
        pf->quant_4x4 = x264_quant_4x4_sse2;
        pf->quant_8x8 = x264_quant_8x8_sse2;
        pf->dequant_4x4 = x264_dequant_4x4_sse2;
        pf->dequant_4x4_dc = x264_dequant_4x4dc_sse2;
        pf->dequant_8x8 = x264_dequant_8x8_sse2;
        if( h->param.i_cqm_preset == X264_CQM_FLAT )
        {
            pf->dequant_4x4 = x264_dequant_4x4_flat16_sse2;
            pf->dequant_8x8 = x264_dequant_8x8_flat16_sse2;
        }
        pf->denoise_dct = x264_denoise_dct_sse2;
        pf->decimate_score15 = x264_decimate_score15_sse2;
        pf->decimate_score16 = x264_decimate_score16_sse2;
        pf->decimate_score64 = x264_decimate_score64_sse2;
        pf->coeff_last[ DCT_LUMA_AC] = x264_coeff_last15_sse2;
        pf->coeff_last[DCT_LUMA_4x4] = x264_coeff_last16_sse2;
        pf->coeff_last[DCT_LUMA_8x8] = x264_coeff_last64_sse2;
        pf->coeff_level_run[ DCT_LUMA_AC] = x264_coeff_level_run15_sse2;
        pf->coeff_level_run[DCT_LUMA_4x4] = x264_coeff_level_run16_sse2;
        if( cpu&X264_CPU_LZCNT )
        {
            pf->coeff_last[ DCT_LUMA_AC] = x264_coeff_last15_sse2_lzcnt;
            pf->coeff_last[DCT_LUMA_4x4] = x264_coeff_last16_sse2_lzcnt;
            pf->coeff_last[DCT_LUMA_8x8] = x264_coeff_last64_sse2_lzcnt;
            pf->coeff_level_run[ DCT_LUMA_AC] = x264_coeff_level_run15_sse2_lzcnt;
            pf->coeff_level_run[DCT_LUMA_4x4] = x264_coeff_level_run16_sse2_lzcnt;
        }
    }
    if( cpu&X264_CPU_SSSE3 )
    {
        pf->quant_2x2_dc = x264_quant_2x2_dc_ssse3;
        pf->quant_4x4_dc = x264_quant_4x4_dc_ssse3;
        pf->quant_4x4 = x264_quant_4x4_ssse3;
        pf->quant_8x8 = x264_quant_8x8_ssse3;
        pf->denoise_dct = x264_denoise_dct_ssse3;
        pf->decimate_score15 = x264_decimate_score15_ssse3;
        pf->decimate_score16 = x264_decimate_score16_ssse3;
        pf->decimate_score64 = x264_decimate_score64_ssse3;
    }
    if( cpu&X264_CPU_SSE4 )
    {
        pf->quant_4x4_dc = x264_quant_4x4_dc_sse4;
        pf->quant_4x4 = x264_quant_4x4_sse4;
        pf->quant_8x8 = x264_quant_8x8_sse4;
    }
#endif // HAVE_MMX
#ifdef ARCH_PPC
    if( cpu&X264_CPU_ALTIVEC ) {
        pf->quant_2x2_dc = x264_quant_2x2_dc_altivec;
        pf->quant_4x4_dc = x264_quant_4x4_dc_altivec;
        pf->quant_4x4 = x264_quant_4x4_altivec;
        pf->quant_8x8 = x264_quant_8x8_altivec;
        pf->dequant_4x4 = x264_dequant_4x4_altivec;
        pf->dequant_8x8 = x264_dequant_8x8_altivec;
    }
#endif
#ifdef HAVE_ARMV6
    if( cpu&X264_CPU_ARMV6 )
        pf->coeff_last[DCT_CHROMA_DC] = x264_coeff_last4_arm;
    if( cpu&X264_CPU_NEON )
    {
        pf->quant_2x2_dc   = x264_quant_2x2_dc_neon;
        pf->quant_4x4      = x264_quant_4x4_neon;
        pf->quant_4x4_dc   = x264_quant_4x4_dc_neon;
        pf->quant_8x8      = x264_quant_8x8_neon;
        pf->dequant_4x4    = x264_dequant_4x4_neon;
        pf->dequant_4x4_dc = x264_dequant_4x4_dc_neon;
        pf->dequant_8x8    = x264_dequant_8x8_neon;
        pf->coeff_last[ DCT_LUMA_AC] = x264_coeff_last15_neon;
        pf->coeff_last[DCT_LUMA_4x4] = x264_coeff_last16_neon;
        pf->coeff_last[DCT_LUMA_8x8] = x264_coeff_last64_neon;
    }
#endif
    pf->coeff_last[  DCT_LUMA_DC] = pf->coeff_last[DCT_LUMA_4x4];
    pf->coeff_last[DCT_CHROMA_AC] = pf->coeff_last[ DCT_LUMA_AC];
    pf->coeff_level_run[  DCT_LUMA_DC] = pf->coeff_level_run[DCT_LUMA_4x4];
    pf->coeff_level_run[DCT_CHROMA_AC] = pf->coeff_level_run[ DCT_LUMA_AC];
}

						