h264idct.c

This the source release kit for the following system configuration(s): - AMD Alchemy(TM) DBAu1200(

/*
 * H.264 IDCT
 * Copyright (c) 2004 Michael Niedermayer <[e-mail]>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg 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
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 *
 */

/**
 * @file h264-idct.c
 * H.264 IDCT.
 * @author Michael Niedermayer <[e-mail]>
 */

#ifndef USE_ASM_VERSION /* [ */
#include "dsputil.h"

/*
#define IDCT_INTERNAL_BLOCK_I(i) \ 
   b0=  block[0 + block_stride*i];         \
   b1=  block[1+ block_stride*i];          \
   b2=  block[2 + block_stride*i];         \
   b3=  block[3 + block_stride*i];         \
    z0=  b0     +  b2;                      \
    z1=  b0     - b2;                       \
    z2= (b1>>1) - b3;                       \
    z3=  b1    + (b3>>1);                   \
     block[0 + block_stride*i]= z0 + z3;    \
     block[1 + block_stride*i]= z1 + z2;    \
     block[2 + block_stride*i]= z1 - z2;    \
     block[3 + block_stride*i]= z0 - z3;    

#define IDCT_INTERNAL_BLOCK_I(i) \ 
{ \
 DCTELEM *p = &block[block_stride*i];\
   b0=  p[0];         \
   b1=  p[1];          \
   b2=  p[2];         \
   b3=  p[3];         \
    z0=  b0     +  b2;                      \
    z1=  b0     - b2;                       \
    z2= (b1>>1) - b3;                       \
    z3=  b1    + (b3>>1);                   \
     p[0]= z0 + z3;    \
     p[1]= z1 + z2;    \
     p[2]= z1 - z2;    \
     p[3]= z0 - z3;    \
}

#define IDCT_INTERNAL_BLOCK_II(i) \ 
   b0=  block[i + block_stride*0];         \
   b1=  block[i+ block_stride*1];          \
   b2=  block[i + block_stride*2];         \
   b3=  block[i + block_stride*3];         \
    z0=  b0     +  b2;                      \
    z1=  b0     - b2;                       \
    z2= (b1>>1) - b3;                       \
    z3=  b1    + (b3>>1);                   \
        dst0=dst[i + 0*stride]; \
  dst1=dst[i + 1*stride];  \
  dst2=dst[i + 2*stride];  \
  dst3=dst[i + 3*stride];  \
        dst[i + 0*stride]= cm[ add*dst0+ ((z0 + z3) >> shift) ];\
        dst[i + 1*stride]= cm[ add*dst1 + ((z1 + z2) >> shift) ];\
        dst[i + 2*stride]= cm[ add*dst2 + ((z1 - z2) >> shift) ];\
        dst[i + 3*stride]= cm[ add*dst3 + ((z0 - z3) >> shift) ];
    

#define FF_H264_IDCT_DC_ADD_M(i) \ 
            dst[0+i*stride] = cm[ dst[0+i*stride] + dc ];  \
      dst[1+i*stride] = cm[ dst[1+i*stride] + dc ];  \
            dst[2+i*stride] = cm[ dst[2+i*stride] + dc ];  \
      dst[3+i*stride] = cm[ dst[3+i*stride] + dc ];  \
                        
#define FF_H264_IDCT8_DC_ADD_M(i) \
      dst[0+i*stride] = cm[ dst[0+i*stride] + dc ];  \
      dst[1+i*stride] = cm[ dst[1+i*stride] + dc ];  \
            dst[2+i*stride] = cm[ dst[2+i*stride] + dc ];  \
      dst[3+i*stride] = cm[ dst[3+i*stride] + dc ];  \
            dst[4+i*stride] = cm[ dst[4+i*stride] + dc ];  \
      dst[5+i*stride] = cm[ dst[5+i*stride] + dc ];  \
            dst[6+i*stride] = cm[ dst[6+i*stride] + dc ];  \
      dst[7+i*stride] = cm[ dst[7+i*stride] + dc ];  
      
      
     
#define H264_IDCT8_DC_ADD_I(i) \
 a0 =  src[i][0] + src[i][4];                                     \ 
 a2 =  src[i][0] - src[i][4];                                     \   
 a4 = (src[i][2]>>1) - src[i][6];                                 \   
 a6 = (src[i][6]>>1) + src[i][2];                                 \   
                                                                  \             
 b0 = a0 + a6;                                                    \   
 b2 = a2 + a4;                                                    \   
 b4 = a2 - a4;                                                    \   
 b6 = a0 - a6;                                                    \   
                                                                  \   
 a1 = -src[i][3] + src[i][5] - src[i][7] - (src[i][7]>>1);        \   
 a3 =  src[i][1] + src[i][7] - src[i][3] - (src[i][3]>>1);        \   
 a5 = -src[i][1] + src[i][7] + src[i][5] + (src[i][5]>>1);        \   
 a7 =  src[i][3] + src[i][5] + src[i][1] + (src[i][1]>>1);        \   
 b1 = (a7>>2) + a1;                                               \   
 b3 =  a3 + (a5>>2);                                              \   
 b5 = (a3>>2) - a5;                                               \   
 b7 =  a7 - (a1>>2);                                              \   
 src[i][0] = b0 + b7;                                             \             
 src[i][7] = b0 - b7;                                             \             
 src[i][1] = b2 + b5;                                             \             
 src[i][6] = b2 - b5;                                             \             
 src[i][2] = b4 + b3;                                             \             
 src[i][5] = b4 - b3;                                             \             
 src[i][3] = b6 + b1;                                             \             
 src[i][4] = b6 - b1;                                            
 
 #define H264_IDCT8_DC_ADD_II(i)\
 a0 =  src[0][i] + src[4][i];                                       \
a2 =  src[0][i] - src[4][i];                                        \
a4 = (src[2][i]>>1) - src[6][i];                                    \
a6 = (src[6][i]>>1) + src[2][i];                                    \
                                                                    \
b0 = a0 + a6;                                                       \
b2 = a2 + a4;                                                       \
b4 = a2 - a4;                                                       \
b6 = a0 - a6;                                                       \
                                                                    \
a1 = -src[3][i] + src[5][i] - src[7][i] - (src[7][i]>>1);           \
a3 =  src[1][i] + src[7][i] - src[3][i] - (src[3][i]>>1);           \
a5 = -src[1][i] + src[7][i] + src[5][i] + (src[5][i]>>1);           \
a7 =  src[3][i] + src[5][i] + src[1][i] + (src[1][i]>>1);           \
                                                                    \
b1 = (a7>>2) + a1;                                                  \
b3 =  a3 + (a5>>2);                                                 \
b5 = (a3>>2) - a5;                                                  \
b7 =  a7 - (a1>>2);                                                 \
dst[i + 0*stride] = cm[ dst[i + 0*stride] + ((b0 + b7) >> 6) ];     \
dst[i + 1*stride] = cm[ dst[i + 1*stride] + ((b2 + b5) >> 6) ];     \
dst[i + 2*stride] = cm[ dst[i + 2*stride] + ((b4 + b3) >> 6) ];     \
dst[i + 3*stride] = cm[ dst[i + 3*stride] + ((b6 + b1) >> 6) ];     \
dst[i + 4*stride] = cm[ dst[i + 4*stride] + ((b6 - b1) >> 6) ];     \
dst[i + 5*stride] = cm[ dst[i + 5*stride] + ((b4 - b3) >> 6) ];     \
dst[i + 6*stride] = cm[ dst[i + 6*stride] + ((b2 - b5) >> 6) ];     \
dst[i + 7*stride] = cm[ dst[i + 7*stride] + ((b0 - b7) >> 6) ];                
              
 */
STATIC_FUNC always_inline void idct_internal(uint8_t *dst, DCTELEM *block, int stride, int block_stride, int shift, int add){
    int i;
    uint8_t *cm = cropTbl + MAX_NEG_CROP;
        
    block[0] += 1<<(shift-1);
        
#if 1   
    for(i=0; i<4; i++){
  const int b0=  block[0 + block_stride*i];
  const int b1=  block[1+ block_stride*i];
  const int b2=  block[2 + block_stride*i];
  const int b3=  block[3 + block_stride*i];
        const int z0=  b0     +  b2;
        const int z1=  b0     - b2;
        const int z2= (b1>>1) - b3;
        const int z3=  b1    + (b3>>1);
        
        block[0 + block_stride*i]= z0 + z3;
        block[1 + block_stride*i]= z1 + z2;
        block[2 + block_stride*i]= z1 - z2;
        block[3 + block_stride*i]= z0 - z3;
    }   
        
  for(i=0; i<4; i++){
 const int b0=block[i + block_stride*0]; 
 const int b1=block[i + block_stride*1];
 const int b2=block[i + block_stride*2]; 
 const int b3=block[i + block_stride*3]; 
        const int z0=  b0    +  b2;
        const int z1=  b0     -  b2;
        const int z2= (b1>>1) -  b3;
        const int z3=  b1     + (b3>>1);
  const int dst0=dst[i + 0*stride];
  const int dst1=dst[i + 1*stride];
  const int dst2=dst[i + 2*stride];
  const int dst3=dst[i + 3*stride];
        dst[i + 0*stride]= cm[ add*dst0+ ((z0 + z3) >> shift) ];
        dst[i + 1*stride]= cm[ add*dst1 + ((z1 + z2) >> shift) ];
        dst[i + 2*stride]= cm[ add*dst2 + ((z1 - z2) >> shift) ];
        dst[i + 3*stride]= cm[ add*dst3 + ((z0 - z3) >> shift) ];
    }
#else
         
            register int b0;
            register int b1;
            register int b2;
            register int b3;
            register int z0;
            register int z1;
            register int z2;
            register int z3;
            register int dst0;
            register int dst1;
            register int dst2;
            register int dst3;
  IDCT_INTERNAL_BLOCK_I(0);
  IDCT_INTERNAL_BLOCK_I(1);
  IDCT_INTERNAL_BLOCK_I(2);
  IDCT_INTERNAL_BLOCK_I(3);
 
  IDCT_INTERNAL_BLOCK_II(0);
   IDCT_INTERNAL_BLOCK_II(1);
  IDCT_INTERNAL_BLOCK_II(2);
  IDCT_INTERNAL_BLOCK_II(3);
#endif
  
}

void ff_h264_idct_add_c(uint8_t *dst, DCTELEM *block, int stride){
    idct_internal(dst, block, stride, 4, 6, 1);
}