* @author Michael Niedermayer <michaelni@gmx.at>
*/
-#include "dsputil.h"
+#include "bit_depth_template.c"
#ifndef AVCODEC_H264IDCT_INTERNAL_H
#define AVCODEC_H264IDCT_INTERNAL_H
//FIXME this table is a duplicate from h264data.h, and will be removed once the tables from, h264 have been split
-static const uint8_t scan8[16 + 2*4]={
- 4+1*8, 5+1*8, 4+2*8, 5+2*8,
- 6+1*8, 7+1*8, 6+2*8, 7+2*8,
- 4+3*8, 5+3*8, 4+4*8, 5+4*8,
- 6+3*8, 7+3*8, 6+4*8, 7+4*8,
- 1+1*8, 2+1*8,
- 1+2*8, 2+2*8,
- 1+4*8, 2+4*8,
- 1+5*8, 2+5*8,
+static const uint8_t scan8[16*3]={
+ 4+ 1*8, 5+ 1*8, 4+ 2*8, 5+ 2*8,
+ 6+ 1*8, 7+ 1*8, 6+ 2*8, 7+ 2*8,
+ 4+ 3*8, 5+ 3*8, 4+ 4*8, 5+ 4*8,
+ 6+ 3*8, 7+ 3*8, 6+ 4*8, 7+ 4*8,
+ 4+ 6*8, 5+ 6*8, 4+ 7*8, 5+ 7*8,
+ 6+ 6*8, 7+ 6*8, 6+ 7*8, 7+ 7*8,
+ 4+ 8*8, 5+ 8*8, 4+ 9*8, 5+ 9*8,
+ 6+ 8*8, 7+ 8*8, 6+ 9*8, 7+ 9*8,
+ 4+11*8, 5+11*8, 4+12*8, 5+12*8,
+ 6+11*8, 7+11*8, 6+12*8, 7+12*8,
+ 4+13*8, 5+13*8, 4+14*8, 5+14*8,
+ 6+13*8, 7+13*8, 6+14*8, 7+14*8
};
#endif
-static av_always_inline void idct_internal(uint8_t *dst, DCTELEM *block, int stride, int block_stride, int shift, int add){
+void FUNCC(ff_h264_idct_add)(uint8_t *_dst, DCTELEM *_block, int stride)
+{
int i;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
+ INIT_CLIP
+ pixel *dst = (pixel*)_dst;
+ dctcoef *block = (dctcoef*)_block;
+ stride /= sizeof(pixel);
- block[0] += 1<<(shift-1);
+ block[0] += 1 << 5;
for(i=0; i<4; i++){
- const int z0= block[i + block_stride*0] + block[i + block_stride*2];
- const int z1= block[i + block_stride*0] - block[i + block_stride*2];
- const int z2= (block[i + block_stride*1]>>1) - block[i + block_stride*3];
- const int z3= block[i + block_stride*1] + (block[i + block_stride*3]>>1);
-
- block[i + block_stride*0]= z0 + z3;
- block[i + block_stride*1]= z1 + z2;
- block[i + block_stride*2]= z1 - z2;
- block[i + block_stride*3]= z0 - z3;
+ const int z0= block[i + 4*0] + block[i + 4*2];
+ const int z1= block[i + 4*0] - block[i + 4*2];
+ const int z2= (block[i + 4*1]>>1) - block[i + 4*3];
+ const int z3= block[i + 4*1] + (block[i + 4*3]>>1);
+
+ block[i + 4*0]= z0 + z3;
+ block[i + 4*1]= z1 + z2;
+ block[i + 4*2]= z1 - z2;
+ block[i + 4*3]= z0 - z3;
}
for(i=0; i<4; i++){
- const int z0= block[0 + block_stride*i] + block[2 + block_stride*i];
- const int z1= block[0 + block_stride*i] - block[2 + block_stride*i];
- const int z2= (block[1 + block_stride*i]>>1) - block[3 + block_stride*i];
- const int z3= block[1 + block_stride*i] + (block[3 + block_stride*i]>>1);
-
- dst[i + 0*stride]= cm[ add*dst[i + 0*stride] + ((z0 + z3) >> shift) ];
- dst[i + 1*stride]= cm[ add*dst[i + 1*stride] + ((z1 + z2) >> shift) ];
- dst[i + 2*stride]= cm[ add*dst[i + 2*stride] + ((z1 - z2) >> shift) ];
- dst[i + 3*stride]= cm[ add*dst[i + 3*stride] + ((z0 - z3) >> shift) ];
+ const int z0= block[0 + 4*i] + block[2 + 4*i];
+ const int z1= block[0 + 4*i] - block[2 + 4*i];
+ const int z2= (block[1 + 4*i]>>1) - block[3 + 4*i];
+ const int z3= block[1 + 4*i] + (block[3 + 4*i]>>1);
+
+ dst[i + 0*stride]= CLIP(dst[i + 0*stride] + ((z0 + z3) >> 6));
+ dst[i + 1*stride]= CLIP(dst[i + 1*stride] + ((z1 + z2) >> 6));
+ dst[i + 2*stride]= CLIP(dst[i + 2*stride] + ((z1 - z2) >> 6));
+ dst[i + 3*stride]= CLIP(dst[i + 3*stride] + ((z0 - z3) >> 6));
}
}
-void ff_h264_idct_add_c(uint8_t *dst, DCTELEM *block, int stride){
- idct_internal(dst, block, stride, 4, 6, 1);
-}
-
-void ff_h264_lowres_idct_add_c(uint8_t *dst, int stride, DCTELEM *block){
- idct_internal(dst, block, stride, 8, 3, 1);
-}
-
-void ff_h264_lowres_idct_put_c(uint8_t *dst, int stride, DCTELEM *block){
- idct_internal(dst, block, stride, 8, 3, 0);
-}
-
-void ff_h264_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride){
+void FUNCC(ff_h264_idct8_add)(uint8_t *_dst, DCTELEM *_block, int stride){
int i;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
+ INIT_CLIP
+ pixel *dst = (pixel*)_dst;
+ dctcoef *block = (dctcoef*)_block;
+ stride /= sizeof(pixel);
block[0] += 32;
const int b5 = (a3>>2) - a5;
const int 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) ];
+ dst[i + 0*stride] = CLIP( dst[i + 0*stride] + ((b0 + b7) >> 6) );
+ dst[i + 1*stride] = CLIP( dst[i + 1*stride] + ((b2 + b5) >> 6) );
+ dst[i + 2*stride] = CLIP( dst[i + 2*stride] + ((b4 + b3) >> 6) );
+ dst[i + 3*stride] = CLIP( dst[i + 3*stride] + ((b6 + b1) >> 6) );
+ dst[i + 4*stride] = CLIP( dst[i + 4*stride] + ((b6 - b1) >> 6) );
+ dst[i + 5*stride] = CLIP( dst[i + 5*stride] + ((b4 - b3) >> 6) );
+ dst[i + 6*stride] = CLIP( dst[i + 6*stride] + ((b2 - b5) >> 6) );
+ dst[i + 7*stride] = CLIP( dst[i + 7*stride] + ((b0 - b7) >> 6) );
}
}
// assumes all AC coefs are 0
-void ff_h264_idct_dc_add_c(uint8_t *dst, DCTELEM *block, int stride){
+void FUNCC(ff_h264_idct_dc_add)(uint8_t *_dst, DCTELEM *block, int stride){
int i, j;
- int dc = (block[0] + 32) >> 6;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP + dc;
+ int dc = (((dctcoef*)block)[0] + 32) >> 6;
+ INIT_CLIP
+ pixel *dst = (pixel*)_dst;
+ stride /= sizeof(pixel);
for( j = 0; j < 4; j++ )
{
for( i = 0; i < 4; i++ )
- dst[i] = cm[ dst[i] ];
+ dst[i] = CLIP( dst[i] + dc );
dst += stride;
}
}
-void ff_h264_idct8_dc_add_c(uint8_t *dst, DCTELEM *block, int stride){
+void FUNCC(ff_h264_idct8_dc_add)(uint8_t *_dst, DCTELEM *block, int stride){
int i, j;
- int dc = (block[0] + 32) >> 6;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP + dc;
+ int dc = (((dctcoef*)block)[0] + 32) >> 6;
+ INIT_CLIP
+ pixel *dst = (pixel*)_dst;
+ stride /= sizeof(pixel);
for( j = 0; j < 8; j++ )
{
for( i = 0; i < 8; i++ )
- dst[i] = cm[ dst[i] ];
+ dst[i] = CLIP( dst[i] + dc );
dst += stride;
}
}
-void ff_h264_idct_add16_c(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+void FUNCC(ff_h264_idct_add16)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
int i;
for(i=0; i<16; i++){
int nnz = nnzc[ scan8[i] ];
if(nnz){
- if(nnz==1 && block[i*16]) ff_h264_idct_dc_add_c(dst + block_offset[i], block + i*16, stride);
- else idct_internal (dst + block_offset[i], block + i*16, stride, 4, 6, 1);
+ if(nnz==1 && ((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
+ else FUNCC(ff_h264_idct_add )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
}
}
}
-void ff_h264_idct_add16intra_c(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+void FUNCC(ff_h264_idct_add16intra)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
int i;
for(i=0; i<16; i++){
- if(nnzc[ scan8[i] ]) idct_internal (dst + block_offset[i], block + i*16, stride, 4, 6, 1);
- else if(block[i*16]) ff_h264_idct_dc_add_c(dst + block_offset[i], block + i*16, stride);
+ if(nnzc[ scan8[i] ]) FUNCC(ff_h264_idct_add )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
+ else if(((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
}
}
-void ff_h264_idct8_add4_c(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+void FUNCC(ff_h264_idct8_add4)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
int i;
for(i=0; i<16; i+=4){
int nnz = nnzc[ scan8[i] ];
if(nnz){
- if(nnz==1 && block[i*16]) ff_h264_idct8_dc_add_c(dst + block_offset[i], block + i*16, stride);
- else ff_h264_idct8_add_c (dst + block_offset[i], block + i*16, stride);
+ if(nnz==1 && ((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct8_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
+ else FUNCC(ff_h264_idct8_add )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
}
}
}
-void ff_h264_idct_add8_c(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
- int i;
- for(i=16; i<16+8; i++){
- if(nnzc[ scan8[i] ])
- ff_h264_idct_add_c (dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
- else if(block[i*16])
- ff_h264_idct_dc_add_c(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
+void FUNCC(ff_h264_idct_add8)(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
+ int i, j;
+ for(j=1; j<3; j++){
+ for(i=j*16; i<j*16+4; i++){
+ if(nnzc[ scan8[i] ])
+ FUNCC(ff_h264_idct_add )(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
+ else if(((dctcoef*)block)[i*16])
+ FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
+ }
}
}
/**
* IDCT transforms the 16 dc values and dequantizes them.
- * @param qp quantization parameter
+ * @param qmul quantization parameter
*/
-void ff_h264_luma_dc_dequant_idct_c(DCTELEM *output, DCTELEM *input, int qmul){
+void FUNCC(ff_h264_luma_dc_dequant_idct)(DCTELEM *_output, DCTELEM *_input, int qmul){
#define stride 16
int i;
int temp[16];
static const uint8_t x_offset[4]={0, 2*stride, 8*stride, 10*stride};
+ dctcoef *input = (dctcoef*)_input;
+ dctcoef *output = (dctcoef*)_output;
for(i=0; i<4; i++){
const int z0= input[4*i+0] + input[4*i+1];
#undef stride
}
-void ff_h264_chroma_dc_dequant_idct_c(DCTELEM *block, int qmul){
+void FUNCC(ff_h264_chroma_dc_dequant_idct)(DCTELEM *_block, int qmul){
const int stride= 16*2;
const int xStride= 16;
int a,b,c,d,e;
+ dctcoef *block = (dctcoef*)_block;
a= block[stride*0 + xStride*0];
b= block[stride*0 + xStride*1];