har0ke/FFmpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Correct rounding values in overlap filtering

Browse Source 
Originally committed as revision 7171 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
Kostya Shishkov 2006-11-26 04:57:31 +00:00
parent 5081f3aad5
commit 61f5b14a8e
3 changed files with 36 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
libavcodec/dsputil.h
View File

@ -397,8 +397,8 @@ typedef struct DSPContext {
void (*vc1_inv_trans_8x4)(DCTELEM *b, int n); void (*vc1_inv_trans_8x4)(DCTELEM *b, int n);
void (*vc1_inv_trans_4x8)(DCTELEM *b, int n); void (*vc1_inv_trans_4x8)(DCTELEM *b, int n);
void (*vc1_inv_trans_4x4)(DCTELEM *b, int n); void (*vc1_inv_trans_4x4)(DCTELEM *b, int n);
void (*vc1_v_overlap)(uint8_t* src, int stride, int rnd); void (*vc1_v_overlap)(uint8_t* src, int stride);
void (*vc1_h_overlap)(uint8_t* src, int stride, int rnd); void (*vc1_h_overlap)(uint8_t* src, int stride);
/* put 8x8 block with bicubic interpolation and quarterpel precision /* put 8x8 block with bicubic interpolation and quarterpel precision
* last argument is actually round value instead of height * last argument is actually round value instead of height
*/ */

56
libavcodec/vc1.c
View File

@ -3397,9 +3397,9 @@ static int vc1_decode_p_mb(VC1Context *v)
s->dsp.put_pixels_clamped(s->block[i], s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2)); s->dsp.put_pixels_clamped(s->block[i], s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
if(v->pq >= 9 && v->overlap) { if(v->pq >= 9 && v->overlap) {
if(v->c_avail) if(v->c_avail)
s->dsp.vc1_h_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), (i<4) ? (i&1) : (s->mb_x&1)); s->dsp.vc1_h_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
if(v->a_avail) if(v->a_avail)
s->dsp.vc1_v_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), (i<4) ? ((i&1)>>1) : (s->mb_y&1)); s->dsp.vc1_v_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
} }
} else if(val) { } else if(val) {
vc1_decode_p_block(v, s->block[i], i, mquant, ttmb, first_block); vc1_decode_p_block(v, s->block[i], i, mquant, ttmb, first_block);
@ -3500,9 +3500,9 @@ static int vc1_decode_p_mb(VC1Context *v)
s->dsp.put_pixels_clamped(s->block[i], s->dest[dst_idx] + off, (i&4)?s->uvlinesize:s->linesize); s->dsp.put_pixels_clamped(s->block[i], s->dest[dst_idx] + off, (i&4)?s->uvlinesize:s->linesize);
if(v->pq >= 9 && v->overlap) { if(v->pq >= 9 && v->overlap) {
if(v->c_avail) if(v->c_avail)
s->dsp.vc1_h_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), (i<4) ? (i&1) : (s->mb_x&1)); s->dsp.vc1_h_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
if(v->a_avail) if(v->a_avail)
s->dsp.vc1_v_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), (i<4) ? ((i&1)>>1) : (s->mb_y&1)); s->dsp.vc1_v_overlap(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
} }
} else if(is_coded[i]) { } else if(is_coded[i]) {
status = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb, first_block); status = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb, first_block);
@ -3769,25 +3769,25 @@ static void vc1_decode_i_blocks(VC1Context *v)
vc1_put_block(v, s->block); vc1_put_block(v, s->block);
if(v->pq >= 9 && v->overlap) { if(v->pq >= 9 && v->overlap) {
if(s->mb_x) { if(s->mb_x) {
s->dsp.vc1_h_overlap(s->dest[0], s->linesize, 0); s->dsp.vc1_h_overlap(s->dest[0], s->linesize);
s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize, s->linesize, 0); s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize, s->linesize);
if(!(s->flags & CODEC_FLAG_GRAY)) { if(!(s->flags & CODEC_FLAG_GRAY)) {
s->dsp.vc1_h_overlap(s->dest[1], s->uvlinesize, s->mb_x&1); s->dsp.vc1_h_overlap(s->dest[1], s->uvlinesize);
s->dsp.vc1_h_overlap(s->dest[2], s->uvlinesize, s->mb_x&1); s->dsp.vc1_h_overlap(s->dest[2], s->uvlinesize);
} }
} }
s->dsp.vc1_h_overlap(s->dest[0] + 8, s->linesize, 1); s->dsp.vc1_h_overlap(s->dest[0] + 8, s->linesize);
s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize, 1); s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize);
if(!s->first_slice_line) { if(!s->first_slice_line) {
s->dsp.vc1_v_overlap(s->dest[0], s->linesize, 0); s->dsp.vc1_v_overlap(s->dest[0], s->linesize);
s->dsp.vc1_v_overlap(s->dest[0] + 8, s->linesize, 0); s->dsp.vc1_v_overlap(s->dest[0] + 8, s->linesize);
if(!(s->flags & CODEC_FLAG_GRAY)) { if(!(s->flags & CODEC_FLAG_GRAY)) {
s->dsp.vc1_v_overlap(s->dest[1], s->uvlinesize, s->mb_y&1); s->dsp.vc1_v_overlap(s->dest[1], s->uvlinesize);
s->dsp.vc1_v_overlap(s->dest[2], s->uvlinesize, s->mb_y&1); s->dsp.vc1_v_overlap(s->dest[2], s->uvlinesize);
} }
} }
s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize, s->linesize, 1); s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize, s->linesize);
s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize, 1); s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize);
} }
if(get_bits_count(&s->gb) > v->bits) { if(get_bits_count(&s->gb) > v->bits) {
@ -3898,25 +3898,25 @@ static void vc1_decode_i_blocks_adv(VC1Context *v)
vc1_put_block(v, s->block); vc1_put_block(v, s->block);
if(overlap) { if(overlap) {
if(s->mb_x) { if(s->mb_x) {
s->dsp.vc1_h_overlap(s->dest[0], s->linesize, 0); s->dsp.vc1_h_overlap(s->dest[0], s->linesize);
s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize, s->linesize, 0); s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize, s->linesize);
if(!(s->flags & CODEC_FLAG_GRAY)) { if(!(s->flags & CODEC_FLAG_GRAY)) {
s->dsp.vc1_h_overlap(s->dest[1], s->uvlinesize, s->mb_x&1); s->dsp.vc1_h_overlap(s->dest[1], s->uvlinesize);
s->dsp.vc1_h_overlap(s->dest[2], s->uvlinesize, s->mb_x&1); s->dsp.vc1_h_overlap(s->dest[2], s->uvlinesize);
} }
} }
s->dsp.vc1_h_overlap(s->dest[0] + 8, s->linesize, 1); s->dsp.vc1_h_overlap(s->dest[0] + 8, s->linesize);
s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize, 1); s->dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize);
if(!s->first_slice_line) { if(!s->first_slice_line) {
s->dsp.vc1_v_overlap(s->dest[0], s->linesize, 0); s->dsp.vc1_v_overlap(s->dest[0], s->linesize);
s->dsp.vc1_v_overlap(s->dest[0] + 8, s->linesize, 0); s->dsp.vc1_v_overlap(s->dest[0] + 8, s->linesize);
if(!(s->flags & CODEC_FLAG_GRAY)) { if(!(s->flags & CODEC_FLAG_GRAY)) {
s->dsp.vc1_v_overlap(s->dest[1], s->uvlinesize, s->mb_y&1); s->dsp.vc1_v_overlap(s->dest[1], s->uvlinesize);
s->dsp.vc1_v_overlap(s->dest[2], s->uvlinesize, s->mb_y&1); s->dsp.vc1_v_overlap(s->dest[2], s->uvlinesize);
} }
} }
s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize, s->linesize, 1); s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize, s->linesize);
s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize, 1); s->dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize);
} }
if(get_bits_count(&s->gb) > v->bits) { if(get_bits_count(&s->gb) > v->bits) {

8
libavcodec/vc1dsp.c
View File

@ -31,11 +31,12 @@
/** Apply overlap transform to horizontal edge /** Apply overlap transform to horizontal edge
*/ */
static void vc1_v_overlap_c(uint8_t* src, int stride, int rnd) static void vc1_v_overlap_c(uint8_t* src, int stride)
{ {
int i; int i;
int a, b, c, d; int a, b, c, d;
int d1, d2; int d1, d2;
int rnd = 1;
for(i = 0; i < 8; i++) { for(i = 0; i < 8; i++) {
a = src[-2*stride]; a = src[-2*stride];
b = src[-stride]; b = src[-stride];
@ -49,16 +50,18 @@ static void vc1_v_overlap_c(uint8_t* src, int stride, int rnd)
src[0] = c + d2; src[0] = c + d2;
src[stride] = d + d1; src[stride] = d + d1;
src++; src++;
rnd = !rnd;
} }
} }
/** Apply overlap transform to vertical edge /** Apply overlap transform to vertical edge
*/ */
static void vc1_h_overlap_c(uint8_t* src, int stride, int rnd) static void vc1_h_overlap_c(uint8_t* src, int stride)
{ {
int i; int i;
int a, b, c, d; int a, b, c, d;
int d1, d2; int d1, d2;
int rnd = 1;
for(i = 0; i < 8; i++) { for(i = 0; i < 8; i++) {
a = src[-2]; a = src[-2];
b = src[-1]; b = src[-1];
@ -72,6 +75,7 @@ static void vc1_h_overlap_c(uint8_t* src, int stride, int rnd)
src[0] = c + d2; src[0] = c + d2;
src[1] = d + d1; src[1] = d + d1;
src += stride; src += stride;
rnd = !rnd;
} }
} }