mirror of
https://git.ffmpeg.org/ffmpeg.git
synced 2024-12-24 16:22:37 +00:00
rewrite h261 loop filter
no malloc(64) memcpy free stuff no filter1 A->B then filter2 A->B (yes not B->A) no incorrect rouding after the 1d filter Originally committed as revision 3177 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
parent
c6148de232
commit
fdbbf2e0fc
@ -2360,32 +2360,29 @@ static void h263_h_loop_filter_c(uint8_t *src, int stride, int qscale){
|
||||
}
|
||||
}
|
||||
|
||||
static void h261_v_loop_filter_c(uint8_t *dest,uint8_t *src, int stride){
|
||||
int i,j,xy,yz;
|
||||
int res;
|
||||
for(i=0; i<8; i++){
|
||||
for(j=1; j<7; j++){
|
||||
xy = j * stride + i;
|
||||
yz = j * 8 + i;
|
||||
res = (int)src[yz-1*8] + ((int)(src[yz+0*8]) * 2) + (int)src[yz+1*8];
|
||||
res +=2;
|
||||
res >>=2;
|
||||
dest[xy] = (uint8_t)res;
|
||||
static void h261_loop_filter_c(uint8_t *src, int stride){
|
||||
int x,y,xy,yz;
|
||||
int temp[64];
|
||||
|
||||
for(x=0; x<8; x++){
|
||||
temp[x ] = 4*src[x ];
|
||||
temp[x + 7*8] = 4*src[x + 7*stride];
|
||||
}
|
||||
for(y=1; y<7; y++){
|
||||
for(x=0; x<8; x++){
|
||||
xy = y * stride + x;
|
||||
yz = y * 8 + x;
|
||||
temp[yz] = src[xy - stride] + 2*src[xy] + src[xy + stride];
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
static void h261_h_loop_filter_c(uint8_t *dest,uint8_t *src, int stride){
|
||||
int i,j,xy,yz;
|
||||
int res;
|
||||
for(i=1; i<7; i++){
|
||||
for(j=0; j<8; j++){
|
||||
xy = j * stride + i;
|
||||
yz = j * 8 + i;
|
||||
res = (int)src[yz-1] + ((int)(src[yz]) *2) + (int)src[yz+1];
|
||||
res+=2;
|
||||
res>>=2;
|
||||
dest[xy] = (uint8_t)res;
|
||||
|
||||
for(y=0; y<8; y++){
|
||||
src[ y*stride] = (temp[ y*8] + 2)>>2;
|
||||
src[7+y*stride] = (temp[7+y*8] + 2)>>2;
|
||||
for(x=1; x<7; x++){
|
||||
xy = y * stride + x;
|
||||
yz = y * 8 + x;
|
||||
src[xy] = (temp[yz-1] + 2*temp[yz] + temp[yz+1] + 8)>>4;
|
||||
}
|
||||
}
|
||||
}
|
||||
@ -3325,8 +3322,7 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
|
||||
c->h263_h_loop_filter= h263_h_loop_filter_c;
|
||||
c->h263_v_loop_filter= h263_v_loop_filter_c;
|
||||
|
||||
c->h261_h_loop_filter= h261_h_loop_filter_c;
|
||||
c->h261_v_loop_filter= h261_v_loop_filter_c;
|
||||
c->h261_loop_filter= h261_loop_filter_c;
|
||||
|
||||
c->try_8x8basis= try_8x8basis_c;
|
||||
c->add_8x8basis= add_8x8basis_c;
|
||||
|
@ -261,8 +261,7 @@ typedef struct DSPContext {
|
||||
void (*h263_v_loop_filter)(uint8_t *src, int stride, int qscale);
|
||||
void (*h263_h_loop_filter)(uint8_t *src, int stride, int qscale);
|
||||
|
||||
void (*h261_v_loop_filter)(uint8_t *dest,uint8_t *src, int stride);
|
||||
void (*h261_h_loop_filter)(uint8_t *dest,uint8_t *src, int stride);
|
||||
void (*h261_loop_filter)(uint8_t *src, int stride);
|
||||
|
||||
/* (I)DCT */
|
||||
void (*fdct)(DCTELEM *block/* align 16*/);
|
||||
|
@ -64,44 +64,13 @@ void ff_h261_loop_filter(H261Context * h){
|
||||
uint8_t *dest_y = s->dest[0];
|
||||
uint8_t *dest_cb= s->dest[1];
|
||||
uint8_t *dest_cr= s->dest[2];
|
||||
uint8_t *src;
|
||||
|
||||
CHECKED_ALLOCZ((src),sizeof(uint8_t) * 64 );
|
||||
|
||||
for(i=0; i<8;i++)
|
||||
memcpy(src+i*8,dest_y+i*linesize,sizeof(uint8_t) * 8 );
|
||||
s->dsp.h261_v_loop_filter(dest_y, src, linesize);
|
||||
s->dsp.h261_h_loop_filter(dest_y, src, linesize);
|
||||
|
||||
for(i=0; i<8;i++)
|
||||
memcpy(src+i*8,dest_y+i*linesize + 8,sizeof(uint8_t) * 8 );
|
||||
s->dsp.h261_v_loop_filter(dest_y + 8, src, linesize);
|
||||
s->dsp.h261_h_loop_filter(dest_y + 8, src, linesize);
|
||||
|
||||
for(i=0; i<8;i++)
|
||||
memcpy(src+i*8,dest_y+(i+8)*linesize,sizeof(uint8_t) * 8 );
|
||||
s->dsp.h261_v_loop_filter(dest_y + 8 * linesize, src, linesize);
|
||||
s->dsp.h261_h_loop_filter(dest_y + 8 * linesize, src, linesize);
|
||||
|
||||
for(i=0; i<8;i++)
|
||||
memcpy(src+i*8,dest_y+(i+8)*linesize + 8,sizeof(uint8_t) * 8 );
|
||||
s->dsp.h261_v_loop_filter(dest_y + 8 * linesize + 8, src, linesize);
|
||||
s->dsp.h261_h_loop_filter(dest_y + 8 * linesize + 8, src, linesize);
|
||||
|
||||
for(i=0; i<8;i++)
|
||||
memcpy(src+i*8,dest_cb+i*uvlinesize,sizeof(uint8_t) * 8 );
|
||||
s->dsp.h261_v_loop_filter(dest_cb, src, uvlinesize);
|
||||
s->dsp.h261_h_loop_filter(dest_cb, src, uvlinesize);
|
||||
|
||||
for(i=0; i<8;i++)
|
||||
memcpy(src+i*8,dest_cr+i*uvlinesize,sizeof(uint8_t) * 8 );
|
||||
s->dsp.h261_v_loop_filter(dest_cr, src, uvlinesize);
|
||||
s->dsp.h261_h_loop_filter(dest_cr, src, uvlinesize);
|
||||
|
||||
fail:
|
||||
av_free(src);
|
||||
|
||||
return;
|
||||
|
||||
s->dsp.h261_loop_filter(dest_y , linesize);
|
||||
s->dsp.h261_loop_filter(dest_y + 8, linesize);
|
||||
s->dsp.h261_loop_filter(dest_y + 8 * linesize , linesize);
|
||||
s->dsp.h261_loop_filter(dest_y + 8 * linesize + 8, linesize);
|
||||
s->dsp.h261_loop_filter(dest_cb, uvlinesize);
|
||||
s->dsp.h261_loop_filter(dest_cr, uvlinesize);
|
||||
}
|
||||
|
||||
static int h261_decode_block(H261Context *h, DCTELEM *block,
|
||||
|
Loading…
Reference in New Issue
Block a user