avfilter/vf_colorlevels: refactor code so all components are processed in same loop

This is also faster.
This commit is contained in:
Paul B Mahol 2021-09-26 13:02:03 +02:00
parent 8983c3d7e4
commit 4727e30ec3

View File

@ -104,56 +104,74 @@ typedef struct ThreadData {
int omin[4];
} ThreadData;
#define LOAD_COMMON \
ColorLevelsContext *s = ctx->priv; \
const ThreadData *td = arg; \
const int process_h = td->h; \
const int slice_start = (process_h * jobnr ) / nb_jobs; \
const int slice_end = (process_h * (jobnr+1)) / nb_jobs; \
const uint8_t *srcrow = td->srcrow; \
uint8_t *dstrow = td->dstrow; \
const int step = s->step;
#define DO_COMMON(type, clip) \
ColorLevelsContext *s = ctx->priv; \
const ThreadData *td = arg; \
const int linesize = s->linesize; \
const int step = s->step; \
const int process_h = td->h; \
const int slice_start = (process_h * jobnr ) / nb_jobs; \
const int slice_end = (process_h * (jobnr+1)) / nb_jobs; \
const int src_linesize = td->src_linesize / sizeof(type); \
const int dst_linesize = td->dst_linesize / sizeof(type); \
const type *srcrow = (const type *)td->srcrow + src_linesize * slice_start; \
type *dstrow = (type *)td->dstrow + dst_linesize * slice_start; \
const uint8_t offset_r = s->rgba_map[R]; \
const uint8_t offset_g = s->rgba_map[G]; \
const uint8_t offset_b = s->rgba_map[B]; \
const uint8_t offset_a = s->rgba_map[A]; \
const int imin_r = td->imin[R]; \
const int imin_g = td->imin[G]; \
const int imin_b = td->imin[B]; \
const int imin_a = td->imin[A]; \
const int omin_r = td->omin[R]; \
const int omin_g = td->omin[G]; \
const int omin_b = td->omin[B]; \
const int omin_a = td->omin[A]; \
const float coeff_r = td->coeff[R]; \
const float coeff_g = td->coeff[G]; \
const float coeff_b = td->coeff[B]; \
const float coeff_a = td->coeff[A]; \
const type *src_r = srcrow + offset_r; \
const type *src_g = srcrow + offset_g; \
const type *src_b = srcrow + offset_b; \
const type *src_a = srcrow + offset_a; \
type *dst_r = dstrow + offset_r; \
type *dst_g = dstrow + offset_g; \
type *dst_b = dstrow + offset_b; \
type *dst_a = dstrow + offset_a; \
\
for (int y = slice_start; y < slice_end; y++) { \
for (int x = 0; x < linesize; x += step) { \
dst_r[x] = clip((src_r[x] - imin_r) * coeff_r + omin_r); \
dst_g[x] = clip((src_g[x] - imin_g) * coeff_g + omin_g); \
dst_b[x] = clip((src_b[x] - imin_b) * coeff_b + omin_b); \
} \
\
for (int x = 0; x < linesize && s->nb_comp == 4; x += step) \
dst_a[x] = clip((src_a[x] - imin_a) * coeff_a + omin_a); \
\
src_r += src_linesize; \
src_g += src_linesize; \
src_b += src_linesize; \
src_a += src_linesize; \
\
dst_r += dst_linesize; \
dst_g += dst_linesize; \
dst_b += dst_linesize; \
dst_a += dst_linesize; \
}
static int colorlevels_slice_8(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
LOAD_COMMON
for (int comp = 0; comp < s->nb_comp; comp++) {
const uint8_t offset = s->rgba_map[comp];
const int imin = td->imin[comp];
const int omin = td->omin[comp];
const float coeff = td->coeff[comp];
for (int y = slice_start; y < slice_end; y++) {
const uint8_t *src = srcrow + y * td->src_linesize;
uint8_t *dst = dstrow + y * td->dst_linesize;
for (int x = 0; x < s->linesize; x += step)
dst[x + offset] = av_clip_uint8((src[x + offset] - imin) * coeff + omin);
}
}
DO_COMMON(uint8_t, av_clip_uint8)
return 0;
}
static int colorlevels_slice_16(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
LOAD_COMMON
for (int comp = 0; comp < s->nb_comp; comp++) {
const uint8_t offset = s->rgba_map[comp];
const int imin = td->imin[comp];
const int omin = td->omin[comp];
const float coeff = td->coeff[comp];
for (int y = slice_start; y < slice_end; y++) {
const uint16_t *src = (const uint16_t *)(srcrow + y * td->src_linesize);
uint16_t *dst = (uint16_t *)(dstrow + y * td->dst_linesize);
for (int x = 0; x < s->linesize; x += step)
dst[x + offset] = av_clip_uint16((src[x + offset] - imin) * coeff + omin);
}
}
DO_COMMON(uint16_t, av_clip_uint16)
return 0;
}