blob: 9857a666ff380834c2e15543aea2e4a460d2dcf7 [file] [log] [blame]
/*
* Copyright (c) 2013 Paul B Mahol
*
* 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
*/
#include <float.h>
#include "libavutil/opt.h"
#include "libavutil/pixdesc.h"
#include "avfilter.h"
#include "drawutils.h"
#include "formats.h"
#include "internal.h"
#include "video.h"
#define R 0
#define G 1
#define B 2
#define A 3
typedef struct ThreadData {
AVFrame *in, *out;
} ThreadData;
typedef struct ColorChannelMixerContext {
const AVClass *class;
double rr, rg, rb, ra;
double gr, gg, gb, ga;
double br, bg, bb, ba;
double ar, ag, ab, aa;
double sr, sg, sb;
double preserve_lightness;
int *lut[4][4];
int *buffer;
uint8_t rgba_map[4];
int (*filter_slice[2])(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs);
} ColorChannelMixerContext;
#define OFFSET(x) offsetof(ColorChannelMixerContext, x)
#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_RUNTIME_PARAM
static const AVOption colorchannelmixer_options[] = {
{ "rr", "set the red gain for the red channel", OFFSET(rr), AV_OPT_TYPE_DOUBLE, {.dbl=1}, -2, 2, FLAGS },
{ "rg", "set the green gain for the red channel", OFFSET(rg), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "rb", "set the blue gain for the red channel", OFFSET(rb), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "ra", "set the alpha gain for the red channel", OFFSET(ra), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "gr", "set the red gain for the green channel", OFFSET(gr), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "gg", "set the green gain for the green channel", OFFSET(gg), AV_OPT_TYPE_DOUBLE, {.dbl=1}, -2, 2, FLAGS },
{ "gb", "set the blue gain for the green channel", OFFSET(gb), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "ga", "set the alpha gain for the green channel", OFFSET(ga), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "br", "set the red gain for the blue channel", OFFSET(br), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "bg", "set the green gain for the blue channel", OFFSET(bg), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "bb", "set the blue gain for the blue channel", OFFSET(bb), AV_OPT_TYPE_DOUBLE, {.dbl=1}, -2, 2, FLAGS },
{ "ba", "set the alpha gain for the blue channel", OFFSET(ba), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "ar", "set the red gain for the alpha channel", OFFSET(ar), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "ag", "set the green gain for the alpha channel", OFFSET(ag), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "ab", "set the blue gain for the alpha channel", OFFSET(ab), AV_OPT_TYPE_DOUBLE, {.dbl=0}, -2, 2, FLAGS },
{ "aa", "set the alpha gain for the alpha channel", OFFSET(aa), AV_OPT_TYPE_DOUBLE, {.dbl=1}, -2, 2, FLAGS },
{ "pl", "preserve lightness", OFFSET(preserve_lightness), AV_OPT_TYPE_DOUBLE, {.dbl=0}, 0, 1, FLAGS },
{ NULL }
};
AVFILTER_DEFINE_CLASS(colorchannelmixer);
static int query_formats(AVFilterContext *ctx)
{
static const enum AVPixelFormat pix_fmts[] = {
AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24,
AV_PIX_FMT_RGBA, AV_PIX_FMT_BGRA,
AV_PIX_FMT_ARGB, AV_PIX_FMT_ABGR,
AV_PIX_FMT_0RGB, AV_PIX_FMT_0BGR,
AV_PIX_FMT_RGB0, AV_PIX_FMT_BGR0,
AV_PIX_FMT_RGB48, AV_PIX_FMT_BGR48,
AV_PIX_FMT_RGBA64, AV_PIX_FMT_BGRA64,
AV_PIX_FMT_GBRP, AV_PIX_FMT_GBRAP,
AV_PIX_FMT_GBRP9,
AV_PIX_FMT_GBRP10, AV_PIX_FMT_GBRAP10,
AV_PIX_FMT_GBRP12, AV_PIX_FMT_GBRAP12,
AV_PIX_FMT_GBRP14,
AV_PIX_FMT_GBRP16, AV_PIX_FMT_GBRAP16,
AV_PIX_FMT_NONE
};
AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts);
if (!fmts_list)
return AVERROR(ENOMEM);
return ff_set_common_formats(ctx, fmts_list);
}
static float lerpf(float v0, float v1, float f)
{
return v0 + (v1 - v0) * f;
}
static void preservel(float *r, float *g, float *b, float lin, float lout)
{
*r *= lout / lin;
*g *= lout / lin;
*b *= lout / lin;
}
static av_always_inline int filter_slice_rgba_planar(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs,
int have_alpha, int pl)
{
ColorChannelMixerContext *s = ctx->priv;
ThreadData *td = arg;
AVFrame *in = td->in;
AVFrame *out = td->out;
const float l = s->preserve_lightness;
const float sr = s->sr;
const float sg = s->sg;
const float sb = s->sb;
const int slice_start = (out->height * jobnr) / nb_jobs;
const int slice_end = (out->height * (jobnr+1)) / nb_jobs;
const uint8_t *srcg = in->data[0] + slice_start * in->linesize[0];
const uint8_t *srcb = in->data[1] + slice_start * in->linesize[1];
const uint8_t *srcr = in->data[2] + slice_start * in->linesize[2];
const uint8_t *srca = in->data[3] + slice_start * in->linesize[3];
uint8_t *dstg = out->data[0] + slice_start * out->linesize[0];
uint8_t *dstb = out->data[1] + slice_start * out->linesize[1];
uint8_t *dstr = out->data[2] + slice_start * out->linesize[2];
uint8_t *dsta = out->data[3] + slice_start * out->linesize[3];
int i, j;
for (i = slice_start; i < slice_end; i++) {
for (j = 0; j < out->width; j++) {
const uint8_t rin = srcr[j];
const uint8_t gin = srcg[j];
const uint8_t bin = srcb[j];
const uint8_t ain = have_alpha ? srca[j] : 0;
int rout, gout, bout;
float lin;
if (pl)
lin = FFMAX3(rin, gin, bin) + FFMIN3(rin, gin, bin);
rout = s->lut[R][R][rin] +
s->lut[R][G][gin] +
s->lut[R][B][bin] +
(have_alpha == 1 ? s->lut[R][A][ain] : 0);
gout = s->lut[G][R][rin] +
s->lut[G][G][gin] +
s->lut[G][B][bin] +
(have_alpha == 1 ? s->lut[G][A][ain] : 0);
bout = s->lut[B][R][rin] +
s->lut[B][G][gin] +
s->lut[B][B][bin] +
(have_alpha == 1 ? s->lut[B][A][ain] : 0);
if (pl) {
float frout = rout / sr;
float fgout = gout / sg;
float fbout = bout / sb;
float lout = FFMAX3(frout, fgout, fbout) + FFMIN3(frout, fgout, fbout);
preservel(&frout, &fgout, &fbout, lin, lout);
rout = lrintf(lerpf(rout, frout, l));
gout = lrintf(lerpf(gout, fgout, l));
bout = lrintf(lerpf(bout, fbout, l));
}
dstr[j] = av_clip_uint8(rout);
dstg[j] = av_clip_uint8(gout);
dstb[j] = av_clip_uint8(bout);
if (have_alpha == 1) {
dsta[j] = av_clip_uint8(s->lut[A][R][rin] +
s->lut[A][G][gin] +
s->lut[A][B][bin] +
s->lut[A][A][ain]);
}
}
srcg += in->linesize[0];
srcb += in->linesize[1];
srcr += in->linesize[2];
srca += in->linesize[3];
dstg += out->linesize[0];
dstb += out->linesize[1];
dstr += out->linesize[2];
dsta += out->linesize[3];
}
return 0;
}
static av_always_inline int filter_slice_rgba16_planar(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs,
int have_alpha, int depth, int pl)
{
ColorChannelMixerContext *s = ctx->priv;
ThreadData *td = arg;
AVFrame *in = td->in;
AVFrame *out = td->out;
const float l = s->preserve_lightness;
const float sr = s->sr;
const float sg = s->sg;
const float sb = s->sb;
const int slice_start = (out->height * jobnr) / nb_jobs;
const int slice_end = (out->height * (jobnr+1)) / nb_jobs;
const uint16_t *srcg = (const uint16_t *)(in->data[0] + slice_start * in->linesize[0]);
const uint16_t *srcb = (const uint16_t *)(in->data[1] + slice_start * in->linesize[1]);
const uint16_t *srcr = (const uint16_t *)(in->data[2] + slice_start * in->linesize[2]);
const uint16_t *srca = (const uint16_t *)(in->data[3] + slice_start * in->linesize[3]);
uint16_t *dstg = (uint16_t *)(out->data[0] + slice_start * out->linesize[0]);
uint16_t *dstb = (uint16_t *)(out->data[1] + slice_start * out->linesize[1]);
uint16_t *dstr = (uint16_t *)(out->data[2] + slice_start * out->linesize[2]);
uint16_t *dsta = (uint16_t *)(out->data[3] + slice_start * out->linesize[3]);
int i, j;
for (i = slice_start; i < slice_end; i++) {
for (j = 0; j < out->width; j++) {
const uint16_t rin = srcr[j];
const uint16_t gin = srcg[j];
const uint16_t bin = srcb[j];
const uint16_t ain = have_alpha ? srca[j] : 0;
int rout, gout, bout;
float lin;
if (pl)
lin = FFMAX3(rin, gin, bin) + FFMIN3(rin, gin, bin);
rout = s->lut[R][R][rin] +
s->lut[R][G][gin] +
s->lut[R][B][bin] +
(have_alpha == 1 ? s->lut[R][A][ain] : 0);
gout = s->lut[G][R][rin] +
s->lut[G][G][gin] +
s->lut[G][B][bin] +
(have_alpha == 1 ? s->lut[G][A][ain] : 0);
bout = s->lut[B][R][rin] +
s->lut[B][G][gin] +
s->lut[B][B][bin] +
(have_alpha == 1 ? s->lut[B][A][ain] : 0);
if (pl) {
float frout = rout / sr;
float fgout = gout / sg;
float fbout = bout / sb;
float lout = FFMAX3(frout, fgout, fbout) + FFMIN3(frout, fgout, fbout);
preservel(&frout, &fgout, &fbout, lin, lout);
rout = lrintf(lerpf(rout, frout, l));
gout = lrintf(lerpf(gout, fgout, l));
bout = lrintf(lerpf(bout, fbout, l));
}
dstr[j] = av_clip_uintp2(rout, depth);
dstg[j] = av_clip_uintp2(gout, depth);
dstb[j] = av_clip_uintp2(bout, depth);
if (have_alpha == 1) {
dsta[j] = av_clip_uintp2(s->lut[A][R][rin] +
s->lut[A][G][gin] +
s->lut[A][B][bin] +
s->lut[A][A][ain], depth);
}
}
srcg += in->linesize[0] / 2;
srcb += in->linesize[1] / 2;
srcr += in->linesize[2] / 2;
srca += in->linesize[3] / 2;
dstg += out->linesize[0] / 2;
dstb += out->linesize[1] / 2;
dstr += out->linesize[2] / 2;
dsta += out->linesize[3] / 2;
}
return 0;
}
static int filter_slice_gbrp(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_planar(ctx, arg, jobnr, nb_jobs, 0, 0);
}
static int filter_slice_gbrap(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_planar(ctx, arg, jobnr, nb_jobs, 1, 0);
}
static int filter_slice_gbrp_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_planar(ctx, arg, jobnr, nb_jobs, 0, 1);
}
static int filter_slice_gbrap_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_planar(ctx, arg, jobnr, nb_jobs, 1, 1);
}
static int filter_slice_gbrp9(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 9, 0);
}
static int filter_slice_gbrp10(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 10, 0);
}
static int filter_slice_gbrap10(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 1, 10, 0);
}
static int filter_slice_gbrp12(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 12, 0);
}
static int filter_slice_gbrap12(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 1, 12, 0);
}
static int filter_slice_gbrp14(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 14, 0);
}
static int filter_slice_gbrp16(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 16, 0);
}
static int filter_slice_gbrap16(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 1, 16, 0);
}
static int filter_slice_gbrp9_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 9, 1);
}
static int filter_slice_gbrp10_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 10, 1);
}
static int filter_slice_gbrap10_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 1, 10, 1);
}
static int filter_slice_gbrp12_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 12, 1);
}
static int filter_slice_gbrap12_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 1, 12, 1);
}
static int filter_slice_gbrp14_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 14, 1);
}
static int filter_slice_gbrp16_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 0, 16, 1);
}
static int filter_slice_gbrap16_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_planar(ctx, arg, jobnr, nb_jobs, 1, 16, 1);
}
static av_always_inline int filter_slice_rgba_packed(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs,
int have_alpha, int step, int pl)
{
ColorChannelMixerContext *s = ctx->priv;
ThreadData *td = arg;
AVFrame *in = td->in;
AVFrame *out = td->out;
const float l = s->preserve_lightness;
const float sr = s->sr;
const float sg = s->sg;
const float sb = s->sb;
const int slice_start = (out->height * jobnr) / nb_jobs;
const int slice_end = (out->height * (jobnr+1)) / nb_jobs;
const uint8_t roffset = s->rgba_map[R];
const uint8_t goffset = s->rgba_map[G];
const uint8_t boffset = s->rgba_map[B];
const uint8_t aoffset = s->rgba_map[A];
const uint8_t *srcrow = in->data[0] + slice_start * in->linesize[0];
uint8_t *dstrow = out->data[0] + slice_start * out->linesize[0];
int i, j;
for (i = slice_start; i < slice_end; i++) {
const uint8_t *src = srcrow;
uint8_t *dst = dstrow;
for (j = 0; j < out->width * step; j += step) {
const uint8_t rin = src[j + roffset];
const uint8_t gin = src[j + goffset];
const uint8_t bin = src[j + boffset];
const uint8_t ain = src[j + aoffset];
int rout, gout, bout;
float lin;
if (pl)
lin = FFMAX3(rin, gin, bin) + FFMIN3(rin, gin, bin);
rout = s->lut[R][R][rin] +
s->lut[R][G][gin] +
s->lut[R][B][bin] +
(have_alpha == 1 ? s->lut[R][A][ain] : 0);
gout = s->lut[G][R][rin] +
s->lut[G][G][gin] +
s->lut[G][B][bin] +
(have_alpha == 1 ? s->lut[G][A][ain] : 0);
bout = s->lut[B][R][rin] +
s->lut[B][G][gin] +
s->lut[B][B][bin] +
(have_alpha == 1 ? s->lut[B][A][ain] : 0);
if (pl) {
float frout = rout / sr;
float fgout = gout / sg;
float fbout = bout / sb;
float lout = FFMAX3(frout, fgout, fbout) + FFMIN3(frout, fgout, fbout);
preservel(&frout, &fgout, &fbout, lin, lout);
rout = lrintf(lerpf(rout, frout, l));
gout = lrintf(lerpf(gout, fgout, l));
bout = lrintf(lerpf(bout, fbout, l));
}
dst[j + roffset] = av_clip_uint8(rout);
dst[j + goffset] = av_clip_uint8(gout);
dst[j + boffset] = av_clip_uint8(bout);
if (have_alpha == 1) {
dst[j + aoffset] = av_clip_uint8(s->lut[A][R][rin] +
s->lut[A][G][gin] +
s->lut[A][B][bin] +
s->lut[A][A][ain]);
} else if (have_alpha == -1 && in != out)
dst[j + aoffset] = 0;
}
srcrow += in->linesize[0];
dstrow += out->linesize[0];
}
return 0;
}
static av_always_inline int filter_slice_rgba16_packed(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs,
int have_alpha, int step, int pl)
{
ColorChannelMixerContext *s = ctx->priv;
ThreadData *td = arg;
AVFrame *in = td->in;
AVFrame *out = td->out;
const float l = s->preserve_lightness;
const float sr = s->sr;
const float sg = s->sg;
const float sb = s->sb;
const int slice_start = (out->height * jobnr) / nb_jobs;
const int slice_end = (out->height * (jobnr+1)) / nb_jobs;
const uint8_t roffset = s->rgba_map[R];
const uint8_t goffset = s->rgba_map[G];
const uint8_t boffset = s->rgba_map[B];
const uint8_t aoffset = s->rgba_map[A];
const uint8_t *srcrow = in->data[0] + slice_start * in->linesize[0];
uint8_t *dstrow = out->data[0] + slice_start * out->linesize[0];
int i, j;
for (i = slice_start; i < slice_end; i++) {
const uint16_t *src = (const uint16_t *)srcrow;
uint16_t *dst = (uint16_t *)dstrow;
for (j = 0; j < out->width * step; j += step) {
const uint16_t rin = src[j + roffset];
const uint16_t gin = src[j + goffset];
const uint16_t bin = src[j + boffset];
const uint16_t ain = src[j + aoffset];
int rout, gout, bout;
float lin;
if (pl)
lin = FFMAX3(rin, gin, bin) + FFMIN3(rin, gin, bin);
rout = s->lut[R][R][rin] +
s->lut[R][G][gin] +
s->lut[R][B][bin] +
(have_alpha == 1 ? s->lut[R][A][ain] : 0);
gout = s->lut[G][R][rin] +
s->lut[G][G][gin] +
s->lut[G][B][bin] +
(have_alpha == 1 ? s->lut[G][A][ain] : 0);
bout = s->lut[B][R][rin] +
s->lut[B][G][gin] +
s->lut[B][B][bin] +
(have_alpha == 1 ? s->lut[B][A][ain] : 0);
if (pl) {
float frout = rout / sr;
float fgout = gout / sg;
float fbout = bout / sb;
float lout = FFMAX3(frout, fgout, fbout) + FFMIN3(frout, fgout, fbout);
preservel(&frout, &fgout, &fbout, lin, lout);
rout = lrintf(lerpf(rout, frout, l));
gout = lrintf(lerpf(gout, fgout, l));
bout = lrintf(lerpf(bout, fbout, l));
}
dst[j + roffset] = av_clip_uint16(rout);
dst[j + goffset] = av_clip_uint16(gout);
dst[j + boffset] = av_clip_uint16(bout);
if (have_alpha == 1) {
dst[j + aoffset] = av_clip_uint16(s->lut[A][R][rin] +
s->lut[A][G][gin] +
s->lut[A][B][bin] +
s->lut[A][A][ain]);
}
}
srcrow += in->linesize[0];
dstrow += out->linesize[0];
}
return 0;
}
static int filter_slice_rgba64(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_packed(ctx, arg, jobnr, nb_jobs, 1, 4, 0);
}
static int filter_slice_rgb48(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_packed(ctx, arg, jobnr, nb_jobs, 0, 3, 0);
}
static int filter_slice_rgba64_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_packed(ctx, arg, jobnr, nb_jobs, 1, 4, 1);
}
static int filter_slice_rgb48_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba16_packed(ctx, arg, jobnr, nb_jobs, 0, 3, 1);
}
static int filter_slice_rgba(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_packed(ctx, arg, jobnr, nb_jobs, 1, 4, 0);
}
static int filter_slice_rgb24(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_packed(ctx, arg, jobnr, nb_jobs, 0, 3, 0);
}
static int filter_slice_rgb0(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_packed(ctx, arg, jobnr, nb_jobs, -1, 4, 0);
}
static int filter_slice_rgba_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_packed(ctx, arg, jobnr, nb_jobs, 1, 4, 1);
}
static int filter_slice_rgb24_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_packed(ctx, arg, jobnr, nb_jobs, 0, 3, 1);
}
static int filter_slice_rgb0_pl(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
return filter_slice_rgba_packed(ctx, arg, jobnr, nb_jobs, -1, 4, 1);
}
static int config_output(AVFilterLink *outlink)
{
AVFilterContext *ctx = outlink->src;
ColorChannelMixerContext *s = ctx->priv;
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(outlink->format);
const int depth = desc->comp[0].depth;
int i, j, size, *buffer = s->buffer;
ff_fill_rgba_map(s->rgba_map, outlink->format);
size = 1 << depth;
if (!s->buffer) {
s->buffer = buffer = av_malloc(16 * size * sizeof(*s->buffer));
if (!s->buffer)
return AVERROR(ENOMEM);
for (i = 0; i < 4; i++)
for (j = 0; j < 4; j++, buffer += size)
s->lut[i][j] = buffer;
}
s->sr = s->rr + s->rg + s->rb + s->ra;
s->sg = s->gr + s->gg + s->gb + s->ga;
s->sb = s->br + s->bg + s->bb + s->ba;
if (fabs(s->sr) <= DBL_EPSILON)
s->sr = 1.;
if (fabs(s->sg) <= DBL_EPSILON)
s->sg = 1.;
if (fabs(s->sb) <= DBL_EPSILON)
s->sb = 1.;
for (i = 0; i < size; i++) {
s->lut[R][R][i] = lrint(i * s->rr);
s->lut[R][G][i] = lrint(i * s->rg);
s->lut[R][B][i] = lrint(i * s->rb);
s->lut[R][A][i] = lrint(i * s->ra);
s->lut[G][R][i] = lrint(i * s->gr);
s->lut[G][G][i] = lrint(i * s->gg);
s->lut[G][B][i] = lrint(i * s->gb);
s->lut[G][A][i] = lrint(i * s->ga);
s->lut[B][R][i] = lrint(i * s->br);
s->lut[B][G][i] = lrint(i * s->bg);
s->lut[B][B][i] = lrint(i * s->bb);
s->lut[B][A][i] = lrint(i * s->ba);
s->lut[A][R][i] = lrint(i * s->ar);
s->lut[A][G][i] = lrint(i * s->ag);
s->lut[A][B][i] = lrint(i * s->ab);
s->lut[A][A][i] = lrint(i * s->aa);
}
switch (outlink->format) {
case AV_PIX_FMT_BGR24:
case AV_PIX_FMT_RGB24:
s->filter_slice[0] = filter_slice_rgb24;
s->filter_slice[1] = filter_slice_rgb24_pl;
break;
case AV_PIX_FMT_0BGR:
case AV_PIX_FMT_0RGB:
case AV_PIX_FMT_BGR0:
case AV_PIX_FMT_RGB0:
s->filter_slice[0] = filter_slice_rgb0;
s->filter_slice[1] = filter_slice_rgb0_pl;
break;
case AV_PIX_FMT_ABGR:
case AV_PIX_FMT_ARGB:
case AV_PIX_FMT_BGRA:
case AV_PIX_FMT_RGBA:
s->filter_slice[0] = filter_slice_rgba;
s->filter_slice[1] = filter_slice_rgba_pl;
break;
case AV_PIX_FMT_BGR48:
case AV_PIX_FMT_RGB48:
s->filter_slice[0] = filter_slice_rgb48;
s->filter_slice[1] = filter_slice_rgb48_pl;
break;
case AV_PIX_FMT_BGRA64:
case AV_PIX_FMT_RGBA64:
s->filter_slice[0] = filter_slice_rgba64;
s->filter_slice[1] = filter_slice_rgba64_pl;
break;
case AV_PIX_FMT_GBRP:
s->filter_slice[0] = filter_slice_gbrp;
s->filter_slice[1] = filter_slice_gbrp_pl;
break;
case AV_PIX_FMT_GBRAP:
s->filter_slice[0] = filter_slice_gbrap;
s->filter_slice[1] = filter_slice_gbrap_pl;
break;
case AV_PIX_FMT_GBRP9:
s->filter_slice[0] = filter_slice_gbrp9;
s->filter_slice[1] = filter_slice_gbrp9_pl;
break;
case AV_PIX_FMT_GBRP10:
s->filter_slice[0] = filter_slice_gbrp10;
s->filter_slice[1] = filter_slice_gbrp10_pl;
break;
case AV_PIX_FMT_GBRAP10:
s->filter_slice[0] = filter_slice_gbrap10;
s->filter_slice[1] = filter_slice_gbrap10_pl;
break;
case AV_PIX_FMT_GBRP12:
s->filter_slice[0] = filter_slice_gbrp12;
s->filter_slice[1] = filter_slice_gbrp12_pl;
break;
case AV_PIX_FMT_GBRAP12:
s->filter_slice[0] = filter_slice_gbrap12;
s->filter_slice[1] = filter_slice_gbrap12_pl;
break;
case AV_PIX_FMT_GBRP14:
s->filter_slice[0] = filter_slice_gbrp14;
s->filter_slice[1] = filter_slice_gbrp14_pl;
break;
case AV_PIX_FMT_GBRP16:
s->filter_slice[0] = filter_slice_gbrp16;
s->filter_slice[1] = filter_slice_gbrp16_pl;
break;
case AV_PIX_FMT_GBRAP16:
s->filter_slice[0] = filter_slice_gbrap16;
s->filter_slice[1] = filter_slice_gbrap16_pl;
break;
}
return 0;
}
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
AVFilterContext *ctx = inlink->dst;
ColorChannelMixerContext *s = ctx->priv;
AVFilterLink *outlink = ctx->outputs[0];
const int pl = s->preserve_lightness > 0.;
ThreadData td;
AVFrame *out;
if (av_frame_is_writable(in)) {
out = in;
} else {
out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
if (!out) {
av_frame_free(&in);
return AVERROR(ENOMEM);
}
av_frame_copy_props(out, in);
}
td.in = in;
td.out = out;
ctx->internal->execute(ctx, s->filter_slice[pl], &td, NULL, FFMIN(outlink->h, ff_filter_get_nb_threads(ctx)));
if (in != out)
av_frame_free(&in);
return ff_filter_frame(outlink, out);
}
static int process_command(AVFilterContext *ctx, const char *cmd, const char *args,
char *res, int res_len, int flags)
{
int ret = ff_filter_process_command(ctx, cmd, args, res, res_len, flags);
if (ret < 0)
return ret;
return config_output(ctx->outputs[0]);
}
static av_cold void uninit(AVFilterContext *ctx)
{
ColorChannelMixerContext *s = ctx->priv;
av_freep(&s->buffer);
}
static const AVFilterPad colorchannelmixer_inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
.filter_frame = filter_frame,
},
{ NULL }
};
static const AVFilterPad colorchannelmixer_outputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
.config_props = config_output,
},
{ NULL }
};
AVFilter ff_vf_colorchannelmixer = {
.name = "colorchannelmixer",
.description = NULL_IF_CONFIG_SMALL("Adjust colors by mixing color channels."),
.priv_size = sizeof(ColorChannelMixerContext),
.priv_class = &colorchannelmixer_class,
.uninit = uninit,
.query_formats = query_formats,
.inputs = colorchannelmixer_inputs,
.outputs = colorchannelmixer_outputs,
.flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC | AVFILTER_FLAG_SLICE_THREADS,
.process_command = process_command,
};