2 * Copyright (c) 2002 Jindrich Makovicka <makovick@gmail.com>
3 * Copyright (c) 2011 Stefano Sabatini
4 * Copyright (c) 2013, 2015 Jean Delvare <jdelvare@suse.com>
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License along
19 * with FFmpeg; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
25 * A very simple tv station logo remover
26 * Originally imported from MPlayer libmpcodecs/vf_delogo.c,
27 * the algorithm was later improved.
30 #include "libavutil/common.h"
31 #include "libavutil/imgutils.h"
32 #include "libavutil/opt.h"
33 #include "libavutil/pixdesc.h"
34 #include "libavutil/eval.h"
39 static const char * const var_names[] = {
44 "n", ///< number of frame
45 "t", ///< timestamp expressed in seconds
58 #define TS2T(ts, tb) ((ts) == AV_NOPTS_VALUE ? NAN : (double)(ts) * av_q2d(tb))
60 static int set_expr(AVExpr **pexpr, const char *expr, const char *option, void *log_ctx)
67 ret = av_expr_parse(pexpr, expr, var_names, NULL, NULL, NULL, NULL, 0, log_ctx);
69 av_log(log_ctx, AV_LOG_ERROR,
70 "Error when parsing the expression '%s' for %s\n",
82 * Apply a simple delogo algorithm to the image in src and put the
85 * The algorithm is only applied to the region specified by the logo
88 * @param w width of the input image
89 * @param h height of the input image
90 * @param logo_x x coordinate of the top left corner of the logo region
91 * @param logo_y y coordinate of the top left corner of the logo region
92 * @param logo_w width of the logo
93 * @param logo_h height of the logo
94 * @param band the size of the band around the processed area
95 * @param show show a rectangle around the processed area, useful for
97 * @param direct if non-zero perform in-place processing
99 static void apply_delogo(uint8_t *dst, int dst_linesize,
100 uint8_t *src, int src_linesize,
101 int w, int h, AVRational sar,
102 int logo_x, int logo_y, int logo_w, int logo_h,
103 unsigned int band, int show, int direct)
106 uint64_t interp, weightl, weightr, weightt, weightb, weight;
107 uint8_t *xdst, *xsrc;
109 uint8_t *topleft, *botleft, *topright;
110 unsigned int left_sample, right_sample;
111 int xclipl, xclipr, yclipt, yclipb;
112 int logo_x1, logo_x2, logo_y1, logo_y2;
114 xclipl = FFMAX(-logo_x, 0);
115 xclipr = FFMAX(logo_x+logo_w-w, 0);
116 yclipt = FFMAX(-logo_y, 0);
117 yclipb = FFMAX(logo_y+logo_h-h, 0);
119 logo_x1 = logo_x + xclipl;
120 logo_x2 = logo_x + logo_w - xclipr - 1;
121 logo_y1 = logo_y + yclipt;
122 logo_y2 = logo_y + logo_h - yclipb - 1;
124 topleft = src+logo_y1 * src_linesize+logo_x1;
125 topright = src+logo_y1 * src_linesize+logo_x2;
126 botleft = src+logo_y2 * src_linesize+logo_x1;
129 av_image_copy_plane(dst, dst_linesize, src, src_linesize, w, h);
131 dst += (logo_y1 + 1) * dst_linesize;
132 src += (logo_y1 + 1) * src_linesize;
134 for (y = logo_y1+1; y < logo_y2; y++) {
135 left_sample = topleft[src_linesize*(y-logo_y1)] +
136 topleft[src_linesize*(y-logo_y1-1)] +
137 topleft[src_linesize*(y-logo_y1+1)];
138 right_sample = topright[src_linesize*(y-logo_y1)] +
139 topright[src_linesize*(y-logo_y1-1)] +
140 topright[src_linesize*(y-logo_y1+1)];
143 xdst = dst+logo_x1+1,
144 xsrc = src+logo_x1+1; x < logo_x2; x++, xdst++, xsrc++) {
146 if (show && (y == logo_y1+1 || y == logo_y2-1 ||
147 x == logo_x1+1 || x == logo_x2-1)) {
152 /* Weighted interpolation based on relative distances, taking SAR into account */
153 weightl = (uint64_t) (logo_x2-x) * (y-logo_y1) * (logo_y2-y) * sar.den;
154 weightr = (uint64_t)(x-logo_x1) * (y-logo_y1) * (logo_y2-y) * sar.den;
155 weightt = (uint64_t)(x-logo_x1) * (logo_x2-x) * (logo_y2-y) * sar.num;
156 weightb = (uint64_t)(x-logo_x1) * (logo_x2-x) * (y-logo_y1) * sar.num;
159 left_sample * weightl
161 right_sample * weightr
163 (topleft[x-logo_x1] +
164 topleft[x-logo_x1-1] +
165 topleft[x-logo_x1+1]) * weightt
167 (botleft[x-logo_x1] +
168 botleft[x-logo_x1-1] +
169 botleft[x-logo_x1+1]) * weightb;
170 weight = (weightl + weightr + weightt + weightb) * 3U;
171 interp = (interp + (weight >> 1)) / weight;
173 if (y >= logo_y+band && y < logo_y+logo_h-band &&
174 x >= logo_x+band && x < logo_x+logo_w-band) {
180 dist = FFMAX(dist, logo_x-x+band);
181 else if (x >= logo_x+logo_w-band)
182 dist = FFMAX(dist, x-(logo_x+logo_w-1-band));
185 dist = FFMAX(dist, logo_y-y+band);
186 else if (y >= logo_y+logo_h-band)
187 dist = FFMAX(dist, y-(logo_y+logo_h-1-band));
189 *xdst = (*xsrc*dist + interp*(band-dist))/band;
198 typedef struct DelogoContext {
199 const AVClass *class;
200 int x, y, w, h, band, show;
201 char *x_expr, *y_expr, *w_expr, *h_expr;
202 AVExpr *x_pexpr, *y_pexpr, *w_pexpr, *h_pexpr;
203 double var_values[VAR_VARS_NB];
206 #define OFFSET(x) offsetof(DelogoContext, x)
207 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
209 static const AVOption delogo_options[]= {
210 { "x", "set logo x position", OFFSET(x_expr), AV_OPT_TYPE_STRING, { .str = "-1" }, CHAR_MIN, CHAR_MAX, FLAGS },
211 { "y", "set logo y position", OFFSET(y_expr), AV_OPT_TYPE_STRING, { .str = "-1" }, CHAR_MIN, CHAR_MAX, FLAGS },
212 { "w", "set logo width", OFFSET(w_expr), AV_OPT_TYPE_STRING, { .str = "-1" }, CHAR_MIN, CHAR_MAX, FLAGS },
213 { "h", "set logo height", OFFSET(h_expr), AV_OPT_TYPE_STRING, { .str = "-1" }, CHAR_MIN, CHAR_MAX, FLAGS },
214 #if LIBAVFILTER_VERSION_MAJOR < 7
215 /* Actual default value for band/t is 1, set in init */
216 { "band", "set delogo area band size", OFFSET(band), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS },
217 { "t", "set delogo area band size", OFFSET(band), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS },
219 { "show", "show delogo area", OFFSET(show), AV_OPT_TYPE_BOOL,{ .i64 = 0 }, 0, 1, FLAGS },
223 AVFILTER_DEFINE_CLASS(delogo);
224 static av_cold void uninit(AVFilterContext *ctx)
226 DelogoContext *s = ctx->priv;
228 av_expr_free(s->x_pexpr); s->x_pexpr = NULL;
229 av_expr_free(s->y_pexpr); s->y_pexpr = NULL;
230 av_expr_free(s->w_pexpr); s->w_pexpr = NULL;
231 av_expr_free(s->h_pexpr); s->h_pexpr = NULL;
235 static int query_formats(AVFilterContext *ctx)
237 static const enum AVPixelFormat pix_fmts[] = {
238 AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV420P,
239 AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV440P,
240 AV_PIX_FMT_YUVA420P, AV_PIX_FMT_GRAY8,
243 AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts);
245 return AVERROR(ENOMEM);
246 return ff_set_common_formats(ctx, fmts_list);
249 static av_cold int init(AVFilterContext *ctx)
251 DelogoContext *s = ctx->priv;
254 if ((ret = set_expr(&s->x_pexpr, s->x_expr, "x", ctx)) < 0 ||
255 (ret = set_expr(&s->y_pexpr, s->y_expr, "y", ctx)) < 0 ||
256 (ret = set_expr(&s->w_pexpr, s->w_expr, "w", ctx)) < 0 ||
257 (ret = set_expr(&s->h_pexpr, s->h_expr, "h", ctx)) < 0 )
260 s->x = av_expr_eval(s->x_pexpr, s->var_values, s);
261 s->y = av_expr_eval(s->y_pexpr, s->var_values, s);
262 s->w = av_expr_eval(s->w_pexpr, s->var_values, s);
263 s->h = av_expr_eval(s->h_pexpr, s->var_values, s);
265 #define CHECK_UNSET_OPT(opt) \
266 if (s->opt == -1) { \
267 av_log(s, AV_LOG_ERROR, "Option %s was not set.\n", #opt); \
268 return AVERROR(EINVAL); \
275 #if LIBAVFILTER_VERSION_MAJOR < 7
276 if (s->band == 0) { /* Unset, use default */
277 av_log(ctx, AV_LOG_WARNING, "Note: default band value was changed from 4 to 1.\n");
279 } else if (s->band != 1) {
280 av_log(ctx, AV_LOG_WARNING, "Option band is deprecated.\n");
285 av_log(ctx, AV_LOG_VERBOSE, "x:%d y:%d, w:%d h:%d band:%d show:%d\n",
286 s->x, s->y, s->w, s->h, s->band, s->show);
296 static int config_input(AVFilterLink *inlink)
298 DelogoContext *s = inlink->dst->priv;
300 /* Check whether the logo area fits in the frame */
301 if (s->x + (s->band - 1) < 0 || s->x + s->w - (s->band*2 - 2) > inlink->w ||
302 s->y + (s->band - 1) < 0 || s->y + s->h - (s->band*2 - 2) > inlink->h) {
303 av_log(s, AV_LOG_ERROR, "Logo area is outside of the frame.\n");
304 return AVERROR(EINVAL);
310 static int filter_frame(AVFilterLink *inlink, AVFrame *in)
312 DelogoContext *s = inlink->dst->priv;
313 AVFilterLink *outlink = inlink->dst->outputs[0];
314 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
316 int hsub0 = desc->log2_chroma_w;
317 int vsub0 = desc->log2_chroma_h;
323 s->var_values[VAR_N] = inlink->frame_count_out;
324 s->var_values[VAR_T] = TS2T(in->pts, inlink->time_base);
325 s->x = av_expr_eval(s->x_pexpr, s->var_values, s);
326 s->y = av_expr_eval(s->y_pexpr, s->var_values, s);
327 s->w = av_expr_eval(s->w_pexpr, s->var_values, s);
328 s->h = av_expr_eval(s->h_pexpr, s->var_values, s);
330 if (s->x + (s->band - 1) <= 0 || s->x + s->w - (s->band*2 - 2) > inlink->w ||
331 s->y + (s->band - 1) <= 0 || s->y + s->h - (s->band*2 - 2) > inlink->h) {
332 av_log(s, AV_LOG_WARNING, "Logo area is outside of the frame,"
333 " auto set the area inside of the frame\n");
336 if (s->x + (s->band - 1) <= 0)
338 if (s->y + (s->band - 1) <= 0)
340 if (s->x + s->w - (s->band*2 - 2) > inlink->w)
341 s->w = inlink->w - s->x - (s->band*2 - 2);
342 if (s->y + s->h - (s->band*2 - 2) > inlink->h)
343 s->h = inlink->h - s->y - (s->band*2 - 2);
345 ret = config_input(inlink);
356 if (av_frame_is_writable(in)) {
360 out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
363 return AVERROR(ENOMEM);
366 av_frame_copy_props(out, in);
369 sar = in->sample_aspect_ratio;
370 /* Assume square pixels if SAR is unknown */
372 sar.num = sar.den = 1;
374 for (plane = 0; plane < desc->nb_components; plane++) {
375 int hsub = plane == 1 || plane == 2 ? hsub0 : 0;
376 int vsub = plane == 1 || plane == 2 ? vsub0 : 0;
378 apply_delogo(out->data[plane], out->linesize[plane],
379 in ->data[plane], in ->linesize[plane],
380 AV_CEIL_RSHIFT(inlink->w, hsub),
381 AV_CEIL_RSHIFT(inlink->h, vsub),
382 sar, s->x>>hsub, s->y>>vsub,
383 /* Up and left borders were rounded down, inject lost bits
384 * into width and height to avoid error accumulation */
385 AV_CEIL_RSHIFT(s->w + (s->x & ((1<<hsub)-1)), hsub),
386 AV_CEIL_RSHIFT(s->h + (s->y & ((1<<vsub)-1)), vsub),
387 s->band>>FFMIN(hsub, vsub),
394 return ff_filter_frame(outlink, out);
397 static const AVFilterPad avfilter_vf_delogo_inputs[] = {
400 .type = AVMEDIA_TYPE_VIDEO,
401 .filter_frame = filter_frame,
402 .config_props = config_input,
407 static const AVFilterPad avfilter_vf_delogo_outputs[] = {
410 .type = AVMEDIA_TYPE_VIDEO,
415 AVFilter ff_vf_delogo = {
417 .description = NULL_IF_CONFIG_SMALL("Remove logo from input video."),
418 .priv_size = sizeof(DelogoContext),
419 .priv_class = &delogo_class,
422 .query_formats = query_formats,
423 .inputs = avfilter_vf_delogo_inputs,
424 .outputs = avfilter_vf_delogo_outputs,
425 .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC,