blob: 065d0936415b3bf0613d03f12665610a7cc39ae7
1 | /* |
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> |
5 | * |
6 | * This file is part of FFmpeg. |
7 | * |
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. |
12 | * |
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. |
17 | * |
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. |
21 | */ |
22 | |
23 | /** |
24 | * @file |
25 | * A very simple tv station logo remover |
26 | * Originally imported from MPlayer libmpcodecs/vf_delogo.c, |
27 | * the algorithm was later improved. |
28 | */ |
29 | |
30 | #include "libavutil/common.h" |
31 | #include "libavutil/imgutils.h" |
32 | #include "libavutil/opt.h" |
33 | #include "libavutil/pixdesc.h" |
34 | #include "avfilter.h" |
35 | #include "formats.h" |
36 | #include "internal.h" |
37 | #include "video.h" |
38 | |
39 | /** |
40 | * Apply a simple delogo algorithm to the image in src and put the |
41 | * result in dst. |
42 | * |
43 | * The algorithm is only applied to the region specified by the logo |
44 | * parameters. |
45 | * |
46 | * @param w width of the input image |
47 | * @param h height of the input image |
48 | * @param logo_x x coordinate of the top left corner of the logo region |
49 | * @param logo_y y coordinate of the top left corner of the logo region |
50 | * @param logo_w width of the logo |
51 | * @param logo_h height of the logo |
52 | * @param band the size of the band around the processed area |
53 | * @param show show a rectangle around the processed area, useful for |
54 | * parameters tweaking |
55 | * @param direct if non-zero perform in-place processing |
56 | */ |
57 | static void apply_delogo(uint8_t *dst, int dst_linesize, |
58 | uint8_t *src, int src_linesize, |
59 | int w, int h, AVRational sar, |
60 | int logo_x, int logo_y, int logo_w, int logo_h, |
61 | unsigned int band, int show, int direct) |
62 | { |
63 | int x, y; |
64 | uint64_t interp, weightl, weightr, weightt, weightb, weight; |
65 | uint8_t *xdst, *xsrc; |
66 | |
67 | uint8_t *topleft, *botleft, *topright; |
68 | unsigned int left_sample, right_sample; |
69 | int xclipl, xclipr, yclipt, yclipb; |
70 | int logo_x1, logo_x2, logo_y1, logo_y2; |
71 | |
72 | xclipl = FFMAX(-logo_x, 0); |
73 | xclipr = FFMAX(logo_x+logo_w-w, 0); |
74 | yclipt = FFMAX(-logo_y, 0); |
75 | yclipb = FFMAX(logo_y+logo_h-h, 0); |
76 | |
77 | logo_x1 = logo_x + xclipl; |
78 | logo_x2 = logo_x + logo_w - xclipr - 1; |
79 | logo_y1 = logo_y + yclipt; |
80 | logo_y2 = logo_y + logo_h - yclipb - 1; |
81 | |
82 | topleft = src+logo_y1 * src_linesize+logo_x1; |
83 | topright = src+logo_y1 * src_linesize+logo_x2; |
84 | botleft = src+logo_y2 * src_linesize+logo_x1; |
85 | |
86 | if (!direct) |
87 | av_image_copy_plane(dst, dst_linesize, src, src_linesize, w, h); |
88 | |
89 | dst += (logo_y1 + 1) * dst_linesize; |
90 | src += (logo_y1 + 1) * src_linesize; |
91 | |
92 | for (y = logo_y1+1; y < logo_y2; y++) { |
93 | left_sample = topleft[src_linesize*(y-logo_y1)] + |
94 | topleft[src_linesize*(y-logo_y1-1)] + |
95 | topleft[src_linesize*(y-logo_y1+1)]; |
96 | right_sample = topright[src_linesize*(y-logo_y1)] + |
97 | topright[src_linesize*(y-logo_y1-1)] + |
98 | topright[src_linesize*(y-logo_y1+1)]; |
99 | |
100 | for (x = logo_x1+1, |
101 | xdst = dst+logo_x1+1, |
102 | xsrc = src+logo_x1+1; x < logo_x2; x++, xdst++, xsrc++) { |
103 | |
104 | if (show && (y == logo_y1+1 || y == logo_y2-1 || |
105 | x == logo_x1+1 || x == logo_x2-1)) { |
106 | *xdst = 0; |
107 | continue; |
108 | } |
109 | |
110 | /* Weighted interpolation based on relative distances, taking SAR into account */ |
111 | weightl = (uint64_t) (logo_x2-x) * (y-logo_y1) * (logo_y2-y) * sar.den; |
112 | weightr = (uint64_t)(x-logo_x1) * (y-logo_y1) * (logo_y2-y) * sar.den; |
113 | weightt = (uint64_t)(x-logo_x1) * (logo_x2-x) * (logo_y2-y) * sar.num; |
114 | weightb = (uint64_t)(x-logo_x1) * (logo_x2-x) * (y-logo_y1) * sar.num; |
115 | |
116 | interp = |
117 | left_sample * weightl |
118 | + |
119 | right_sample * weightr |
120 | + |
121 | (topleft[x-logo_x1] + |
122 | topleft[x-logo_x1-1] + |
123 | topleft[x-logo_x1+1]) * weightt |
124 | + |
125 | (botleft[x-logo_x1] + |
126 | botleft[x-logo_x1-1] + |
127 | botleft[x-logo_x1+1]) * weightb; |
128 | weight = (weightl + weightr + weightt + weightb) * 3U; |
129 | interp = ROUNDED_DIV(interp, weight); |
130 | |
131 | if (y >= logo_y+band && y < logo_y+logo_h-band && |
132 | x >= logo_x+band && x < logo_x+logo_w-band) { |
133 | *xdst = interp; |
134 | } else { |
135 | unsigned dist = 0; |
136 | |
137 | if (x < logo_x+band) |
138 | dist = FFMAX(dist, logo_x-x+band); |
139 | else if (x >= logo_x+logo_w-band) |
140 | dist = FFMAX(dist, x-(logo_x+logo_w-1-band)); |
141 | |
142 | if (y < logo_y+band) |
143 | dist = FFMAX(dist, logo_y-y+band); |
144 | else if (y >= logo_y+logo_h-band) |
145 | dist = FFMAX(dist, y-(logo_y+logo_h-1-band)); |
146 | |
147 | *xdst = (*xsrc*dist + interp*(band-dist))/band; |
148 | } |
149 | } |
150 | |
151 | dst += dst_linesize; |
152 | src += src_linesize; |
153 | } |
154 | } |
155 | |
156 | typedef struct DelogoContext { |
157 | const AVClass *class; |
158 | int x, y, w, h, band, show; |
159 | } DelogoContext; |
160 | |
161 | #define OFFSET(x) offsetof(DelogoContext, x) |
162 | #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM |
163 | |
164 | static const AVOption delogo_options[]= { |
165 | { "x", "set logo x position", OFFSET(x), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, |
166 | { "y", "set logo y position", OFFSET(y), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, |
167 | { "w", "set logo width", OFFSET(w), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, |
168 | { "h", "set logo height", OFFSET(h), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, |
169 | #if LIBAVFILTER_VERSION_MAJOR < 7 |
170 | /* Actual default value for band/t is 1, set in init */ |
171 | { "band", "set delogo area band size", OFFSET(band), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS }, |
172 | { "t", "set delogo area band size", OFFSET(band), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS }, |
173 | #endif |
174 | { "show", "show delogo area", OFFSET(show), AV_OPT_TYPE_BOOL,{ .i64 = 0 }, 0, 1, FLAGS }, |
175 | { NULL } |
176 | }; |
177 | |
178 | AVFILTER_DEFINE_CLASS(delogo); |
179 | |
180 | static int query_formats(AVFilterContext *ctx) |
181 | { |
182 | static const enum AVPixelFormat pix_fmts[] = { |
183 | AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV420P, |
184 | AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV440P, |
185 | AV_PIX_FMT_YUVA420P, AV_PIX_FMT_GRAY8, |
186 | AV_PIX_FMT_NONE |
187 | }; |
188 | AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts); |
189 | if (!fmts_list) |
190 | return AVERROR(ENOMEM); |
191 | return ff_set_common_formats(ctx, fmts_list); |
192 | } |
193 | |
194 | static av_cold int init(AVFilterContext *ctx) |
195 | { |
196 | DelogoContext *s = ctx->priv; |
197 | |
198 | #define CHECK_UNSET_OPT(opt) \ |
199 | if (s->opt == -1) { \ |
200 | av_log(s, AV_LOG_ERROR, "Option %s was not set.\n", #opt); \ |
201 | return AVERROR(EINVAL); \ |
202 | } |
203 | CHECK_UNSET_OPT(x); |
204 | CHECK_UNSET_OPT(y); |
205 | CHECK_UNSET_OPT(w); |
206 | CHECK_UNSET_OPT(h); |
207 | |
208 | #if LIBAVFILTER_VERSION_MAJOR < 7 |
209 | if (s->band == 0) { /* Unset, use default */ |
210 | av_log(ctx, AV_LOG_WARNING, "Note: default band value was changed from 4 to 1.\n"); |
211 | s->band = 1; |
212 | } else if (s->band != 1) { |
213 | av_log(ctx, AV_LOG_WARNING, "Option band is deprecated.\n"); |
214 | } |
215 | #else |
216 | s->band = 1; |
217 | #endif |
218 | av_log(ctx, AV_LOG_VERBOSE, "x:%d y:%d, w:%d h:%d band:%d show:%d\n", |
219 | s->x, s->y, s->w, s->h, s->band, s->show); |
220 | |
221 | s->w += s->band*2; |
222 | s->h += s->band*2; |
223 | s->x -= s->band; |
224 | s->y -= s->band; |
225 | |
226 | return 0; |
227 | } |
228 | |
229 | static int config_input(AVFilterLink *inlink) |
230 | { |
231 | DelogoContext *s = inlink->dst->priv; |
232 | |
233 | /* Check whether the logo area fits in the frame */ |
234 | if (s->x + (s->band - 1) < 0 || s->x + s->w - (s->band*2 - 2) > inlink->w || |
235 | s->y + (s->band - 1) < 0 || s->y + s->h - (s->band*2 - 2) > inlink->h) { |
236 | av_log(s, AV_LOG_ERROR, "Logo area is outside of the frame.\n"); |
237 | return AVERROR(EINVAL); |
238 | } |
239 | |
240 | return 0; |
241 | } |
242 | |
243 | static int filter_frame(AVFilterLink *inlink, AVFrame *in) |
244 | { |
245 | DelogoContext *s = inlink->dst->priv; |
246 | AVFilterLink *outlink = inlink->dst->outputs[0]; |
247 | const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format); |
248 | AVFrame *out; |
249 | int hsub0 = desc->log2_chroma_w; |
250 | int vsub0 = desc->log2_chroma_h; |
251 | int direct = 0; |
252 | int plane; |
253 | AVRational sar; |
254 | |
255 | if (av_frame_is_writable(in)) { |
256 | direct = 1; |
257 | out = in; |
258 | } else { |
259 | out = ff_get_video_buffer(outlink, outlink->w, outlink->h); |
260 | if (!out) { |
261 | av_frame_free(&in); |
262 | return AVERROR(ENOMEM); |
263 | } |
264 | |
265 | av_frame_copy_props(out, in); |
266 | } |
267 | |
268 | sar = in->sample_aspect_ratio; |
269 | /* Assume square pixels if SAR is unknown */ |
270 | if (!sar.num) |
271 | sar.num = sar.den = 1; |
272 | |
273 | for (plane = 0; plane < desc->nb_components; plane++) { |
274 | int hsub = plane == 1 || plane == 2 ? hsub0 : 0; |
275 | int vsub = plane == 1 || plane == 2 ? vsub0 : 0; |
276 | |
277 | apply_delogo(out->data[plane], out->linesize[plane], |
278 | in ->data[plane], in ->linesize[plane], |
279 | AV_CEIL_RSHIFT(inlink->w, hsub), |
280 | AV_CEIL_RSHIFT(inlink->h, vsub), |
281 | sar, s->x>>hsub, s->y>>vsub, |
282 | /* Up and left borders were rounded down, inject lost bits |
283 | * into width and height to avoid error accumulation */ |
284 | AV_CEIL_RSHIFT(s->w + (s->x & ((1<<hsub)-1)), hsub), |
285 | AV_CEIL_RSHIFT(s->h + (s->y & ((1<<vsub)-1)), vsub), |
286 | s->band>>FFMIN(hsub, vsub), |
287 | s->show, direct); |
288 | } |
289 | |
290 | if (!direct) |
291 | av_frame_free(&in); |
292 | |
293 | return ff_filter_frame(outlink, out); |
294 | } |
295 | |
296 | static const AVFilterPad avfilter_vf_delogo_inputs[] = { |
297 | { |
298 | .name = "default", |
299 | .type = AVMEDIA_TYPE_VIDEO, |
300 | .filter_frame = filter_frame, |
301 | .config_props = config_input, |
302 | }, |
303 | { NULL } |
304 | }; |
305 | |
306 | static const AVFilterPad avfilter_vf_delogo_outputs[] = { |
307 | { |
308 | .name = "default", |
309 | .type = AVMEDIA_TYPE_VIDEO, |
310 | }, |
311 | { NULL } |
312 | }; |
313 | |
314 | AVFilter ff_vf_delogo = { |
315 | .name = "delogo", |
316 | .description = NULL_IF_CONFIG_SMALL("Remove logo from input video."), |
317 | .priv_size = sizeof(DelogoContext), |
318 | .priv_class = &delogo_class, |
319 | .init = init, |
320 | .query_formats = query_formats, |
321 | .inputs = avfilter_vf_delogo_inputs, |
322 | .outputs = avfilter_vf_delogo_outputs, |
323 | .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC, |
324 | }; |
325 |