Commit | Line | Data |
---|---|---|
2ba45a60 DM |
1 | /* |
2 | * Copyright (c) 2002 Jindrich Makovicka <makovick@gmail.com> | |
3 | * Copyright (c) 2011 Stefano Sabatini | |
4 | * Copyright (c) 2013 Jean Delvare <khali@linux-fr.org> | |
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; | |
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; | |
79 | logo_y1 = logo_y + yclipt; | |
80 | logo_y2 = logo_y + logo_h - yclipb; | |
81 | ||
82 | topleft = src+logo_y1 * src_linesize+logo_x1; | |
83 | topright = src+logo_y1 * src_linesize+logo_x2-1; | |
84 | botleft = src+(logo_y2-1) * 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_y2-1; 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_x2-1; x++, xdst++, xsrc++) { | |
103 | ||
104 | /* Weighted interpolation based on relative distances, taking SAR into account */ | |
105 | weightl = (uint64_t) (logo_x2-1-x) * (y-logo_y1) * (logo_y2-1-y) * sar.den; | |
106 | weightr = (uint64_t)(x-logo_x1) * (y-logo_y1) * (logo_y2-1-y) * sar.den; | |
107 | weightt = (uint64_t)(x-logo_x1) * (logo_x2-1-x) * (logo_y2-1-y) * sar.num; | |
108 | weightb = (uint64_t)(x-logo_x1) * (logo_x2-1-x) * (y-logo_y1) * sar.num; | |
109 | ||
110 | interp = | |
111 | left_sample * weightl | |
112 | + | |
113 | right_sample * weightr | |
114 | + | |
115 | (topleft[x-logo_x1] + | |
116 | topleft[x-logo_x1-1] + | |
117 | topleft[x-logo_x1+1]) * weightt | |
118 | + | |
119 | (botleft[x-logo_x1] + | |
120 | botleft[x-logo_x1-1] + | |
121 | botleft[x-logo_x1+1]) * weightb; | |
122 | interp /= (weightl + weightr + weightt + weightb) * 3U; | |
123 | ||
124 | if (y >= logo_y+band && y < logo_y+logo_h-band && | |
125 | x >= logo_x+band && x < logo_x+logo_w-band) { | |
126 | *xdst = interp; | |
127 | } else { | |
128 | unsigned dist = 0; | |
129 | ||
130 | if (x < logo_x+band) | |
131 | dist = FFMAX(dist, logo_x-x+band); | |
132 | else if (x >= logo_x+logo_w-band) | |
133 | dist = FFMAX(dist, x-(logo_x+logo_w-1-band)); | |
134 | ||
135 | if (y < logo_y+band) | |
136 | dist = FFMAX(dist, logo_y-y+band); | |
137 | else if (y >= logo_y+logo_h-band) | |
138 | dist = FFMAX(dist, y-(logo_y+logo_h-1-band)); | |
139 | ||
140 | *xdst = (*xsrc*dist + interp*(band-dist))/band; | |
141 | if (show && (dist == band-1)) | |
142 | *xdst = 0; | |
143 | } | |
144 | } | |
145 | ||
146 | dst += dst_linesize; | |
147 | src += src_linesize; | |
148 | } | |
149 | } | |
150 | ||
151 | typedef struct DelogoContext { | |
152 | const AVClass *class; | |
153 | int x, y, w, h, band, show; | |
154 | } DelogoContext; | |
155 | ||
156 | #define OFFSET(x) offsetof(DelogoContext, x) | |
157 | #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM | |
158 | ||
159 | static const AVOption delogo_options[]= { | |
160 | { "x", "set logo x position", OFFSET(x), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, | |
161 | { "y", "set logo y position", OFFSET(y), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, | |
162 | { "w", "set logo width", OFFSET(w), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, | |
163 | { "h", "set logo height", OFFSET(h), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS }, | |
164 | { "band", "set delogo area band size", OFFSET(band), AV_OPT_TYPE_INT, { .i64 = 4 }, 1, INT_MAX, FLAGS }, | |
165 | { "t", "set delogo area band size", OFFSET(band), AV_OPT_TYPE_INT, { .i64 = 4 }, 1, INT_MAX, FLAGS }, | |
166 | { "show", "show delogo area", OFFSET(show), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, FLAGS }, | |
167 | { NULL } | |
168 | }; | |
169 | ||
170 | AVFILTER_DEFINE_CLASS(delogo); | |
171 | ||
172 | static int query_formats(AVFilterContext *ctx) | |
173 | { | |
174 | static const enum AVPixelFormat pix_fmts[] = { | |
175 | AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV420P, | |
176 | AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV440P, | |
177 | AV_PIX_FMT_YUVA420P, AV_PIX_FMT_GRAY8, | |
178 | AV_PIX_FMT_NONE | |
179 | }; | |
180 | ||
181 | ff_set_common_formats(ctx, ff_make_format_list(pix_fmts)); | |
182 | return 0; | |
183 | } | |
184 | ||
185 | static av_cold int init(AVFilterContext *ctx) | |
186 | { | |
187 | DelogoContext *s = ctx->priv; | |
188 | ||
189 | #define CHECK_UNSET_OPT(opt) \ | |
190 | if (s->opt == -1) { \ | |
191 | av_log(s, AV_LOG_ERROR, "Option %s was not set.\n", #opt); \ | |
192 | return AVERROR(EINVAL); \ | |
193 | } | |
194 | CHECK_UNSET_OPT(x); | |
195 | CHECK_UNSET_OPT(y); | |
196 | CHECK_UNSET_OPT(w); | |
197 | CHECK_UNSET_OPT(h); | |
198 | ||
199 | av_log(ctx, AV_LOG_VERBOSE, "x:%d y:%d, w:%d h:%d band:%d show:%d\n", | |
200 | s->x, s->y, s->w, s->h, s->band, s->show); | |
201 | ||
202 | s->w += s->band*2; | |
203 | s->h += s->band*2; | |
204 | s->x -= s->band; | |
205 | s->y -= s->band; | |
206 | ||
207 | return 0; | |
208 | } | |
209 | ||
210 | static int filter_frame(AVFilterLink *inlink, AVFrame *in) | |
211 | { | |
212 | DelogoContext *s = inlink->dst->priv; | |
213 | AVFilterLink *outlink = inlink->dst->outputs[0]; | |
214 | const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format); | |
215 | AVFrame *out; | |
216 | int hsub0 = desc->log2_chroma_w; | |
217 | int vsub0 = desc->log2_chroma_h; | |
218 | int direct = 0; | |
219 | int plane; | |
220 | AVRational sar; | |
221 | ||
222 | if (av_frame_is_writable(in)) { | |
223 | direct = 1; | |
224 | out = in; | |
225 | } else { | |
226 | out = ff_get_video_buffer(outlink, outlink->w, outlink->h); | |
227 | if (!out) { | |
228 | av_frame_free(&in); | |
229 | return AVERROR(ENOMEM); | |
230 | } | |
231 | ||
232 | av_frame_copy_props(out, in); | |
233 | } | |
234 | ||
235 | sar = in->sample_aspect_ratio; | |
236 | /* Assume square pixels if SAR is unknown */ | |
237 | if (!sar.num) | |
238 | sar.num = sar.den = 1; | |
239 | ||
240 | for (plane = 0; plane < 4 && in->data[plane] && in->linesize[plane]; plane++) { | |
241 | int hsub = plane == 1 || plane == 2 ? hsub0 : 0; | |
242 | int vsub = plane == 1 || plane == 2 ? vsub0 : 0; | |
243 | ||
244 | apply_delogo(out->data[plane], out->linesize[plane], | |
245 | in ->data[plane], in ->linesize[plane], | |
246 | FF_CEIL_RSHIFT(inlink->w, hsub), | |
247 | FF_CEIL_RSHIFT(inlink->h, vsub), | |
248 | sar, s->x>>hsub, s->y>>vsub, | |
249 | /* Up and left borders were rounded down, inject lost bits | |
250 | * into width and height to avoid error accumulation */ | |
251 | FF_CEIL_RSHIFT(s->w + (s->x & ((1<<hsub)-1)), hsub), | |
252 | FF_CEIL_RSHIFT(s->h + (s->y & ((1<<vsub)-1)), vsub), | |
253 | s->band>>FFMIN(hsub, vsub), | |
254 | s->show, direct); | |
255 | } | |
256 | ||
257 | if (!direct) | |
258 | av_frame_free(&in); | |
259 | ||
260 | return ff_filter_frame(outlink, out); | |
261 | } | |
262 | ||
263 | static const AVFilterPad avfilter_vf_delogo_inputs[] = { | |
264 | { | |
265 | .name = "default", | |
266 | .type = AVMEDIA_TYPE_VIDEO, | |
267 | .filter_frame = filter_frame, | |
268 | }, | |
269 | { NULL } | |
270 | }; | |
271 | ||
272 | static const AVFilterPad avfilter_vf_delogo_outputs[] = { | |
273 | { | |
274 | .name = "default", | |
275 | .type = AVMEDIA_TYPE_VIDEO, | |
276 | }, | |
277 | { NULL } | |
278 | }; | |
279 | ||
280 | AVFilter ff_vf_delogo = { | |
281 | .name = "delogo", | |
282 | .description = NULL_IF_CONFIG_SMALL("Remove logo from input video."), | |
283 | .priv_size = sizeof(DelogoContext), | |
284 | .priv_class = &delogo_class, | |
285 | .init = init, | |
286 | .query_formats = query_formats, | |
287 | .inputs = avfilter_vf_delogo_inputs, | |
288 | .outputs = avfilter_vf_delogo_outputs, | |
289 | .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC, | |
290 | }; |