1 /*
2 * Copyright (c) 2002 A'rpi
3 * This file is part of FFmpeg.
4 *
5 * FFmpeg is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * FFmpeg is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with FFmpeg; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
18 */
19
20 /**
21 * @file
22 * border detection filter
23 * Ported from MPlayer libmpcodecs/vf_cropdetect.c.
24 */
25
26 #include "libavutil/imgutils.h"
27 #include "libavutil/internal.h"
28 #include "libavutil/opt.h"
29
30 #include "avfilter.h"
31 #include "formats.h"
32 #include "internal.h"
33 #include "video.h"
34
35 typedef struct CropDetectContext {
36 const AVClass *class;
37 int x1, y1, x2, y2;
38 float limit;
39 int round;
40 int skip;
41 int reset_count;
42 int frame_nb;
43 int max_pixsteps[4];
44 int max_outliers;
45 } CropDetectContext;
46
47 static const enum AVPixelFormat pix_fmts[] = {
48 AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUVJ420P,
49 AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUVJ422P,
50 AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUVJ444P,
51 AV_PIX_FMT_YUV411P, AV_PIX_FMT_GRAY8,
52 AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUV410P,
53 AV_PIX_FMT_YUV420P9 , AV_PIX_FMT_YUV422P9 , AV_PIX_FMT_YUV444P9,
54 AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
55 AV_PIX_FMT_YUV420P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV444P12,
56 AV_PIX_FMT_YUV420P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV444P14,
57 AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
58 AV_PIX_FMT_NV12, AV_PIX_FMT_NV21,
59 AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24,
60 AV_PIX_FMT_RGBA, AV_PIX_FMT_BGRA,
61 AV_PIX_FMT_NONE
62 };
63
checkline(void * ctx,const unsigned char * src,int stride,int len,int bpp)64 static int checkline(void *ctx, const unsigned char *src, int stride, int len, int bpp)
65 {
66 int total = 0;
67 int div = len;
68 const uint16_t *src16 = (const uint16_t *)src;
69
70 switch (bpp) {
71 case 1:
72 while (len >= 8) {
73 total += src[ 0] + src[ stride] + src[2*stride] + src[3*stride]
74 + src[4*stride] + src[5*stride] + src[6*stride] + src[7*stride];
75 src += 8*stride;
76 len -= 8;
77 }
78 while (--len >= 0) {
79 total += src[0];
80 src += stride;
81 }
82 break;
83 case 2:
84 stride >>= 1;
85 while (len >= 8) {
86 total += src16[ 0] + src16[ stride] + src16[2*stride] + src16[3*stride]
87 + src16[4*stride] + src16[5*stride] + src16[6*stride] + src16[7*stride];
88 src16 += 8*stride;
89 len -= 8;
90 }
91 while (--len >= 0) {
92 total += src16[0];
93 src16 += stride;
94 }
95 break;
96 case 3:
97 case 4:
98 while (len >= 4) {
99 total += src[0] + src[1 ] + src[2 ]
100 + src[ stride] + src[1+ stride] + src[2+ stride]
101 + src[2*stride] + src[1+2*stride] + src[2+2*stride]
102 + src[3*stride] + src[1+3*stride] + src[2+3*stride];
103 src += 4*stride;
104 len -= 4;
105 }
106 while (--len >= 0) {
107 total += src[0] + src[1] + src[2];
108 src += stride;
109 }
110 div *= 3;
111 break;
112 }
113 total /= div;
114
115 av_log(ctx, AV_LOG_DEBUG, "total:%d\n", total);
116 return total;
117 }
118
init(AVFilterContext * ctx)119 static av_cold int init(AVFilterContext *ctx)
120 {
121 CropDetectContext *s = ctx->priv;
122
123 s->frame_nb = -1 * s->skip;
124
125 av_log(ctx, AV_LOG_VERBOSE, "limit:%f round:%d skip:%d reset_count:%d\n",
126 s->limit, s->round, s->skip, s->reset_count);
127
128 return 0;
129 }
130
config_input(AVFilterLink * inlink)131 static int config_input(AVFilterLink *inlink)
132 {
133 AVFilterContext *ctx = inlink->dst;
134 CropDetectContext *s = ctx->priv;
135 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
136
137 av_image_fill_max_pixsteps(s->max_pixsteps, NULL, desc);
138
139 if (s->limit < 1.0)
140 s->limit *= (1 << desc->comp[0].depth) - 1;
141
142 s->x1 = inlink->w - 1;
143 s->y1 = inlink->h - 1;
144 s->x2 = 0;
145 s->y2 = 0;
146
147 return 0;
148 }
149
150 #define SET_META(key, value) \
151 av_dict_set_int(metadata, key, value, 0)
152
filter_frame(AVFilterLink * inlink,AVFrame * frame)153 static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
154 {
155 AVFilterContext *ctx = inlink->dst;
156 CropDetectContext *s = ctx->priv;
157 int bpp = s->max_pixsteps[0];
158 int w, h, x, y, shrink_by;
159 AVDictionary **metadata;
160 int outliers, last_y;
161 int limit = lrint(s->limit);
162
163 // ignore first s->skip frames
164 if (++s->frame_nb > 0) {
165 metadata = &frame->metadata;
166
167 // Reset the crop area every reset_count frames, if reset_count is > 0
168 if (s->reset_count > 0 && s->frame_nb > s->reset_count) {
169 s->x1 = frame->width - 1;
170 s->y1 = frame->height - 1;
171 s->x2 = 0;
172 s->y2 = 0;
173 s->frame_nb = 1;
174 }
175
176 #define FIND(DST, FROM, NOEND, INC, STEP0, STEP1, LEN) \
177 outliers = 0;\
178 for (last_y = y = FROM; NOEND; y = y INC) {\
179 if (checkline(ctx, frame->data[0] + STEP0 * y, STEP1, LEN, bpp) > limit) {\
180 if (++outliers > s->max_outliers) { \
181 DST = last_y;\
182 break;\
183 }\
184 } else\
185 last_y = y INC;\
186 }
187
188 FIND(s->y1, 0, y < s->y1, +1, frame->linesize[0], bpp, frame->width);
189 FIND(s->y2, frame->height - 1, y > FFMAX(s->y2, s->y1), -1, frame->linesize[0], bpp, frame->width);
190 FIND(s->x1, 0, y < s->x1, +1, bpp, frame->linesize[0], frame->height);
191 FIND(s->x2, frame->width - 1, y > FFMAX(s->x2, s->x1), -1, bpp, frame->linesize[0], frame->height);
192
193
194 // round x and y (up), important for yuv colorspaces
195 // make sure they stay rounded!
196 x = (s->x1+1) & ~1;
197 y = (s->y1+1) & ~1;
198
199 w = s->x2 - x + 1;
200 h = s->y2 - y + 1;
201
202 // w and h must be divisible by 2 as well because of yuv
203 // colorspace problems.
204 if (s->round <= 1)
205 s->round = 16;
206 if (s->round % 2)
207 s->round *= 2;
208
209 shrink_by = w % s->round;
210 w -= shrink_by;
211 x += (shrink_by/2 + 1) & ~1;
212
213 shrink_by = h % s->round;
214 h -= shrink_by;
215 y += (shrink_by/2 + 1) & ~1;
216
217 SET_META("lavfi.cropdetect.x1", s->x1);
218 SET_META("lavfi.cropdetect.x2", s->x2);
219 SET_META("lavfi.cropdetect.y1", s->y1);
220 SET_META("lavfi.cropdetect.y2", s->y2);
221 SET_META("lavfi.cropdetect.w", w);
222 SET_META("lavfi.cropdetect.h", h);
223 SET_META("lavfi.cropdetect.x", x);
224 SET_META("lavfi.cropdetect.y", y);
225
226 av_log(ctx, AV_LOG_INFO,
227 "x1:%d x2:%d y1:%d y2:%d w:%d h:%d x:%d y:%d pts:%"PRId64" t:%f crop=%d:%d:%d:%d\n",
228 s->x1, s->x2, s->y1, s->y2, w, h, x, y, frame->pts,
229 frame->pts == AV_NOPTS_VALUE ? -1 : frame->pts * av_q2d(inlink->time_base),
230 w, h, x, y);
231 }
232
233 return ff_filter_frame(inlink->dst->outputs[0], frame);
234 }
235
236 #define OFFSET(x) offsetof(CropDetectContext, x)
237 #define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
238
239 static const AVOption cropdetect_options[] = {
240 { "limit", "Threshold below which the pixel is considered black", OFFSET(limit), AV_OPT_TYPE_FLOAT, { .dbl = 24.0/255 }, 0, 65535, FLAGS },
241 { "round", "Value by which the width/height should be divisible", OFFSET(round), AV_OPT_TYPE_INT, { .i64 = 16 }, 0, INT_MAX, FLAGS },
242 { "reset", "Recalculate the crop area after this many frames", OFFSET(reset_count), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS },
243 { "skip", "Number of initial frames to skip", OFFSET(skip), AV_OPT_TYPE_INT, { .i64 = 2 }, 0, INT_MAX, FLAGS },
244 { "reset_count", "Recalculate the crop area after this many frames",OFFSET(reset_count),AV_OPT_TYPE_INT,{ .i64 = 0 }, 0, INT_MAX, FLAGS },
245 { "max_outliers", "Threshold count of outliers", OFFSET(max_outliers),AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS },
246 { NULL }
247 };
248
249 AVFILTER_DEFINE_CLASS(cropdetect);
250
251 static const AVFilterPad avfilter_vf_cropdetect_inputs[] = {
252 {
253 .name = "default",
254 .type = AVMEDIA_TYPE_VIDEO,
255 .config_props = config_input,
256 .filter_frame = filter_frame,
257 },
258 };
259
260 static const AVFilterPad avfilter_vf_cropdetect_outputs[] = {
261 {
262 .name = "default",
263 .type = AVMEDIA_TYPE_VIDEO
264 },
265 };
266
267 const AVFilter ff_vf_cropdetect = {
268 .name = "cropdetect",
269 .description = NULL_IF_CONFIG_SMALL("Auto-detect crop size."),
270 .priv_size = sizeof(CropDetectContext),
271 .priv_class = &cropdetect_class,
272 .init = init,
273 FILTER_INPUTS(avfilter_vf_cropdetect_inputs),
274 FILTER_OUTPUTS(avfilter_vf_cropdetect_outputs),
275 FILTER_PIXFMTS_ARRAY(pix_fmts),
276 .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC | AVFILTER_FLAG_METADATA_ONLY,
277 };
278