1 /*
2 * This file is part of FFmpeg.
3 *
4 * FFmpeg is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
8 *
9 * FFmpeg is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
13 *
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with FFmpeg; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17 */
18
19 /**
20 * @file
21 * video scene change detection filter
22 */
23
24 #include "libavutil/avassert.h"
25 #include "libavutil/imgutils.h"
26 #include "libavutil/opt.h"
27 #include "libavutil/pixdesc.h"
28 #include "libavutil/timestamp.h"
29
30 #include "avfilter.h"
31 #include "filters.h"
32 #include "scene_sad.h"
33
34 typedef struct SCDetContext {
35 const AVClass *class;
36
37 ptrdiff_t width[4];
38 ptrdiff_t height[4];
39 int nb_planes;
40 int bitdepth;
41 ff_scene_sad_fn sad;
42 double prev_mafd;
43 double scene_score;
44 AVFrame *prev_picref;
45 double threshold;
46 int sc_pass;
47 } SCDetContext;
48
49 #define OFFSET(x) offsetof(SCDetContext, x)
50 #define V AV_OPT_FLAG_VIDEO_PARAM
51 #define F AV_OPT_FLAG_FILTERING_PARAM
52
53 static const AVOption scdet_options[] = {
54 { "threshold", "set scene change detect threshold", OFFSET(threshold), AV_OPT_TYPE_DOUBLE, {.dbl = 10.}, 0, 100., V|F },
55 { "t", "set scene change detect threshold", OFFSET(threshold), AV_OPT_TYPE_DOUBLE, {.dbl = 10.}, 0, 100., V|F },
56 { "sc_pass", "Set the flag to pass scene change frames", OFFSET(sc_pass), AV_OPT_TYPE_BOOL, {.dbl = 0 }, 0, 1, V|F },
57 { "s", "Set the flag to pass scene change frames", OFFSET(sc_pass), AV_OPT_TYPE_BOOL, {.dbl = 0 }, 0, 1, V|F },
58 {NULL}
59 };
60
61 AVFILTER_DEFINE_CLASS(scdet);
62
query_formats(AVFilterContext * ctx)63 static int query_formats(AVFilterContext *ctx)
64 {
65 static const enum AVPixelFormat pix_fmts[] = {
66 AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24, AV_PIX_FMT_RGBA,
67 AV_PIX_FMT_ABGR, AV_PIX_FMT_BGRA, AV_PIX_FMT_GRAY8,
68 AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUVJ420P,
69 AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUVJ422P,
70 AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUVJ440P,
71 AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUVJ444P,
72 AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV420P12,
73 AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV422P12,
74 AV_PIX_FMT_YUV444P9, AV_PIX_FMT_YUV444P10, AV_PIX_FMT_YUV444P12,
75 AV_PIX_FMT_NONE
76 };
77
78 AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts);
79 if (!fmts_list)
80 return AVERROR(ENOMEM);
81 return ff_set_common_formats(ctx, fmts_list);
82 }
83
config_input(AVFilterLink * inlink)84 static int config_input(AVFilterLink *inlink)
85 {
86 AVFilterContext *ctx = inlink->dst;
87 SCDetContext *s = ctx->priv;
88 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
89 int is_yuv = !(desc->flags & AV_PIX_FMT_FLAG_RGB) &&
90 (desc->flags & AV_PIX_FMT_FLAG_PLANAR) &&
91 desc->nb_components >= 3;
92
93 s->bitdepth = desc->comp[0].depth;
94 s->nb_planes = is_yuv ? 1 : av_pix_fmt_count_planes(inlink->format);
95
96 for (int plane = 0; plane < 4; plane++) {
97 ptrdiff_t line_size = av_image_get_linesize(inlink->format, inlink->w, plane);
98 s->width[plane] = line_size >> (s->bitdepth > 8);
99 s->height[plane] = inlink->h >> ((plane == 1 || plane == 2) ? desc->log2_chroma_h : 0);
100 }
101
102 s->sad = ff_scene_sad_get_fn(s->bitdepth == 8 ? 8 : 16);
103 if (!s->sad)
104 return AVERROR(EINVAL);
105
106 return 0;
107 }
108
uninit(AVFilterContext * ctx)109 static av_cold void uninit(AVFilterContext *ctx)
110 {
111 SCDetContext *s = ctx->priv;
112
113 av_frame_free(&s->prev_picref);
114 }
115
get_scene_score(AVFilterContext * ctx,AVFrame * frame)116 static double get_scene_score(AVFilterContext *ctx, AVFrame *frame)
117 {
118 double ret = 0;
119 SCDetContext *s = ctx->priv;
120 AVFrame *prev_picref = s->prev_picref;
121
122 if (prev_picref && frame->height == prev_picref->height
123 && frame->width == prev_picref->width) {
124 uint64_t sad = 0;
125 double mafd, diff;
126 uint64_t count = 0;
127
128 for (int plane = 0; plane < s->nb_planes; plane++) {
129 uint64_t plane_sad;
130 s->sad(prev_picref->data[plane], prev_picref->linesize[plane],
131 frame->data[plane], frame->linesize[plane],
132 s->width[plane], s->height[plane], &plane_sad);
133 sad += plane_sad;
134 count += s->width[plane] * s->height[plane];
135 }
136
137 emms_c();
138 mafd = (double)sad * 100. / count / (1ULL << s->bitdepth);
139 diff = fabs(mafd - s->prev_mafd);
140 ret = av_clipf(FFMIN(mafd, diff), 0, 100.);
141 s->prev_mafd = mafd;
142 av_frame_free(&prev_picref);
143 }
144 s->prev_picref = av_frame_clone(frame);
145 return ret;
146 }
147
set_meta(SCDetContext * s,AVFrame * frame,const char * key,const char * value)148 static int set_meta(SCDetContext *s, AVFrame *frame, const char *key, const char *value)
149 {
150 return av_dict_set(&frame->metadata, key, value, 0);
151 }
152
activate(AVFilterContext * ctx)153 static int activate(AVFilterContext *ctx)
154 {
155 int ret;
156 AVFilterLink *inlink = ctx->inputs[0];
157 AVFilterLink *outlink = ctx->outputs[0];
158 SCDetContext *s = ctx->priv;
159 AVFrame *frame;
160
161 FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
162
163 ret = ff_inlink_consume_frame(inlink, &frame);
164 if (ret < 0)
165 return ret;
166
167 if (frame) {
168 char buf[64];
169 s->scene_score = get_scene_score(ctx, frame);
170 snprintf(buf, sizeof(buf), "%0.3f", s->prev_mafd);
171 set_meta(s, frame, "lavfi.scd.mafd", buf);
172 snprintf(buf, sizeof(buf), "%0.3f", s->scene_score);
173 set_meta(s, frame, "lavfi.scd.score", buf);
174
175 if (s->scene_score > s->threshold) {
176 av_log(s, AV_LOG_INFO, "lavfi.scd.score: %.3f, lavfi.scd.time: %s\n",
177 s->scene_score, av_ts2timestr(frame->pts, &inlink->time_base));
178 set_meta(s, frame, "lavfi.scd.time",
179 av_ts2timestr(frame->pts, &inlink->time_base));
180 }
181 if (s->sc_pass) {
182 if (s->scene_score > s->threshold)
183 return ff_filter_frame(outlink, frame);
184 else {
185 av_frame_free(&frame);
186 }
187 } else
188 return ff_filter_frame(outlink, frame);
189 }
190
191 FF_FILTER_FORWARD_STATUS(inlink, outlink);
192 FF_FILTER_FORWARD_WANTED(outlink, inlink);
193
194 return FFERROR_NOT_READY;
195 }
196
197 static const AVFilterPad scdet_inputs[] = {
198 {
199 .name = "default",
200 .type = AVMEDIA_TYPE_VIDEO,
201 .config_props = config_input,
202 },
203 { NULL }
204 };
205
206 static const AVFilterPad scdet_outputs[] = {
207 {
208 .name = "default",
209 .type = AVMEDIA_TYPE_VIDEO,
210 },
211 { NULL }
212 };
213
214 AVFilter ff_vf_scdet = {
215 .name = "scdet",
216 .description = NULL_IF_CONFIG_SMALL("Detect video scene change"),
217 .priv_size = sizeof(SCDetContext),
218 .priv_class = &scdet_class,
219 .uninit = uninit,
220 .query_formats = query_formats,
221 .inputs = scdet_inputs,
222 .outputs = scdet_outputs,
223 .activate = activate,
224 };
225