1 /*
2 * Copyright (c) 2010 Stefano Sabatini
3 * Copyright (c) 2008 Victor Paesa
4 *
5 * This file is part of FFmpeg.
6 *
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22 /**
23 * @file
24 * video presentation timestamp (PTS) modification filter
25 */
26
27 #include <inttypes.h>
28
29 #include "libavutil/eval.h"
30 #include "libavutil/internal.h"
31 #include "libavutil/mathematics.h"
32 #include "libavutil/opt.h"
33 #include "libavutil/time.h"
34 #include "audio.h"
35 #include "avfilter.h"
36 #include "filters.h"
37 #include "internal.h"
38 #include "video.h"
39
40 static const char *const var_names[] = {
41 "FRAME_RATE", ///< defined only for constant frame-rate video
42 "INTERLACED", ///< tell if the current frame is interlaced
43 "N", ///< frame / sample number (starting at zero)
44 "NB_CONSUMED_SAMPLES", ///< number of samples consumed by the filter (only audio)
45 "NB_SAMPLES", ///< number of samples in the current frame (only audio)
46 "POS", ///< original position in the file of the frame
47 "PREV_INPTS", ///< previous input PTS
48 "PREV_INT", ///< previous input time in seconds
49 "PREV_OUTPTS", ///< previous output PTS
50 "PREV_OUTT", ///< previous output time in seconds
51 "PTS", ///< original pts in the file of the frame
52 "SAMPLE_RATE", ///< sample rate (only audio)
53 "STARTPTS", ///< PTS at start of movie
54 "STARTT", ///< time at start of movie
55 "T", ///< original time in the file of the frame
56 "TB", ///< timebase
57 "RTCTIME", ///< wallclock (RTC) time in micro seconds
58 "RTCSTART", ///< wallclock (RTC) time at the start of the movie in micro seconds
59 "S", // Number of samples in the current frame
60 "SR", // Audio sample rate
61 "FR", ///< defined only for constant frame-rate video
62 NULL
63 };
64
65 enum var_name {
66 VAR_FRAME_RATE,
67 VAR_INTERLACED,
68 VAR_N,
69 VAR_NB_CONSUMED_SAMPLES,
70 VAR_NB_SAMPLES,
71 VAR_POS,
72 VAR_PREV_INPTS,
73 VAR_PREV_INT,
74 VAR_PREV_OUTPTS,
75 VAR_PREV_OUTT,
76 VAR_PTS,
77 VAR_SAMPLE_RATE,
78 VAR_STARTPTS,
79 VAR_STARTT,
80 VAR_T,
81 VAR_TB,
82 VAR_RTCTIME,
83 VAR_RTCSTART,
84 VAR_S,
85 VAR_SR,
86 VAR_FR,
87 VAR_VARS_NB
88 };
89
90 typedef struct SetPTSContext {
91 const AVClass *class;
92 char *expr_str;
93 AVExpr *expr;
94 double var_values[VAR_VARS_NB];
95 enum AVMediaType type;
96 } SetPTSContext;
97
init(AVFilterContext * ctx)98 static av_cold int init(AVFilterContext *ctx)
99 {
100 SetPTSContext *setpts = ctx->priv;
101 int ret;
102
103 if ((ret = av_expr_parse(&setpts->expr, setpts->expr_str,
104 var_names, NULL, NULL, NULL, NULL, 0, ctx)) < 0) {
105 av_log(ctx, AV_LOG_ERROR, "Error while parsing expression '%s'\n", setpts->expr_str);
106 return ret;
107 }
108
109 setpts->var_values[VAR_N] = 0.0;
110 setpts->var_values[VAR_S] = 0.0;
111 setpts->var_values[VAR_PREV_INPTS] = NAN;
112 setpts->var_values[VAR_PREV_INT] = NAN;
113 setpts->var_values[VAR_PREV_OUTPTS] = NAN;
114 setpts->var_values[VAR_PREV_OUTT] = NAN;
115 setpts->var_values[VAR_STARTPTS] = NAN;
116 setpts->var_values[VAR_STARTT] = NAN;
117 return 0;
118 }
119
config_input(AVFilterLink * inlink)120 static int config_input(AVFilterLink *inlink)
121 {
122 AVFilterContext *ctx = inlink->dst;
123 SetPTSContext *setpts = ctx->priv;
124
125 setpts->type = inlink->type;
126 setpts->var_values[VAR_TB] = av_q2d(inlink->time_base);
127 setpts->var_values[VAR_RTCSTART] = av_gettime();
128
129 setpts->var_values[VAR_SR] =
130 setpts->var_values[VAR_SAMPLE_RATE] =
131 setpts->type == AVMEDIA_TYPE_AUDIO ? inlink->sample_rate : NAN;
132
133 setpts->var_values[VAR_FRAME_RATE] =
134 setpts->var_values[VAR_FR] = inlink->frame_rate.num &&
135 inlink->frame_rate.den ?
136 av_q2d(inlink->frame_rate) : NAN;
137
138 av_log(inlink->src, AV_LOG_VERBOSE, "TB:%f FRAME_RATE:%f SAMPLE_RATE:%f\n",
139 setpts->var_values[VAR_TB],
140 setpts->var_values[VAR_FRAME_RATE],
141 setpts->var_values[VAR_SAMPLE_RATE]);
142 return 0;
143 }
144
145 #define BUF_SIZE 64
146
double2int64str(char * buf,double v)147 static inline char *double2int64str(char *buf, double v)
148 {
149 if (isnan(v)) snprintf(buf, BUF_SIZE, "nan");
150 else snprintf(buf, BUF_SIZE, "%"PRId64, (int64_t)v);
151 return buf;
152 }
153
eval_pts(SetPTSContext * setpts,AVFilterLink * inlink,AVFrame * frame,int64_t pts)154 static double eval_pts(SetPTSContext *setpts, AVFilterLink *inlink, AVFrame *frame, int64_t pts)
155 {
156 if (isnan(setpts->var_values[VAR_STARTPTS])) {
157 setpts->var_values[VAR_STARTPTS] = TS2D(pts);
158 setpts->var_values[VAR_STARTT ] = TS2T(pts, inlink->time_base);
159 }
160 setpts->var_values[VAR_PTS ] = TS2D(pts);
161 setpts->var_values[VAR_T ] = TS2T(pts, inlink->time_base);
162 setpts->var_values[VAR_POS ] = !frame || frame->pkt_pos == -1 ? NAN : frame->pkt_pos;
163 setpts->var_values[VAR_RTCTIME ] = av_gettime();
164
165 if (frame) {
166 if (inlink->type == AVMEDIA_TYPE_VIDEO) {
167 setpts->var_values[VAR_INTERLACED] = frame->interlaced_frame;
168 } else if (inlink->type == AVMEDIA_TYPE_AUDIO) {
169 setpts->var_values[VAR_S] = frame->nb_samples;
170 setpts->var_values[VAR_NB_SAMPLES] = frame->nb_samples;
171 }
172 }
173
174 return av_expr_eval(setpts->expr, setpts->var_values, NULL);
175 }
176 #define d2istr(v) double2int64str((char[BUF_SIZE]){0}, v)
177
filter_frame(AVFilterLink * inlink,AVFrame * frame)178 static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
179 {
180 SetPTSContext *setpts = inlink->dst->priv;
181 int64_t in_pts = frame->pts;
182 double d;
183
184 d = eval_pts(setpts, inlink, frame, frame->pts);
185 frame->pts = D2TS(d);
186
187 av_log(inlink->dst, AV_LOG_TRACE,
188 "N:%"PRId64" PTS:%s T:%f POS:%s",
189 (int64_t)setpts->var_values[VAR_N],
190 d2istr(setpts->var_values[VAR_PTS]),
191 setpts->var_values[VAR_T],
192 d2istr(setpts->var_values[VAR_POS]));
193 switch (inlink->type) {
194 case AVMEDIA_TYPE_VIDEO:
195 av_log(inlink->dst, AV_LOG_TRACE, " INTERLACED:%"PRId64,
196 (int64_t)setpts->var_values[VAR_INTERLACED]);
197 break;
198 case AVMEDIA_TYPE_AUDIO:
199 av_log(inlink->dst, AV_LOG_TRACE, " NB_SAMPLES:%"PRId64" NB_CONSUMED_SAMPLES:%"PRId64,
200 (int64_t)setpts->var_values[VAR_NB_SAMPLES],
201 (int64_t)setpts->var_values[VAR_NB_CONSUMED_SAMPLES]);
202 break;
203 }
204 av_log(inlink->dst, AV_LOG_TRACE, " -> PTS:%s T:%f\n", d2istr(d), TS2T(d, inlink->time_base));
205
206 if (inlink->type == AVMEDIA_TYPE_VIDEO) {
207 setpts->var_values[VAR_N] += 1.0;
208 } else {
209 setpts->var_values[VAR_N] += frame->nb_samples;
210 }
211
212 setpts->var_values[VAR_PREV_INPTS ] = TS2D(in_pts);
213 setpts->var_values[VAR_PREV_INT ] = TS2T(in_pts, inlink->time_base);
214 setpts->var_values[VAR_PREV_OUTPTS] = TS2D(frame->pts);
215 setpts->var_values[VAR_PREV_OUTT] = TS2T(frame->pts, inlink->time_base);
216 if (setpts->type == AVMEDIA_TYPE_AUDIO) {
217 setpts->var_values[VAR_NB_CONSUMED_SAMPLES] += frame->nb_samples;
218 }
219 return ff_filter_frame(inlink->dst->outputs[0], frame);
220 }
221
activate(AVFilterContext * ctx)222 static int activate(AVFilterContext *ctx)
223 {
224 SetPTSContext *setpts = ctx->priv;
225 AVFilterLink *inlink = ctx->inputs[0];
226 AVFilterLink *outlink = ctx->outputs[0];
227 AVFrame *in;
228 int status;
229 int64_t pts;
230 int ret;
231
232 FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
233
234 ret = ff_inlink_consume_frame(inlink, &in);
235 if (ret < 0)
236 return ret;
237 if (ret > 0)
238 return filter_frame(inlink, in);
239
240 if (ff_inlink_acknowledge_status(inlink, &status, &pts)) {
241 double d = eval_pts(setpts, inlink, NULL, pts);
242
243 av_log(ctx, AV_LOG_TRACE, "N:EOF PTS:%s T:%f POS:%s -> PTS:%s T:%f\n",
244 d2istr(setpts->var_values[VAR_PTS]),
245 setpts->var_values[VAR_T],
246 d2istr(setpts->var_values[VAR_POS]),
247 d2istr(d), TS2T(d, inlink->time_base));
248 ff_outlink_set_status(outlink, status, D2TS(d));
249 return 0;
250 }
251
252 FF_FILTER_FORWARD_WANTED(outlink, inlink);
253
254 return FFERROR_NOT_READY;
255 }
256
uninit(AVFilterContext * ctx)257 static av_cold void uninit(AVFilterContext *ctx)
258 {
259 SetPTSContext *setpts = ctx->priv;
260 av_expr_free(setpts->expr);
261 setpts->expr = NULL;
262 }
263
264 #define OFFSET(x) offsetof(SetPTSContext, x)
265 #define V AV_OPT_FLAG_VIDEO_PARAM
266 #define A AV_OPT_FLAG_AUDIO_PARAM
267 #define F AV_OPT_FLAG_FILTERING_PARAM
268
269 #if CONFIG_SETPTS_FILTER
270 static const AVOption setpts_options[] = {
271 { "expr", "Expression determining the frame timestamp", OFFSET(expr_str), AV_OPT_TYPE_STRING, { .str = "PTS" }, .flags = V|F },
272 { NULL }
273 };
274 AVFILTER_DEFINE_CLASS(setpts);
275
276 static const AVFilterPad avfilter_vf_setpts_inputs[] = {
277 {
278 .name = "default",
279 .type = AVMEDIA_TYPE_VIDEO,
280 .config_props = config_input,
281 },
282 { NULL }
283 };
284
285 static const AVFilterPad avfilter_vf_setpts_outputs[] = {
286 {
287 .name = "default",
288 .type = AVMEDIA_TYPE_VIDEO,
289 },
290 { NULL }
291 };
292
293 AVFilter ff_vf_setpts = {
294 .name = "setpts",
295 .description = NULL_IF_CONFIG_SMALL("Set PTS for the output video frame."),
296 .init = init,
297 .activate = activate,
298 .uninit = uninit,
299
300 .priv_size = sizeof(SetPTSContext),
301 .priv_class = &setpts_class,
302
303 .inputs = avfilter_vf_setpts_inputs,
304 .outputs = avfilter_vf_setpts_outputs,
305 };
306 #endif /* CONFIG_SETPTS_FILTER */
307
308 #if CONFIG_ASETPTS_FILTER
309
310 static const AVOption asetpts_options[] = {
311 { "expr", "Expression determining the frame timestamp", OFFSET(expr_str), AV_OPT_TYPE_STRING, { .str = "PTS" }, .flags = A|F },
312 { NULL }
313 };
314 AVFILTER_DEFINE_CLASS(asetpts);
315
316 static const AVFilterPad asetpts_inputs[] = {
317 {
318 .name = "default",
319 .type = AVMEDIA_TYPE_AUDIO,
320 .config_props = config_input,
321 },
322 { NULL }
323 };
324
325 static const AVFilterPad asetpts_outputs[] = {
326 {
327 .name = "default",
328 .type = AVMEDIA_TYPE_AUDIO,
329 },
330 { NULL }
331 };
332
333 AVFilter ff_af_asetpts = {
334 .name = "asetpts",
335 .description = NULL_IF_CONFIG_SMALL("Set PTS for the output audio frame."),
336 .init = init,
337 .activate = activate,
338 .uninit = uninit,
339 .priv_size = sizeof(SetPTSContext),
340 .priv_class = &asetpts_class,
341 .inputs = asetpts_inputs,
342 .outputs = asetpts_outputs,
343 };
344 #endif /* CONFIG_ASETPTS_FILTER */
345