2 * Copyright (c) 2014 Martin Storsjo
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24 #include "libavformat/avformat.h"
25 #include "libavutil/avstring.h"
26 #include "libavutil/intreadwrite.h"
27 #include "libavutil/mathematics.h"
29 static int usage(const char *argv0, int ret)
31 fprintf(stderr, "%s -out foo.mpd file1\n", argv0);
40 int is_audio, is_video;
42 int sample_rate, channels;
45 int64_t sidx_start, sidx_length;
46 int64_t earliest_presentation;
47 uint32_t earliest_presentation_timescale;
53 struct Track **tracks;
54 int multiple_tracks_per_file;
57 static void set_codec_str(AVCodecContext *codec, char *str, int size)
59 switch (codec->codec_id) {
60 case AV_CODEC_ID_H264:
61 snprintf(str, size, "avc1");
62 if (codec->extradata_size >= 4 && codec->extradata[0] == 1) {
63 av_strlcatf(str, size, ".%02x%02x%02x",
64 codec->extradata[1], codec->extradata[2], codec->extradata[3]);
68 snprintf(str, size, "mp4a.40"); // 0x40 is the mp4 object type for AAC
69 if (codec->extradata_size >= 2) {
70 int aot = codec->extradata[0] >> 3;
72 aot = ((AV_RB16(codec->extradata) >> 5) & 0x3f) + 32;
73 av_strlcatf(str, size, ".%d", aot);
79 static int find_sidx(struct Tracks *tracks, int start_index,
83 AVIOContext *f = NULL;
86 if ((err = avio_open2(&f, file, AVIO_FLAG_READ, NULL, NULL)) < 0)
89 while (!f->eof_reached) {
90 int64_t pos = avio_tell(f);
97 if (tag == MKBETAG('s', 'i', 'd', 'x')) {
98 int version, track_id;
100 int64_t earliest_presentation;
101 version = avio_r8(f);
102 avio_rb24(f); /* flags */
103 track_id = avio_rb32(f);
104 timescale = avio_rb32(f);
105 earliest_presentation = version ? avio_rb64(f) : avio_rb32(f);
106 for (i = start_index; i < tracks->nb_tracks; i++) {
107 struct Track *track = tracks->tracks[i];
108 if (!track->sidx_start) {
109 track->sidx_start = pos;
110 track->sidx_length = size;
111 } else if (pos == track->sidx_start + track->sidx_length) {
112 track->sidx_length = pos + size - track->sidx_start;
114 if (track->track_id == track_id) {
115 track->earliest_presentation = earliest_presentation;
116 track->earliest_presentation_timescale = timescale;
120 if (avio_seek(f, pos + size, SEEK_SET) != pos + size)
130 static int handle_file(struct Tracks *tracks, const char *file)
132 AVFormatContext *ctx = NULL;
133 int err = 0, i, orig_tracks = tracks->nb_tracks;
134 char errbuf[50], *ptr;
137 err = avformat_open_input(&ctx, file, NULL, NULL);
139 av_strerror(err, errbuf, sizeof(errbuf));
140 fprintf(stderr, "Unable to open %s: %s\n", file, errbuf);
144 err = avformat_find_stream_info(ctx, NULL);
146 av_strerror(err, errbuf, sizeof(errbuf));
147 fprintf(stderr, "Unable to identify %s: %s\n", file, errbuf);
151 if (ctx->nb_streams < 1) {
152 fprintf(stderr, "No streams found in %s\n", file);
155 if (ctx->nb_streams > 1)
156 tracks->multiple_tracks_per_file = 1;
158 for (i = 0; i < ctx->nb_streams; i++) {
160 AVStream *st = ctx->streams[i];
162 if (st->codec->bit_rate == 0) {
163 fprintf(stderr, "Skipping track %d in %s as it has zero bitrate\n",
168 track = av_mallocz(sizeof(*track));
170 err = AVERROR(ENOMEM);
173 temp = av_realloc(tracks->tracks,
174 sizeof(*tracks->tracks) * (tracks->nb_tracks + 1));
177 err = AVERROR(ENOMEM);
180 tracks->tracks = temp;
181 tracks->tracks[tracks->nb_tracks] = track;
184 if ((ptr = strrchr(file, '/')))
185 track->name = ptr + 1;
187 track->bitrate = st->codec->bit_rate;
188 track->track_id = st->id;
189 track->timescale = st->time_base.den;
190 track->duration = st->duration;
191 track->is_audio = st->codec->codec_type == AVMEDIA_TYPE_AUDIO;
192 track->is_video = st->codec->codec_type == AVMEDIA_TYPE_VIDEO;
194 if (!track->is_audio && !track->is_video) {
196 "Track %d in %s is neither video nor audio, skipping\n",
197 track->track_id, file);
198 av_freep(&tracks->tracks[tracks->nb_tracks]);
202 tracks->duration = FFMAX(tracks->duration,
203 av_rescale_rnd(track->duration, AV_TIME_BASE,
204 track->timescale, AV_ROUND_UP));
206 if (track->is_audio) {
207 track->channels = st->codec->channels;
208 track->sample_rate = st->codec->sample_rate;
210 if (track->is_video) {
211 track->width = st->codec->width;
212 track->height = st->codec->height;
214 set_codec_str(st->codec, track->codec_str, sizeof(track->codec_str));
219 avformat_close_input(&ctx);
221 err = find_sidx(tracks, orig_tracks, file);
225 avformat_close_input(&ctx);
229 static void write_time(FILE *out, int64_t time, int decimals, enum AVRounding round)
231 int seconds = time / AV_TIME_BASE;
232 int fractions = time % AV_TIME_BASE;
233 int minutes = seconds / 60;
234 int hours = minutes / 60;
235 fractions = av_rescale_rnd(fractions, pow(10, decimals), AV_TIME_BASE, round);
240 fprintf(out, "%dH", hours);
241 if (hours || minutes)
242 fprintf(out, "%dM", minutes);
243 fprintf(out, "%d.%0*dS", seconds, decimals, fractions);
246 static int output_mpd(struct Tracks *tracks, const char *filename)
250 struct Track **adaptation_sets_buf[2] = { NULL };
251 struct Track ***adaptation_sets;
252 int nb_tracks_buf[2] = { 0 };
256 if (!tracks->multiple_tracks_per_file) {
257 adaptation_sets = adaptation_sets_buf;
258 nb_tracks = nb_tracks_buf;
260 for (i = 0; i < 2; i++) {
261 adaptation_sets[i] = av_malloc(sizeof(*adaptation_sets[i]) * tracks->nb_tracks);
262 if (!adaptation_sets[i]) {
263 ret = AVERROR(ENOMEM);
267 for (i = 0; i < tracks->nb_tracks; i++) {
269 if (tracks->tracks[i]->is_video)
271 else if (tracks->tracks[i]->is_audio)
275 adaptation_sets[set_index][nb_tracks[set_index]++] = tracks->tracks[i];
278 adaptation_sets = &tracks->tracks;
279 nb_tracks = &tracks->nb_tracks;
283 out = fopen(filename, "w");
285 ret = AVERROR(errno);
289 fprintf(out, "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n");
290 fprintf(out, "<MPD xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"\n"
291 "\txmlns=\"urn:mpeg:dash:schema:mpd:2011\"\n"
292 "\txmlns:xlink=\"http://www.w3.org/1999/xlink\"\n"
293 "\txsi:schemaLocation=\"urn:mpeg:DASH:schema:MPD:2011 http://standards.iso.org/ittf/PubliclyAvailableStandards/MPEG-DASH_schema_files/DASH-MPD.xsd\"\n"
294 "\tprofiles=\"urn:mpeg:dash:profile:isoff-on-demand:2011\"\n"
295 "\ttype=\"static\"\n");
296 fprintf(out, "\tmediaPresentationDuration=\"");
297 write_time(out, tracks->duration, 1, AV_ROUND_DOWN);
298 fprintf(out, "\"\n");
299 fprintf(out, "\tminBufferTime=\"PT5S\">\n");
301 fprintf(out, "\t<Period start=\"PT0.0S\">\n");
303 for (set = 0; set < nb_sets; set++) {
304 if (nb_tracks[set] == 0)
306 fprintf(out, "\t\t<AdaptationSet segmentAlignment=\"true\">\n");
308 for (i = 0; i < nb_tracks[set]; i++) {
309 struct Track *track = adaptation_sets[set][i];
310 if (strcmp(track->name, adaptation_sets[set][0]->name))
312 fprintf(out, "\t\t\t<ContentComponent id=\"%d\" contentType=\"%s\" />\n", track->track_id, track->is_audio ? "audio" : "video");
316 for (i = 0; i < nb_tracks[set]; ) {
317 struct Track *first_track = adaptation_sets[set][i];
318 int width = 0, height = 0, sample_rate = 0, channels = 0, bitrate = 0;
319 fprintf(out, "\t\t\t<Representation id=\"%d\" codecs=\"", i);
320 for (j = i; j < nb_tracks[set]; j++) {
321 struct Track *track = adaptation_sets[set][j];
322 if (strcmp(track->name, first_track->name))
324 if (track->is_audio) {
325 sample_rate = track->sample_rate;
326 channels = track->channels;
328 if (track->is_video) {
329 width = track->width;
330 height = track->height;
332 bitrate += track->bitrate;
335 fprintf(out, "%s", track->codec_str);
337 fprintf(out, "\" mimeType=\"%s/mp4\" bandwidth=\"%d\"",
338 width ? "video" : "audio", bitrate);
339 if (width > 0 && height > 0)
340 fprintf(out, " width=\"%d\" height=\"%d\"", width, height);
342 fprintf(out, " audioSamplingRate=\"%d\"", sample_rate);
345 fprintf(out, "\t\t\t\t<AudioChannelConfiguration schemeIdUri=\"urn:mpeg:dash:23003:3:audio_channel_configuration:2011\" value=\"%d\" />\n", channels);
346 fprintf(out, "\t\t\t\t<BaseURL>%s</BaseURL>\n", first_track->name);
347 fprintf(out, "\t\t\t\t<SegmentBase indexRange=\"%"PRId64"-%"PRId64"\" />\n", first_track->sidx_start, first_track->sidx_start + first_track->sidx_length - 1);
348 fprintf(out, "\t\t\t</Representation>\n");
351 fprintf(out, "\t\t</AdaptationSet>\n");
353 fprintf(out, "\t</Period>\n");
354 fprintf(out, "</MPD>\n");
358 for (i = 0; i < 2; i++)
359 av_free(adaptation_sets_buf[i]);
363 static void clean_tracks(struct Tracks *tracks)
366 for (i = 0; i < tracks->nb_tracks; i++) {
367 av_freep(&tracks->tracks[i]);
369 av_freep(&tracks->tracks);
370 tracks->nb_tracks = 0;
373 int main(int argc, char **argv)
375 const char *out = NULL;
376 struct Tracks tracks = { 0 };
381 for (i = 1; i < argc; i++) {
382 if (!strcmp(argv[i], "-out")) {
385 } else if (argv[i][0] == '-') {
386 return usage(argv[0], 1);
388 if (handle_file(&tracks, argv[i]))
392 if (!tracks.nb_tracks || !out)
393 return usage(argv[0], 1);
395 output_mpd(&tracks, out);
397 clean_tracks(&tracks);