2 * Copyright (c) 2010 Nicolas George
3 * Copyright (c) 2011 Stefano Sabatini
4 * Copyright (c) 2014 Andrey Utkin
6 * Permission is hereby granted, free of charge, to any person obtaining a copy
7 * of this software and associated documentation files (the "Software"), to deal
8 * in the Software without restriction, including without limitation the rights
9 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
10 * copies of the Software, and to permit persons to whom the Software is
11 * furnished to do so, subject to the following conditions:
13 * The above copyright notice and this permission notice shall be included in
14 * all copies or substantial portions of the Software.
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
19 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
27 * API example for demuxing, decoding, filtering, encoding and muxing
28 * @example doc/examples/transcoding.c
31 #include <libavcodec/avcodec.h>
32 #include <libavformat/avformat.h>
33 #include <libavfilter/avfiltergraph.h>
34 #include <libavfilter/avcodec.h>
35 #include <libavfilter/buffersink.h>
36 #include <libavfilter/buffersrc.h>
37 #include <libavutil/opt.h>
38 #include <libavutil/pixdesc.h>
40 static AVFormatContext *ifmt_ctx;
41 static AVFormatContext *ofmt_ctx;
42 typedef struct FilteringContext {
43 AVFilterContext *buffersink_ctx;
44 AVFilterContext *buffersrc_ctx;
45 AVFilterGraph *filter_graph;
47 static FilteringContext *filter_ctx;
49 static int open_input_file(const char *filename)
55 if ((ret = avformat_open_input(&ifmt_ctx, filename, NULL, NULL)) < 0) {
56 av_log(NULL, AV_LOG_ERROR, "Cannot open input file\n");
60 if ((ret = avformat_find_stream_info(ifmt_ctx, NULL)) < 0) {
61 av_log(NULL, AV_LOG_ERROR, "Cannot find stream information\n");
65 for (i = 0; i < ifmt_ctx->nb_streams; i++) {
67 AVCodecContext *codec_ctx;
68 stream = ifmt_ctx->streams[i];
69 codec_ctx = stream->codec;
70 /* Reencode video & audio and remux subtitles etc. */
71 if (codec_ctx->codec_type == AVMEDIA_TYPE_VIDEO
72 || codec_ctx->codec_type == AVMEDIA_TYPE_AUDIO) {
74 ret = avcodec_open2(codec_ctx,
75 avcodec_find_decoder(codec_ctx->codec_id), NULL);
77 av_log(NULL, AV_LOG_ERROR, "Failed to open decoder for stream #%u\n", i);
83 av_dump_format(ifmt_ctx, 0, filename, 0);
87 static int open_output_file(const char *filename)
91 AVCodecContext *dec_ctx, *enc_ctx;
97 avformat_alloc_output_context2(&ofmt_ctx, NULL, NULL, filename);
99 av_log(NULL, AV_LOG_ERROR, "Could not create output context\n");
100 return AVERROR_UNKNOWN;
104 for (i = 0; i < ifmt_ctx->nb_streams; i++) {
105 out_stream = avformat_new_stream(ofmt_ctx, NULL);
107 av_log(NULL, AV_LOG_ERROR, "Failed allocating output stream\n");
108 return AVERROR_UNKNOWN;
111 in_stream = ifmt_ctx->streams[i];
112 dec_ctx = in_stream->codec;
113 enc_ctx = out_stream->codec;
115 if (dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO
116 || dec_ctx->codec_type == AVMEDIA_TYPE_AUDIO) {
117 /* in this example, we choose transcoding to same codec */
118 encoder = avcodec_find_encoder(dec_ctx->codec_id);
120 /* In this example, we transcode to same properties (picture size,
121 * sample rate etc.). These properties can be changed for output
122 * streams easily using filters */
123 if (dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO) {
124 enc_ctx->height = dec_ctx->height;
125 enc_ctx->width = dec_ctx->width;
126 enc_ctx->sample_aspect_ratio = dec_ctx->sample_aspect_ratio;
127 /* take first format from list of supported formats */
128 enc_ctx->pix_fmt = encoder->pix_fmts[0];
129 /* video time_base can be set to whatever is handy and supported by encoder */
130 enc_ctx->time_base = dec_ctx->time_base;
132 enc_ctx->sample_rate = dec_ctx->sample_rate;
133 enc_ctx->channel_layout = dec_ctx->channel_layout;
134 enc_ctx->channels = av_get_channel_layout_nb_channels(enc_ctx->channel_layout);
135 /* take first format from list of supported formats */
136 enc_ctx->sample_fmt = encoder->sample_fmts[0];
137 enc_ctx->time_base = (AVRational){1, enc_ctx->sample_rate};
140 /* Third parameter can be used to pass settings to encoder */
141 ret = avcodec_open2(enc_ctx, encoder, NULL);
143 av_log(NULL, AV_LOG_ERROR, "Cannot open video encoder for stream #%u\n", i);
146 } else if (dec_ctx->codec_type == AVMEDIA_TYPE_UNKNOWN) {
147 av_log(NULL, AV_LOG_FATAL, "Elementary stream #%d is of unknown type, cannot proceed\n", i);
148 return AVERROR_INVALIDDATA;
150 /* if this stream must be remuxed */
151 ret = avcodec_copy_context(ofmt_ctx->streams[i]->codec,
152 ifmt_ctx->streams[i]->codec);
154 av_log(NULL, AV_LOG_ERROR, "Copying stream context failed\n");
159 if (ofmt_ctx->oformat->flags & AVFMT_GLOBALHEADER)
160 enc_ctx->flags |= CODEC_FLAG_GLOBAL_HEADER;
163 av_dump_format(ofmt_ctx, 0, filename, 1);
165 if (!(ofmt_ctx->oformat->flags & AVFMT_NOFILE)) {
166 ret = avio_open(&ofmt_ctx->pb, filename, AVIO_FLAG_WRITE);
168 av_log(NULL, AV_LOG_ERROR, "Could not open output file '%s'", filename);
173 /* init muxer, write output file header */
174 ret = avformat_write_header(ofmt_ctx, NULL);
176 av_log(NULL, AV_LOG_ERROR, "Error occurred when opening output file\n");
183 static int init_filter(FilteringContext* fctx, AVCodecContext *dec_ctx,
184 AVCodecContext *enc_ctx, const char *filter_spec)
188 AVFilter *buffersrc = NULL;
189 AVFilter *buffersink = NULL;
190 AVFilterContext *buffersrc_ctx = NULL;
191 AVFilterContext *buffersink_ctx = NULL;
192 AVFilterInOut *outputs = avfilter_inout_alloc();
193 AVFilterInOut *inputs = avfilter_inout_alloc();
194 AVFilterGraph *filter_graph = avfilter_graph_alloc();
196 if (!outputs || !inputs || !filter_graph) {
197 ret = AVERROR(ENOMEM);
201 if (dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO) {
202 buffersrc = avfilter_get_by_name("buffer");
203 buffersink = avfilter_get_by_name("buffersink");
204 if (!buffersrc || !buffersink) {
205 av_log(NULL, AV_LOG_ERROR, "filtering source or sink element not found\n");
206 ret = AVERROR_UNKNOWN;
210 snprintf(args, sizeof(args),
211 "video_size=%dx%d:pix_fmt=%d:time_base=%d/%d:pixel_aspect=%d/%d",
212 dec_ctx->width, dec_ctx->height, dec_ctx->pix_fmt,
213 dec_ctx->time_base.num, dec_ctx->time_base.den,
214 dec_ctx->sample_aspect_ratio.num,
215 dec_ctx->sample_aspect_ratio.den);
217 ret = avfilter_graph_create_filter(&buffersrc_ctx, buffersrc, "in",
218 args, NULL, filter_graph);
220 av_log(NULL, AV_LOG_ERROR, "Cannot create buffer source\n");
224 ret = avfilter_graph_create_filter(&buffersink_ctx, buffersink, "out",
225 NULL, NULL, filter_graph);
227 av_log(NULL, AV_LOG_ERROR, "Cannot create buffer sink\n");
231 ret = av_opt_set_bin(buffersink_ctx, "pix_fmts",
232 (uint8_t*)&enc_ctx->pix_fmt, sizeof(enc_ctx->pix_fmt),
233 AV_OPT_SEARCH_CHILDREN);
235 av_log(NULL, AV_LOG_ERROR, "Cannot set output pixel format\n");
238 } else if (dec_ctx->codec_type == AVMEDIA_TYPE_AUDIO) {
239 buffersrc = avfilter_get_by_name("abuffer");
240 buffersink = avfilter_get_by_name("abuffersink");
241 if (!buffersrc || !buffersink) {
242 av_log(NULL, AV_LOG_ERROR, "filtering source or sink element not found\n");
243 ret = AVERROR_UNKNOWN;
247 if (!dec_ctx->channel_layout)
248 dec_ctx->channel_layout =
249 av_get_default_channel_layout(dec_ctx->channels);
250 snprintf(args, sizeof(args),
251 "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%"PRIx64,
252 dec_ctx->time_base.num, dec_ctx->time_base.den, dec_ctx->sample_rate,
253 av_get_sample_fmt_name(dec_ctx->sample_fmt),
254 dec_ctx->channel_layout);
255 ret = avfilter_graph_create_filter(&buffersrc_ctx, buffersrc, "in",
256 args, NULL, filter_graph);
258 av_log(NULL, AV_LOG_ERROR, "Cannot create audio buffer source\n");
262 ret = avfilter_graph_create_filter(&buffersink_ctx, buffersink, "out",
263 NULL, NULL, filter_graph);
265 av_log(NULL, AV_LOG_ERROR, "Cannot create audio buffer sink\n");
269 ret = av_opt_set_bin(buffersink_ctx, "sample_fmts",
270 (uint8_t*)&enc_ctx->sample_fmt, sizeof(enc_ctx->sample_fmt),
271 AV_OPT_SEARCH_CHILDREN);
273 av_log(NULL, AV_LOG_ERROR, "Cannot set output sample format\n");
277 ret = av_opt_set_bin(buffersink_ctx, "channel_layouts",
278 (uint8_t*)&enc_ctx->channel_layout,
279 sizeof(enc_ctx->channel_layout), AV_OPT_SEARCH_CHILDREN);
281 av_log(NULL, AV_LOG_ERROR, "Cannot set output channel layout\n");
285 ret = av_opt_set_bin(buffersink_ctx, "sample_rates",
286 (uint8_t*)&enc_ctx->sample_rate, sizeof(enc_ctx->sample_rate),
287 AV_OPT_SEARCH_CHILDREN);
289 av_log(NULL, AV_LOG_ERROR, "Cannot set output sample rate\n");
293 ret = AVERROR_UNKNOWN;
297 /* Endpoints for the filter graph. */
298 outputs->name = av_strdup("in");
299 outputs->filter_ctx = buffersrc_ctx;
300 outputs->pad_idx = 0;
301 outputs->next = NULL;
303 inputs->name = av_strdup("out");
304 inputs->filter_ctx = buffersink_ctx;
308 if (!outputs->name || !inputs->name) {
309 ret = AVERROR(ENOMEM);
313 if ((ret = avfilter_graph_parse_ptr(filter_graph, filter_spec,
314 &inputs, &outputs, NULL)) < 0)
317 if ((ret = avfilter_graph_config(filter_graph, NULL)) < 0)
320 /* Fill FilteringContext */
321 fctx->buffersrc_ctx = buffersrc_ctx;
322 fctx->buffersink_ctx = buffersink_ctx;
323 fctx->filter_graph = filter_graph;
326 avfilter_inout_free(&inputs);
327 avfilter_inout_free(&outputs);
332 static int init_filters(void)
334 const char *filter_spec;
337 filter_ctx = av_malloc_array(ifmt_ctx->nb_streams, sizeof(*filter_ctx));
339 return AVERROR(ENOMEM);
341 for (i = 0; i < ifmt_ctx->nb_streams; i++) {
342 filter_ctx[i].buffersrc_ctx = NULL;
343 filter_ctx[i].buffersink_ctx = NULL;
344 filter_ctx[i].filter_graph = NULL;
345 if (!(ifmt_ctx->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO
346 || ifmt_ctx->streams[i]->codec->codec_type == AVMEDIA_TYPE_VIDEO))
350 if (ifmt_ctx->streams[i]->codec->codec_type == AVMEDIA_TYPE_VIDEO)
351 filter_spec = "null"; /* passthrough (dummy) filter for video */
353 filter_spec = "anull"; /* passthrough (dummy) filter for audio */
354 ret = init_filter(&filter_ctx[i], ifmt_ctx->streams[i]->codec,
355 ofmt_ctx->streams[i]->codec, filter_spec);
362 static int encode_write_frame(AVFrame *filt_frame, unsigned int stream_index, int *got_frame) {
366 int (*enc_func)(AVCodecContext *, AVPacket *, const AVFrame *, int *) =
367 (ifmt_ctx->streams[stream_index]->codec->codec_type ==
368 AVMEDIA_TYPE_VIDEO) ? avcodec_encode_video2 : avcodec_encode_audio2;
371 got_frame = &got_frame_local;
373 av_log(NULL, AV_LOG_INFO, "Encoding frame\n");
374 /* encode filtered frame */
377 av_init_packet(&enc_pkt);
378 ret = enc_func(ofmt_ctx->streams[stream_index]->codec, &enc_pkt,
379 filt_frame, got_frame);
380 av_frame_free(&filt_frame);
386 /* prepare packet for muxing */
387 enc_pkt.stream_index = stream_index;
388 enc_pkt.dts = av_rescale_q_rnd(enc_pkt.dts,
389 ofmt_ctx->streams[stream_index]->codec->time_base,
390 ofmt_ctx->streams[stream_index]->time_base,
391 AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
392 enc_pkt.pts = av_rescale_q_rnd(enc_pkt.pts,
393 ofmt_ctx->streams[stream_index]->codec->time_base,
394 ofmt_ctx->streams[stream_index]->time_base,
395 AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
396 enc_pkt.duration = av_rescale_q(enc_pkt.duration,
397 ofmt_ctx->streams[stream_index]->codec->time_base,
398 ofmt_ctx->streams[stream_index]->time_base);
400 av_log(NULL, AV_LOG_DEBUG, "Muxing frame\n");
401 /* mux encoded frame */
402 ret = av_interleaved_write_frame(ofmt_ctx, &enc_pkt);
406 static int filter_encode_write_frame(AVFrame *frame, unsigned int stream_index)
411 av_log(NULL, AV_LOG_INFO, "Pushing decoded frame to filters\n");
412 /* push the decoded frame into the filtergraph */
413 ret = av_buffersrc_add_frame_flags(filter_ctx[stream_index].buffersrc_ctx,
416 av_log(NULL, AV_LOG_ERROR, "Error while feeding the filtergraph\n");
420 /* pull filtered frames from the filtergraph */
422 filt_frame = av_frame_alloc();
424 ret = AVERROR(ENOMEM);
427 av_log(NULL, AV_LOG_INFO, "Pulling filtered frame from filters\n");
428 ret = av_buffersink_get_frame(filter_ctx[stream_index].buffersink_ctx,
431 /* if no more frames for output - returns AVERROR(EAGAIN)
432 * if flushed and no more frames for output - returns AVERROR_EOF
433 * rewrite retcode to 0 to show it as normal procedure completion
435 if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF)
437 av_frame_free(&filt_frame);
441 filt_frame->pict_type = AV_PICTURE_TYPE_NONE;
442 ret = encode_write_frame(filt_frame, stream_index, NULL);
450 static int flush_encoder(unsigned int stream_index)
455 if (!(ofmt_ctx->streams[stream_index]->codec->codec->capabilities &
460 av_log(NULL, AV_LOG_INFO, "Flushing stream #%u encoder\n", stream_index);
461 ret = encode_write_frame(NULL, stream_index, &got_frame);
470 int main(int argc, char **argv)
473 AVPacket packet = { .data = NULL, .size = 0 };
474 AVFrame *frame = NULL;
475 enum AVMediaType type;
476 unsigned int stream_index;
479 int (*dec_func)(AVCodecContext *, AVFrame *, int *, const AVPacket *);
482 av_log(NULL, AV_LOG_ERROR, "Usage: %s <input file> <output file>\n", argv[0]);
487 avfilter_register_all();
489 if ((ret = open_input_file(argv[1])) < 0)
491 if ((ret = open_output_file(argv[2])) < 0)
493 if ((ret = init_filters()) < 0)
496 /* read all packets */
498 if ((ret = av_read_frame(ifmt_ctx, &packet)) < 0)
500 stream_index = packet.stream_index;
501 type = ifmt_ctx->streams[packet.stream_index]->codec->codec_type;
502 av_log(NULL, AV_LOG_DEBUG, "Demuxer gave frame of stream_index %u\n",
505 if (filter_ctx[stream_index].filter_graph) {
506 av_log(NULL, AV_LOG_DEBUG, "Going to reencode&filter the frame\n");
507 frame = av_frame_alloc();
509 ret = AVERROR(ENOMEM);
512 packet.dts = av_rescale_q_rnd(packet.dts,
513 ifmt_ctx->streams[stream_index]->time_base,
514 ifmt_ctx->streams[stream_index]->codec->time_base,
515 AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
516 packet.pts = av_rescale_q_rnd(packet.pts,
517 ifmt_ctx->streams[stream_index]->time_base,
518 ifmt_ctx->streams[stream_index]->codec->time_base,
519 AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
520 dec_func = (type == AVMEDIA_TYPE_VIDEO) ? avcodec_decode_video2 :
521 avcodec_decode_audio4;
522 ret = dec_func(ifmt_ctx->streams[stream_index]->codec, frame,
523 &got_frame, &packet);
525 av_frame_free(&frame);
526 av_log(NULL, AV_LOG_ERROR, "Decoding failed\n");
531 frame->pts = av_frame_get_best_effort_timestamp(frame);
532 ret = filter_encode_write_frame(frame, stream_index);
533 av_frame_free(&frame);
537 av_frame_free(&frame);
540 /* remux this frame without reencoding */
541 packet.dts = av_rescale_q_rnd(packet.dts,
542 ifmt_ctx->streams[stream_index]->time_base,
543 ofmt_ctx->streams[stream_index]->time_base,
544 AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
545 packet.pts = av_rescale_q_rnd(packet.pts,
546 ifmt_ctx->streams[stream_index]->time_base,
547 ofmt_ctx->streams[stream_index]->time_base,
548 AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
550 ret = av_interleaved_write_frame(ofmt_ctx, &packet);
554 av_free_packet(&packet);
557 /* flush filters and encoders */
558 for (i = 0; i < ifmt_ctx->nb_streams; i++) {
560 if (!filter_ctx[i].filter_graph)
562 ret = filter_encode_write_frame(NULL, i);
564 av_log(NULL, AV_LOG_ERROR, "Flushing filter failed\n");
569 ret = flush_encoder(i);
571 av_log(NULL, AV_LOG_ERROR, "Flushing encoder failed\n");
576 av_write_trailer(ofmt_ctx);
578 av_free_packet(&packet);
579 av_frame_free(&frame);
580 for (i = 0; i < ifmt_ctx->nb_streams; i++) {
581 avcodec_close(ifmt_ctx->streams[i]->codec);
582 if (ofmt_ctx && ofmt_ctx->nb_streams > i && ofmt_ctx->streams[i] && ofmt_ctx->streams[i]->codec)
583 avcodec_close(ofmt_ctx->streams[i]->codec);
584 if (filter_ctx && filter_ctx[i].filter_graph)
585 avfilter_graph_free(&filter_ctx[i].filter_graph);
588 avformat_close_input(&ifmt_ctx);
589 if (ofmt_ctx && !(ofmt_ctx->oformat->flags & AVFMT_NOFILE))
590 avio_close(ofmt_ctx->pb);
591 avformat_free_context(ofmt_ctx);
594 av_log(NULL, AV_LOG_ERROR, "Error occurred: %s\n", av_err2str(ret));