transcode.c 11.6 KB
Newer Older
Julian Rother's avatar
Julian Rother committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178
#include <stdio.h>
#include <unistd.h>
#include <libavutil/opt.h>
#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>
#include <libavfilter/avfilter.h>
#include <libavfilter/buffersink.h>
#include <libavfilter/buffersrc.h>

#include "util.h"
#include "config.h"

void assert_empty_opts(AVDictionary *opts)
{
	AVDictionaryEntry *e;
	e = 0;
	while (e = av_dict_get(opts, "", e, AV_DICT_IGNORE_SUFFIX))
		job_failed("Unrecognized option: %s", e->key);
}

static void setup_input_stream(char *stream, AVFormatContext *demux,
		AVCodecContext **decs, AVFilterGraph *fg, AVFilterContext **srcs,
		AVFilterInOut **pads)
{
	char *p;
	int idx, err;
	enum AVMediaType type;
	AVDictionary *opts;
	AVFilterInOut *_inpads;
	AVFilterInOut *_outpads;
	type = AVMEDIA_TYPE_VIDEO;
	if (!strcmp(jstr(jlookup(stream, "type"), ""), "audio"))
		type = AVMEDIA_TYPE_AUDIO;
	idx = av_find_best_stream(demux, type, jint(jlookup(stream, "index"), -1), -1, 0, 0);
	if ((idx == -1 || decs[idx]) && jint(jlookup(stream, "index"), -1) == -1)
		for (idx = 0; idx < demux->nb_streams
				&& (decs[idx] || demux->streams[idx]->codecpar->codec_type == type); idx ++);
	if (idx == -1 || decs[idx])
		job_failed("Could not find suitable input stream");

	/* Prepare Decoder */
	decs[idx] = avcodec_alloc_context3(avcodec_find_decoder(demux->streams[idx]->codecpar->codec_id));
	if (!decs[idx])
		job_failed("Could not allocate decoder context");
	avcodec_parameters_to_context(decs[idx], demux->streams[idx]->codecpar);
	opts = 0;
	parse_dict(&opts, jlookup(stream, "options"));
	if ((err = avcodec_open2(decs[idx], decs[idx]->codec, &opts)) < 0)
		job_failed("Opening decoder failed: %s", av_err2str(err));
	assert_empty_opts(opts);

	/* Prepare Filter Source */
	if (type == AVMEDIA_TYPE_VIDEO)
	{
		p = mprintf("video_size=%dx%d:pix_fmt=%d:time_base=%d/%d:pixel_aspect=%d/%d",
			decs[idx]->width, decs[idx]->height, decs[idx]->pix_fmt,
			demux->streams[idx]->time_base.num, demux->streams[idx]->time_base.den,
			decs[idx]->sample_aspect_ratio.num, decs[idx]->sample_aspect_ratio.den);
		err = avfilter_graph_create_filter(&srcs[idx],
				avfilter_get_by_name("buffer"), 0, p, 0, fg);
		if (err < 0)
			job_failed("Creating buffer source failed: %s", av_err2str(err));
		free(p);
	}
	else if (type == AVMEDIA_TYPE_AUDIO)
	{
		p = mprintf("time_base=%d/%d:sample_rate=%d:sample_fmt=%d:channel_layout=%d:channels=%d",
			demux->streams[idx]->time_base.num, demux->streams[idx]->time_base.den,
			decs[idx]->sample_rate, decs[idx]->sample_fmt, decs[idx]->channel_layout,
			decs[idx]->channels);
		err = avfilter_graph_create_filter(&srcs[idx],
				avfilter_get_by_name("abuffer"), 0, p, 0, fg);
		if (err < 0)
			job_failed("Creating abuffer source failed: %s", av_err2str(err));
		free(p);
	}
	for (; *pads; pads = &(*pads)->next);
	*pads = avfilter_inout_alloc();
	(*pads)->filter_ctx = srcs[idx];
	(*pads)->pad_idx = 0;
	(*pads)->next = 0;
	for (p = jenter(jlookup(stream, "filters")); p; p = jnext(p))
	{
		(*pads)->name = av_strdup("in");
		if ((err = avfilter_graph_parse_ptr(fg, jstr(p, 0), 0, pads, 0)) < 0)
			job_failed("Parsing filter string \"%s\" failed: %s", jstr(p, 0), av_err2str(err));
		if ((*pads)->next)
			job_failed("Unconnected filter pad\n");
	}
	(*pads)->name = av_strdup(jstr(jlookup(stream, "name"), "unknown"));
}

static void prepare_output_stream(char *stream, AVStream *st,
		AVCodecContext **encs, AVFilterGraph *fg, AVFilterContext **sinks,
		AVFilterInOut **pads)
{
	int idx, err;
	char *p;
	AVFilterInOut *_inpads;
	AVFilterInOut *_outpads;
	idx = st->index;
	encs[idx] = avcodec_alloc_context3(avcodec_find_encoder_by_name(
				jstr(jlookup(stream, "codec"), 0)));
	if (!encs[idx])
		job_failed("Could not allocate decoder context");

	/* Prepare Filter Sink */
	if (encs[idx]->codec_type == AVMEDIA_TYPE_VIDEO)
	{
		err = avfilter_graph_create_filter(&sinks[idx],
				avfilter_get_by_name("buffersink"), 0, 0, 0, fg);
		if (err < 0)
			job_failed("Creating buffer sink failed: %s", av_err2str(err));
		av_opt_set_int_list(sinks[idx], "pix_fmts", encs[idx]->codec->pix_fmts,
				AV_PIX_FMT_NONE, AV_OPT_SEARCH_CHILDREN);
	}
	else if (encs[idx]->codec_type == AVMEDIA_TYPE_AUDIO)
	{
		err = avfilter_graph_create_filter(&sinks[idx],
				avfilter_get_by_name("abuffersink"), 0, 0, 0, fg);
		if (err < 0)
			job_failed("Creating abuffer sink failed: %s", av_err2str(err));
		av_opt_set_int_list(sinks[idx], "sample_fmts", encs[idx]->codec->sample_fmts,
				AV_SAMPLE_FMT_NONE, AV_OPT_SEARCH_CHILDREN);
		av_opt_set_int_list(sinks[idx], "sample_rates",
				encs[idx]->codec->supported_samplerates, 0, AV_OPT_SEARCH_CHILDREN);
		av_opt_set_int_list(sinks[idx], "channel_layouts",
				encs[idx]->codec->channel_layouts, 0, AV_OPT_SEARCH_CHILDREN);
	}
	else
		job_failed("Requested codec has unsupported type\n");
	for (; *pads; pads = &(*pads)->next);
	*pads = avfilter_inout_alloc();
	(*pads)->filter_ctx = sinks[idx];
	(*pads)->pad_idx = 0;
	(*pads)->next = 0;
	for (p = jenter(jlookup(stream, "filters")); p; p = jnext(p))
	{
		(*pads)->name = av_strdup("out");
		if ((err = avfilter_graph_parse_ptr(fg, jstr(p, 0), pads, 0, 0)) < 0)
			job_failed("Parsing filter string \"%s\" failed: %s", jstr(p, 0), av_err2str(err));
		if ((*pads)->next)
			job_failed("Unconnected filter pad\n");
	}
	(*pads)->name = av_strdup(jstr(jlookup(stream, "name"), "unknown"));
}

static void setup_output_stream(char *stream, AVStream *st, AVCodecContext *enc,
		AVFilterContext *sink)
{
	int err;
	AVFilterLink *l;
	AVDictionary *opts;
	l = sink->inputs[0];
	if (enc->codec_type == AVMEDIA_TYPE_VIDEO)
	{
		/* Why not use av_buffersink_get_* here? */
		enc->height = l->h;
		enc->width = l->w;
		enc->sample_aspect_ratio = l->sample_aspect_ratio;
		enc->pix_fmt = l->format;
	}
	else if (enc->codec_type == AVMEDIA_TYPE_AUDIO)
	{
		enc->sample_rate = l->sample_rate;
		enc->channel_layout = l->channel_layout;
		enc->channels = l->channels;
		enc->sample_fmt = l->format;
	}
	enc->time_base = l->time_base;
	opts = 0;
	parse_dict(&opts, jlookup(stream, "options"));
	if ((err = avcodec_open2(enc, enc->codec, &opts)) < 0)
		job_failed("Opening encoder failed: %s", av_err2str(err));
	assert_empty_opts(opts);
	avcodec_parameters_from_context(st->codecpar, enc);
}

179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
static void flush_encoder(AVFormatContext *mux, AVFilterContext *sink, AVCodecContext *enc, int idx)
{
	int err;
	AVPacket pkt;
	pkt.data = 0;
	pkt.size = 0;
	av_init_packet(&pkt);
	while (!avcodec_receive_packet(enc, &pkt))
	{
		pkt.stream_index = idx;
		av_packet_rescale_ts(&pkt, sink->inputs[0]->time_base,
				mux->streams[idx]->time_base);
		if (err = av_interleaved_write_frame(mux, &pkt))
			job_failed("Could not write frame: %s", av_err2str(err));
	}
}

static void flush_filtergraph(AVFormatContext *mux, AVFilterContext **sinks, AVCodecContext **encs)
{
	int i;
	AVFrame *frame;
	if (!(frame = av_frame_alloc()))
		exit(99);
	for (i = 0; i < mux->nb_streams; i ++)
		while (av_buffersink_get_frame(sinks[i], frame) >= 0)
		{
			avcodec_send_frame(encs[i], frame);
			av_frame_unref(frame);
			flush_encoder(mux, sinks[i], encs[i], i);
		}
	av_frame_free(&frame);
}

static void filtergraph_send(AVFilterContext *src, AVCodecContext *dec, AVPacket *pkt)
{
	int err;
	AVFrame *frame;
	if (!dec || !src)
		return;
	if (!(frame = av_frame_alloc()))
		exit(99);
	avcodec_send_packet(dec, pkt);
	while (!avcodec_receive_frame(dec, frame))
		if ((err = av_buffersrc_add_frame(src, frame)) < 0)
			job_failed("Could not insert frame into filter graph: %s", av_err2str(err));
	av_frame_free(&frame);
}

Julian Rother's avatar
Julian Rother committed
227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
int main(int argc, char *argv[])
{
	int err, i, progress, _progress;
	char *p, *input, *output, *inpath, *outpath, *tmppath;
	AVFormatContext *demux, *mux;
	AVCodecContext **decs, **encs;
	AVFilterContext **srcs, **sinks;
	AVFilterGraph *fg;
	AVFilterInOut *inpads, *outpads;
	AVPacket pkt;
	AVStream *stream;
	AVDictionary *opts;
	if (argc != 5)
		return 1;
	av_register_all();
	avfilter_register_all();
	init_avlogbuf();
244 245
	pkt.data = 0;
	pkt.size = 0;
Julian Rother's avatar
Julian Rother committed
246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
	av_init_packet(&pkt);
	if (!(fg = avfilter_graph_alloc()))
		return 99;

	jobid = atoi(argv[1]);
	input = jlookup(argv[4], "input");
	inpath = mprintf("%s/%s", CONFIG_VIDEOS_RAW, jstr(jlookup(input, "path"), ""));
	output = jlookup(argv[4], "output");
	outpath = mprintf("%s/%s", CONFIG_VIDEOS_RELEASED, jstr(jlookup(output, "path"), ""));
	tmppath = mprintf("%s/.tmp-%i", CONFIG_VIDEOS_TMP, jobid);

	demux = 0;
	opts = 0;
	parse_dict(&opts, jlookup(input, "options"));
	if (err = avformat_open_input(&demux, inpath, 0, &opts))
		job_failed("Opening input file failed: %s", av_err2str(err));
	assert_empty_opts(opts);
	avformat_find_stream_info(demux, 0);
	decs = xmalloc(sizeof(AVCodecContext *)*demux->nb_streams);
	srcs = xmalloc(sizeof(AVFilterContext *)*demux->nb_streams);
	inpads = 0;
	for (p = jenter(jlookup(input, "streams")); p; p = jnext(p))
		setup_input_stream(p, demux, decs, fg, srcs, &inpads);

270 271 272
	err = avformat_alloc_output_context2(&mux,
			av_guess_format(jstr(jlookup(output, "format"), 0), outpath, 0), 0,
			tmppath);
Julian Rother's avatar
Julian Rother committed
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
	if (err < 0)
		job_failed("Error allocating muxer context: %s", av_err2str(err));
	parse_dict(&mux->metadata, jlookup(output, "metadata"));
	parse_chapters(mux, jlookup(output, "chapters"), demux->duration);
	for (i = 0, p = jenter(jlookup(output, "streams")); p; i ++, p = jnext(p));
	sinks = xmalloc(sizeof(AVFilterContext *)*i);
	encs = xmalloc(sizeof(AVCodecContext *)*i);
	outpads = 0;
	for (i = 0, p = jenter(jlookup(output, "streams")); p; i ++, p = jnext(p))
		prepare_output_stream(p, avformat_new_stream(mux, 0), encs, fg, sinks, &outpads);

	/* TODO: Connect pads of same name before applying any filter strings */
	for (p = jenter(jlookup(argv[4], "filters")); p; p = jnext(p))
		if ((err = avfilter_graph_parse_ptr(fg, jstr(p, 0), &outpads, &inpads, 0)) < 0)
			job_failed("Parsing filter string \"%s\" failed: %s", jstr(p, 0), av_err2str(err));
	if (avfilter_graph_config(fg, 0) < 0)
		job_failed("Error configuring filter graph: %s", av_err2str(err));

	for (i = 0, p = jenter(jlookup(output, "streams")); p; i ++, p = jnext(p))
		setup_output_stream(p, mux->streams[i], encs[i], sinks[i]);
	if ((err = avio_open(&mux->pb, tmppath, AVIO_FLAG_WRITE)) < 0)
		job_failed("Opening temporary file \"%s\" failed: %s", tmppath, av_err2str(err));
	opts = 0;
	parse_dict(&opts, jlookup(output, "options"));
	if ((err = avformat_write_header(mux, &opts)) < 0)
		job_failed("Writing temporary file failed: %s", av_err2str(err));
	assert_empty_opts(opts);

	progress = 0;
	ping_job(jobid, "running", 0);
303

Julian Rother's avatar
Julian Rother committed
304 305 306 307 308 309 310 311 312 313
	while (!av_read_frame(demux, &pkt))
	{
		i = pkt.stream_index;
		_progress = av_rescale_q(pkt.pts, demux->streams[i]->time_base, AV_TIME_BASE_Q)*100/demux->duration;
		if (_progress > progress)
		{
			ping_job(jobid, "running", "{\"progress\": %i, \"log\": \"%s\"}", _progress,
					jescape(get_avlogbuf()));
			progress = _progress;
		}
314 315
		filtergraph_send(srcs[i], decs[i], &pkt);
		flush_filtergraph(mux, sinks, encs);
Julian Rother's avatar
Julian Rother committed
316 317 318
	}

	/* Flush */
319 320 321 322 323 324 325 326
	for (i = 0; i < demux->nb_streams; i ++)
	{
		if (decs[i])
			continue;
		filtergraph_send(srcs[i], decs[i], 0);
		!av_buffersrc_add_frame(srcs[i], 0);
	}
	flush_filtergraph(mux, sinks, encs);
Julian Rother's avatar
Julian Rother committed
327 328 329
	for (i = 0; i < mux->nb_streams; i ++)
	{
		avcodec_send_frame(encs[i], 0);
330
		flush_encoder(mux, sinks[i], encs[i], i);
Julian Rother's avatar
Julian Rother committed
331 332
	}
	av_interleaved_write_frame(mux, 0);
333 334

	avformat_close_input(&demux);
Julian Rother's avatar
Julian Rother committed
335 336 337 338 339 340 341 342 343 344
	if (err = av_write_trailer(mux))
		job_failed("Error writing trailer to temporary file", av_err2str(err));
	avio_closep(&mux->pb);
	if (rename(tmppath, outpath))
		job_failed("Overwriting output file \"%s\" failed: %s", outpath, strerror(errno));
	unlink(tmppath);
	ping_job(jobid, "finished", "{\"hash\": \"%s\", \"duration\": %f, \"filesize\": %i, \"log\": \"%s\"}",
			hashfile(outpath), fileduration(outpath), filesize(outpath), jescape(get_avlogbuf()));
	return 0;
}