You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

388 lines
14KB

  1. /*
  2. * Copyright (c) 2012 Stefano Sabatini
  3. *
  4. * Permission is hereby granted, free of charge, to any person obtaining a copy
  5. * of this software and associated documentation files (the "Software"), to deal
  6. * in the Software without restriction, including without limitation the rights
  7. * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
  8. * copies of the Software, and to permit persons to whom the Software is
  9. * furnished to do so, subject to the following conditions:
  10. *
  11. * The above copyright notice and this permission notice shall be included in
  12. * all copies or substantial portions of the Software.
  13. *
  14. * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  15. * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  16. * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
  17. * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  18. * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  19. * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  20. * THE SOFTWARE.
  21. */
  22. /**
  23. * @file
  24. * Demuxing and decoding example.
  25. *
  26. * Show how to use the libavformat and libavcodec API to demux and
  27. * decode audio and video data.
  28. * @example demuxing_decoding.c
  29. */
  30. #include <libavutil/imgutils.h>
  31. #include <libavutil/samplefmt.h>
  32. #include <libavutil/timestamp.h>
  33. #include <libavformat/avformat.h>
  34. static AVFormatContext *fmt_ctx = NULL;
  35. static AVCodecContext *video_dec_ctx = NULL, *audio_dec_ctx;
  36. static AVStream *video_stream = NULL, *audio_stream = NULL;
  37. static const char *src_filename = NULL;
  38. static const char *video_dst_filename = NULL;
  39. static const char *audio_dst_filename = NULL;
  40. static FILE *video_dst_file = NULL;
  41. static FILE *audio_dst_file = NULL;
  42. static uint8_t *video_dst_data[4] = {NULL};
  43. static int video_dst_linesize[4];
  44. static int video_dst_bufsize;
  45. static int video_stream_idx = -1, audio_stream_idx = -1;
  46. static AVFrame *frame = NULL;
  47. static AVPacket pkt;
  48. static int video_frame_count = 0;
  49. static int audio_frame_count = 0;
  50. /* The different ways of decoding and managing data memory. You are not
  51. * supposed to support all the modes in your application but pick the one most
  52. * appropriate to your needs. Look for the use of api_mode in this example to
  53. * see what are the differences of API usage between them */
  54. enum {
  55. API_MODE_OLD = 0, /* old method, deprecated */
  56. API_MODE_NEW_API_REF_COUNT = 1, /* new method, using the frame reference counting */
  57. API_MODE_NEW_API_NO_REF_COUNT = 2, /* new method, without reference counting */
  58. };
  59. static int api_mode = API_MODE_OLD;
  60. static int decode_packet(int *got_frame, int cached)
  61. {
  62. int ret = 0;
  63. int decoded = pkt.size;
  64. *got_frame = 0;
  65. if (pkt.stream_index == video_stream_idx) {
  66. /* decode video frame */
  67. ret = avcodec_decode_video2(video_dec_ctx, frame, got_frame, &pkt);
  68. if (ret < 0) {
  69. fprintf(stderr, "Error decoding video frame (%s)\n", av_err2str(ret));
  70. return ret;
  71. }
  72. if (*got_frame) {
  73. printf("video_frame%s n:%d coded_n:%d pts:%s\n",
  74. cached ? "(cached)" : "",
  75. video_frame_count++, frame->coded_picture_number,
  76. av_ts2timestr(frame->pts, &video_dec_ctx->time_base));
  77. /* copy decoded frame to destination buffer:
  78. * this is required since rawvideo expects non aligned data */
  79. av_image_copy(video_dst_data, video_dst_linesize,
  80. (const uint8_t **)(frame->data), frame->linesize,
  81. video_dec_ctx->pix_fmt, video_dec_ctx->width, video_dec_ctx->height);
  82. /* write to rawvideo file */
  83. fwrite(video_dst_data[0], 1, video_dst_bufsize, video_dst_file);
  84. }
  85. } else if (pkt.stream_index == audio_stream_idx) {
  86. /* decode audio frame */
  87. ret = avcodec_decode_audio4(audio_dec_ctx, frame, got_frame, &pkt);
  88. if (ret < 0) {
  89. fprintf(stderr, "Error decoding audio frame (%s)\n", av_err2str(ret));
  90. return ret;
  91. }
  92. /* Some audio decoders decode only part of the packet, and have to be
  93. * called again with the remainder of the packet data.
  94. * Sample: fate-suite/lossless-audio/luckynight-partial.shn
  95. * Also, some decoders might over-read the packet. */
  96. decoded = FFMIN(ret, pkt.size);
  97. if (*got_frame) {
  98. size_t unpadded_linesize = frame->nb_samples * av_get_bytes_per_sample(frame->format);
  99. printf("audio_frame%s n:%d nb_samples:%d pts:%s\n",
  100. cached ? "(cached)" : "",
  101. audio_frame_count++, frame->nb_samples,
  102. av_ts2timestr(frame->pts, &audio_dec_ctx->time_base));
  103. /* Write the raw audio data samples of the first plane. This works
  104. * fine for packed formats (e.g. AV_SAMPLE_FMT_S16). However,
  105. * most audio decoders output planar audio, which uses a separate
  106. * plane of audio samples for each channel (e.g. AV_SAMPLE_FMT_S16P).
  107. * In other words, this code will write only the first audio channel
  108. * in these cases.
  109. * You should use libswresample or libavfilter to convert the frame
  110. * to packed data. */
  111. fwrite(frame->extended_data[0], 1, unpadded_linesize, audio_dst_file);
  112. }
  113. }
  114. /* If we use the new API with reference counting, we own the data and need
  115. * to de-reference it when we don't use it anymore */
  116. if (*got_frame && api_mode == API_MODE_NEW_API_REF_COUNT)
  117. av_frame_unref(frame);
  118. return decoded;
  119. }
  120. static int open_codec_context(int *stream_idx,
  121. AVFormatContext *fmt_ctx, enum AVMediaType type)
  122. {
  123. int ret, stream_index;
  124. AVStream *st;
  125. AVCodecContext *dec_ctx = NULL;
  126. AVCodec *dec = NULL;
  127. AVDictionary *opts = NULL;
  128. ret = av_find_best_stream(fmt_ctx, type, -1, -1, NULL, 0);
  129. if (ret < 0) {
  130. fprintf(stderr, "Could not find %s stream in input file '%s'\n",
  131. av_get_media_type_string(type), src_filename);
  132. return ret;
  133. } else {
  134. stream_index = ret;
  135. st = fmt_ctx->streams[stream_index];
  136. /* find decoder for the stream */
  137. dec_ctx = st->codec;
  138. dec = avcodec_find_decoder(dec_ctx->codec_id);
  139. if (!dec) {
  140. fprintf(stderr, "Failed to find %s codec\n",
  141. av_get_media_type_string(type));
  142. return AVERROR(EINVAL);
  143. }
  144. /* Init the decoders, with or without reference counting */
  145. if (api_mode == API_MODE_NEW_API_REF_COUNT)
  146. av_dict_set(&opts, "refcounted_frames", "1", 0);
  147. if ((ret = avcodec_open2(dec_ctx, dec, &opts)) < 0) {
  148. fprintf(stderr, "Failed to open %s codec\n",
  149. av_get_media_type_string(type));
  150. return ret;
  151. }
  152. *stream_idx = stream_index;
  153. }
  154. return 0;
  155. }
  156. static int get_format_from_sample_fmt(const char **fmt,
  157. enum AVSampleFormat sample_fmt)
  158. {
  159. int i;
  160. struct sample_fmt_entry {
  161. enum AVSampleFormat sample_fmt; const char *fmt_be, *fmt_le;
  162. } sample_fmt_entries[] = {
  163. { AV_SAMPLE_FMT_U8, "u8", "u8" },
  164. { AV_SAMPLE_FMT_S16, "s16be", "s16le" },
  165. { AV_SAMPLE_FMT_S32, "s32be", "s32le" },
  166. { AV_SAMPLE_FMT_FLT, "f32be", "f32le" },
  167. { AV_SAMPLE_FMT_DBL, "f64be", "f64le" },
  168. };
  169. *fmt = NULL;
  170. for (i = 0; i < FF_ARRAY_ELEMS(sample_fmt_entries); i++) {
  171. struct sample_fmt_entry *entry = &sample_fmt_entries[i];
  172. if (sample_fmt == entry->sample_fmt) {
  173. *fmt = AV_NE(entry->fmt_be, entry->fmt_le);
  174. return 0;
  175. }
  176. }
  177. fprintf(stderr,
  178. "sample format %s is not supported as output format\n",
  179. av_get_sample_fmt_name(sample_fmt));
  180. return -1;
  181. }
  182. int main (int argc, char **argv)
  183. {
  184. int ret = 0, got_frame;
  185. if (argc != 4 && argc != 5) {
  186. fprintf(stderr, "usage: %s [-refcount=<old|new_norefcount|new_refcount>] "
  187. "input_file video_output_file audio_output_file\n"
  188. "API example program to show how to read frames from an input file.\n"
  189. "This program reads frames from a file, decodes them, and writes decoded\n"
  190. "video frames to a rawvideo file named video_output_file, and decoded\n"
  191. "audio frames to a rawaudio file named audio_output_file.\n\n"
  192. "If the -refcount option is specified, the program use the\n"
  193. "reference counting frame system which allows keeping a copy of\n"
  194. "the data for longer than one decode call. If unset, it's using\n"
  195. "the classic old method.\n"
  196. "\n", argv[0]);
  197. exit(1);
  198. }
  199. if (argc == 5) {
  200. const char *mode = argv[1] + strlen("-refcount=");
  201. if (!strcmp(mode, "old")) api_mode = API_MODE_OLD;
  202. else if (!strcmp(mode, "new_norefcount")) api_mode = API_MODE_NEW_API_NO_REF_COUNT;
  203. else if (!strcmp(mode, "new_refcount")) api_mode = API_MODE_NEW_API_REF_COUNT;
  204. else {
  205. fprintf(stderr, "unknow mode '%s'\n", mode);
  206. exit(1);
  207. }
  208. argv++;
  209. }
  210. src_filename = argv[1];
  211. video_dst_filename = argv[2];
  212. audio_dst_filename = argv[3];
  213. /* register all formats and codecs */
  214. av_register_all();
  215. /* open input file, and allocate format context */
  216. if (avformat_open_input(&fmt_ctx, src_filename, NULL, NULL) < 0) {
  217. fprintf(stderr, "Could not open source file %s\n", src_filename);
  218. exit(1);
  219. }
  220. /* retrieve stream information */
  221. if (avformat_find_stream_info(fmt_ctx, NULL) < 0) {
  222. fprintf(stderr, "Could not find stream information\n");
  223. exit(1);
  224. }
  225. if (open_codec_context(&video_stream_idx, fmt_ctx, AVMEDIA_TYPE_VIDEO) >= 0) {
  226. video_stream = fmt_ctx->streams[video_stream_idx];
  227. video_dec_ctx = video_stream->codec;
  228. video_dst_file = fopen(video_dst_filename, "wb");
  229. if (!video_dst_file) {
  230. fprintf(stderr, "Could not open destination file %s\n", video_dst_filename);
  231. ret = 1;
  232. goto end;
  233. }
  234. /* allocate image where the decoded image will be put */
  235. ret = av_image_alloc(video_dst_data, video_dst_linesize,
  236. video_dec_ctx->width, video_dec_ctx->height,
  237. video_dec_ctx->pix_fmt, 1);
  238. if (ret < 0) {
  239. fprintf(stderr, "Could not allocate raw video buffer\n");
  240. goto end;
  241. }
  242. video_dst_bufsize = ret;
  243. }
  244. if (open_codec_context(&audio_stream_idx, fmt_ctx, AVMEDIA_TYPE_AUDIO) >= 0) {
  245. audio_stream = fmt_ctx->streams[audio_stream_idx];
  246. audio_dec_ctx = audio_stream->codec;
  247. audio_dst_file = fopen(audio_dst_filename, "wb");
  248. if (!audio_dst_file) {
  249. fprintf(stderr, "Could not open destination file %s\n", audio_dst_filename);
  250. ret = 1;
  251. goto end;
  252. }
  253. }
  254. /* dump input information to stderr */
  255. av_dump_format(fmt_ctx, 0, src_filename, 0);
  256. if (!audio_stream && !video_stream) {
  257. fprintf(stderr, "Could not find audio or video stream in the input, aborting\n");
  258. ret = 1;
  259. goto end;
  260. }
  261. /* When using the new API, you need to use the libavutil/frame.h API, while
  262. * the classic frame management is available in libavcodec */
  263. if (api_mode == API_MODE_OLD)
  264. frame = avcodec_alloc_frame();
  265. else
  266. frame = av_frame_alloc();
  267. if (!frame) {
  268. fprintf(stderr, "Could not allocate frame\n");
  269. ret = AVERROR(ENOMEM);
  270. goto end;
  271. }
  272. /* initialize packet, set data to NULL, let the demuxer fill it */
  273. av_init_packet(&pkt);
  274. pkt.data = NULL;
  275. pkt.size = 0;
  276. if (video_stream)
  277. printf("Demuxing video from file '%s' into '%s'\n", src_filename, video_dst_filename);
  278. if (audio_stream)
  279. printf("Demuxing audio from file '%s' into '%s'\n", src_filename, audio_dst_filename);
  280. /* read frames from the file */
  281. while (av_read_frame(fmt_ctx, &pkt) >= 0) {
  282. AVPacket orig_pkt = pkt;
  283. do {
  284. ret = decode_packet(&got_frame, 0);
  285. if (ret < 0)
  286. break;
  287. pkt.data += ret;
  288. pkt.size -= ret;
  289. } while (pkt.size > 0);
  290. av_free_packet(&orig_pkt);
  291. }
  292. /* flush cached frames */
  293. pkt.data = NULL;
  294. pkt.size = 0;
  295. do {
  296. decode_packet(&got_frame, 1);
  297. } while (got_frame);
  298. printf("Demuxing succeeded.\n");
  299. if (video_stream) {
  300. printf("Play the output video file with the command:\n"
  301. "ffplay -f rawvideo -pix_fmt %s -video_size %dx%d %s\n",
  302. av_get_pix_fmt_name(video_dec_ctx->pix_fmt), video_dec_ctx->width, video_dec_ctx->height,
  303. video_dst_filename);
  304. }
  305. if (audio_stream) {
  306. enum AVSampleFormat sfmt = audio_dec_ctx->sample_fmt;
  307. int n_channels = audio_dec_ctx->channels;
  308. const char *fmt;
  309. if (av_sample_fmt_is_planar(sfmt)) {
  310. const char *packed = av_get_sample_fmt_name(sfmt);
  311. printf("Warning: the sample format the decoder produced is planar "
  312. "(%s). This example will output the first channel only.\n",
  313. packed ? packed : "?");
  314. sfmt = av_get_packed_sample_fmt(sfmt);
  315. n_channels = 1;
  316. }
  317. if ((ret = get_format_from_sample_fmt(&fmt, sfmt)) < 0)
  318. goto end;
  319. printf("Play the output audio file with the command:\n"
  320. "ffplay -f %s -ac %d -ar %d %s\n",
  321. fmt, n_channels, audio_dec_ctx->sample_rate,
  322. audio_dst_filename);
  323. }
  324. end:
  325. avcodec_close(video_dec_ctx);
  326. avcodec_close(audio_dec_ctx);
  327. avformat_close_input(&fmt_ctx);
  328. if (video_dst_file)
  329. fclose(video_dst_file);
  330. if (audio_dst_file)
  331. fclose(audio_dst_file);
  332. if (api_mode == API_MODE_OLD)
  333. avcodec_free_frame(&frame);
  334. else
  335. av_frame_free(&frame);
  336. av_free(video_dst_data[0]);
  337. return ret < 0;
  338. }