FFMPEG音视频同步-音频实时采集编码封装成MP4
FFMPEG-音频实时采集编码封装成MP4
//-------------------------------------------------------------------------------------------------
参考链接1、https://blog.csdn.net/leixiaohua1020/article/details/39702113
参考链接2、https://blog.csdn.net/li_wen01/article/details/67631687
//-------------------------------------------------------------------------------------------------
音视频同步录制相关文章
//-------------------------------------------------------------------------------------------------
1、 ffmpeg-摄像头采集保存
2、 ffmpeg音视频同步-摄像头采集编码封装
3、 ffmpeg-音频正弦产生并编码封装
4、 ffmpeg-音频实时采集保存
5、 ffmpeg音视频同步-音频实时采集编码封装
6、 ffmpeg音视频同步-音视频实时采集编码封装
7、 ffmpeg音视频同步-音视频实时采集编码推流
8、 ffmpeg音视频同步-音视频实时采集编码推流-优化版本
//---------------------------------------------------------------
系统环境：
系统版本：lubuntu 16.04
Ffmpge版本：ffmpeg version N-93527-g1125277
摄像头：1.3M HD WebCan
虚拟机：Oracle VM VirtualBox 5.2.22
指令查看设备 ffmpeg -devices
指令播放实时音频 ffplay -f alsa -showmode 1 -ac 2 -i default -ar 44100
https://blog.csdn.net/Tang_Chuanlin/article/details/86775881
指令录制实时音频ffmpeg -f alsa -ar 44100 -i hw:0,0 audio.wav
ffmpeg -f alsa -ar 44100 -i default ffmpeg_record_audio.wav
https://blog.csdn.net/dingjianfeng2014/article/details/57424473
指令分离音视频:ffmpeg -i test.mp4 -vn -y -acodec copy test.aac
ffmpeg -i test.mp4 -vn -y -avcodec copy test.h264
本章文档基于ubuntu alsa格式。实时采集音频一帧数据并将其转化为pcm,编码成AAC,封装成mp4保存下来
1.简介
FFmpeg中有一个和多媒体设备交互的类库：Libavdevice。使用这个库可以读取电脑（或者其他设备上）的多媒体设备的数据，或者输出数据到指定的多媒体设备上。
ffmpeg中的音频相关内容
ffmpeg中AVPacket中可能会含有多个音频帧（AVFrame）,一个音频帧含有多个采样，采样率决定了1s有多少个采样（或者说一个音频帧可以播放多久）。对于aac,一帧有1024个采样，mp3一帧则固定为1152个字节。
对于音频，由于有多声道的音频，那么音频解码出来的数据不同声道也储存在不同的指针，如data[0]是左声道,data[1]是右声道，由于各个声道的数据长度是一样的，所以linesize[0]就代表了所有声道数据的长度。
成员extended_data则指向了data，是一个拓展，上面可以看到data 是包含8个指针的数组，也就是说对于音频，最多只支持8个声道
如果是以下格式：
AV_SAMPLE_FMT_U8, ///< unsigned 8 bits
AV_SAMPLE_FMT_S16, ///< signed 16 bits
AV_SAMPLE_FMT_S32, ///< signed 32 bits
AV_SAMPLE_FMT_FLT, ///< float
AV_SAMPLE_FMT_DBL, ///< double
只能保存再AVFrame的uint8_t *data[0]; 音频保持格式如下：
LRLRLR。。。。
如果是以下格式
AV_SAMPLE_FMT_U8P, ///< unsigned 8 bits, planar
AV_SAMPLE_FMT_S16P, ///< signed 16 bits, planar
AV_SAMPLE_FMT_S32P, ///< signed 32 bits, planar
AV_SAMPLE_FMT_FLTP, ///< float, planar //aac只支持此格式
AV_SAMPLE_FMT_DBLP, ///< double, planar
plane 0: LLLLLLLLLLLLLLLLLLLLLLLLLL…
plane 1: RRRRRRRRRRRRRRRRRRRR…
plane 0对于uint8_t *data[0];
plane 1对于uint8_t *data[1];
2.源码
最简单的基于Libavdevice的摄像头数据读取一帧帧pcm数据，经过音频重采样获取目标AAC的音频源数据参数，保存成MP4文件
1. #include <stdlib.h>
2. #include <stdio.h>
3. #include <string.h>
4. #include <math.h>
5.
6. #include <libavutil/avassert.h>
7. #include <libavutil/channel_layout.h>
8. #include <libavutil/opt.h>
9. #include <libavutil/mathematics.h>
10. #include <libavutil/timestamp.h>
11. #include <libavformat/avformat.h>
12. #include <libswscale/swscale.h>
13. #include <libswresample/swresample.h>
14.
15. #define STREAM_DURATION 10.0
16. #define STREAM_FRAME_RATE 25 /* 25 images/s */
17. #define STREAM_PIX_FMT AV_PIX_FMT_YUV420P /* default pix_fmt */
18.
19. #define SCALE_FLAGS SWS_BICUBIC
20. #define OUTPUT_PCM 1
21. #define ALSA_GET_ENABLE 1
22. #define MAX_AUDIO_FRAME_SIZE 192000
23.
24.
25. // a wrapper around a single output AVStream
26. typedef struct OutputStream {
27. AVStream *st;
28. AVCodecContext *enc;
29.
30. /* pts of the next frame that will be generated */
31. int64_t next_pts;
32. int samples_count;
33.
34. AVFrame *frame;
35. AVFrame *tmp_frame;
36.
37. float t, tincr, tincr2;
38.
39. struct SwsContext *sws_ctx;
40. struct SwrContext *swr_ctx;
41. } OutputStream;
42.
43.
44. typedef struct IntputDev {
45.
46. AVCodecContext *pCodecCtx;
47. AVCodec *pCodec;
48. AVFormatContext *a_ifmtCtx;
49. int audioindex;
50. AVFrame *pAudioFrame;
51. AVPacket *in_packet;
52. struct SwrContext *audio_convert_ctx;
53. uint8_t *dst_buffer;
54. int out_buffer_size;
55. }IntputDev;
56.
57. static void log_packet(const AVFormatContext *fmt_ctx, const AVPacket *pkt)
58. {
59. AVRational *time_base = &fmt_ctx->streams[pkt->stream_index]->time_base;
60.
61. printf("pts:%s pts_time:%s dts:%s dts_time:%s duration:%s duration_time:%s stream_index:%d\n",
62. av_ts2str(pkt->pts), av_ts2timestr(pkt->pts, time_base),
63. av_ts2str(pkt->dts), av_ts2timestr(pkt->dts, time_base),
64. av_ts2str(pkt->duration), av_ts2timestr(pkt->duration, time_base),
65. pkt->stream_index);
66. }
67.
68. static int write_frame(AVFormatContext *fmt_ctx, const AVRational *time_base, AVStream *st, AVPacket *pkt)
69. {
70. /* rescale output packet timestamp values from codec to stream timebase */
71. av_packet_rescale_ts(pkt, *time_base, st->time_base);
72. pkt->stream_index = st->index;
73.
74. /* Write the compressed frame to the media file. */
75. log_packet(fmt_ctx, pkt);
76. return av_interleaved_write_frame(fmt_ctx, pkt);
77. }
78.
79. /* Add an output stream. */
80. static void add_stream(OutputStream *ost, AVFormatContext *oc,
81. AVCodec **codec,
82. enum AVCodecID codec_id)
83. {
84. AVCodecContext *c;
85. int i;
86.
87. /* find the encoder */
88. *codec = avcodec_find_encoder(codec_id);
89. if (!(*codec)) {
90. fprintf(stderr, "Could not find encoder for '%s'\n",
91. avcodec_get_name(codec_id));
92. exit(1);
93. }
94.
95. ost->st = avformat_new_stream(oc, NULL);
96. if (!ost->st) {
97. fprintf(stderr, "Could not allocate stream\n");
98. exit(1);
99. }
100. ost->st->id = oc->nb_streams-1;
101. c = avcodec_alloc_context3(*codec);
102. if (!c) {
103. fprintf(stderr, "Could not alloc an encoding context\n");
104. exit(1);
105. }
106. ost->enc = c;
107.
108. switch ((*codec)->type) {
109. int default_sample_rate=48000;//44100
110. case AVMEDIA_TYPE_AUDIO:
111. c->sample_fmt = (*codec)->sample_fmts ?
112. (*codec)->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
113. c->bit_rate = 64000;
114. c->sample_rate = 48000;
115. if ((*codec)->supported_samplerates) {
116. c->sample_rate = (*codec)->supported_samplerates[0];
117. for (i = 0; (*codec)->supported_samplerates[i]; i++) {
118. if ((*codec)->supported_samplerates[i] == 48000)
119. c->sample_rate = 48000;
120. }
121. }
122. c->channels = av_get_channel_layout_nb_channels(c->channel_layout);
123. c->channel_layout = AV_CH_LAYOUT_STEREO;
124. if ((*codec)->channel_layouts) {
125. c->channel_layout = (*codec)->channel_layouts[0];
126. for (i = 0; (*codec)->channel_layouts[i]; i++) {
127. if ((*codec)->channel_layouts[i] == AV_CH_LAYOUT_STEREO)
128. c->channel_layout = AV_CH_LAYOUT_STEREO;
129. }
130. }
131. c->channels = av_get_channel_layout_nb_channels(c->channel_layout);
132. ost->st->time_base = (AVRational){ 1, c->sample_rate };
133. break;
134.
135. case AVMEDIA_TYPE_VIDEO:
136. c->codec_id = codec_id;
137.
138. c->bit_rate = 400000;
139. /* Resolution must be a multiple of two. */
140. c->width = 352;
141. c->height = 288;
142. /* timebase: This is the fundamental unit of time (in seconds) in terms
143. * of which frame timestamps are represented. For fixed-fps content,
144. * timebase should be 1/framerate and timestamp increments should be
145. * identical to 1. */
146. ost->st->time_base = (AVRational){ 1, STREAM_FRAME_RATE };
147. c->time_base = ost->st->time_base;
148.
149. c->gop_size = 12; /* emit one intra frame every twelve frames at most */
150. c->pix_fmt = STREAM_PIX_FMT;
151. if (c->codec_id == AV_CODEC_ID_MPEG2VIDEO) {
152. /* just for testing, we also add B-frames */
153. c->max_b_frames = 2;
154. }
155. if (c->codec_id == AV_CODEC_ID_MPEG1VIDEO) {
156. /* Needed to avoid using macroblocks in which some coeffs overflow.
157. * This does not happen with normal video, it just happens here as
158. * the motion of the chroma plane does not match the luma plane. */
159. c->mb_decision = 2;
160. }
161. break;
162.
163. default:
164. break;
165. }
166.
167. /* Some formats want stream headers to be separate. */
168. if (oc->oformat->flags & AVFMT_GLOBALHEADER)
169. c->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
170. }
171.
172. /**************************************************************/
173. /* audio output */
174.
175. static AVFrame *alloc_audio_frame(enum AVSampleFormat sample_fmt,
176. uint64_t channel_layout,
177. int sample_rate, int nb_samples)
178. {
179. AVFrame *frame = av_frame_alloc();
180. int ret;
181.
182. if (!frame) {
183. fprintf(stderr, "Error allocating an audio frame\n");
184. exit(1);
185. }
186.
187. frame->format = sample_fmt;
188. frame->channel_layout = channel_layout;
189. frame->sample_rate = sample_rate;
190. frame->nb_samples = nb_samples;
191.
192. if (nb_samples) {
193. ret = av_frame_get_buffer(frame, 0);
194. if (ret < 0) {
195. fprintf(stderr, "Error allocating an audio buffer\n");
196. exit(1);
197. }
198. }
199.
200. return frame;
201. }
202.
203. static void open_audio(AVFormatContext *oc, AVCodec *codec, OutputStream *ost, AVDictionary *opt_arg)
204. {
205. AVCodecContext *c;
206. int nb_samples;
207. int ret;
208. AVDictionary *opt = NULL;
209.
210. c = ost->enc;
211.
212. /* open it */
213. av_dict_copy(&opt, opt_arg, 0);
214. ret = avcodec_open2(c, codec, &opt);
215. av_dict_free(&opt);
216. if (ret < 0) {
217. fprintf(stderr, "Could not open audio codec: %s\n", av_err2str(ret));
218. exit(1);
219. }
220.
221. /* init signal generator */
222. ost->t = 0;
223. ost->tincr = 2 * M_PI * 110.0 / c->sample_rate;
224. /* increment frequency by 110 Hz per second */
225. ost->tincr2 = 2 * M_PI * 110.0 / c->sample_rate / c->sample_rate;
226.
227. if (c->codec->capabilities & AV_CODEC_CAP_VARIABLE_FRAME_SIZE)
228. nb_samples = 10000;
229. else
230. nb_samples = c->frame_size;
231.
232. ost->frame = alloc_audio_frame(c->sample_fmt, c->channel_layout,
233. c->sample_rate, nb_samples);
234. ost->tmp_frame = alloc_audio_frame(AV_SAMPLE_FMT_S16, c->channel_layout,
235. c->sample_rate, nb_samples);
236.
237. printf( "c->channel_layout=%s channel=%d c->sample_fmt=%d c->sample_rate=%d nb_samples=%d\n",
238. av_ts2str(c->channel_layout),c->channels,c->sample_rate,nb_samples,c->sample_fmt);
239.
240. /* copy the stream parameters to the muxer */
241. ret = avcodec_parameters_from_context(ost->st->codecpar, c);
242. if (ret < 0) {
243. fprintf(stderr, "Could not copy the stream parameters\n");
244. exit(1);
245. }
246.
247. /* create resampler context */
248. ost->swr_ctx = swr_alloc();
249. if (!ost->swr_ctx) {
250. fprintf(stderr, "Could not allocate resampler context\n");
251. exit(1);
252. }
253.
254. /* set options */
255. av_opt_set_int (ost->swr_ctx, "in_channel_count", c->channels, 0);
256. av_opt_set_int (ost->swr_ctx, "in_sample_rate", c->sample_rate, 0);
257. av_opt_set_sample_fmt(ost->swr_ctx, "in_sample_fmt", AV_SAMPLE_FMT_S16, 0);
258. av_opt_set_int (ost->swr_ctx, "out_channel_count", c->channels, 0);
259. av_opt_set_int (ost->swr_ctx, "out_sample_rate", c->sample_rate, 0);
260. av_opt_set_sample_fmt(ost->swr_ctx, "out_sample_fmt", c->sample_fmt, 0);
261.
262. /* initialize the resampling context */
263. if ((ret = swr_init(ost->swr_ctx)) < 0) {
264. fprintf(stderr, "Failed to initialize the resampling context\n");
265. exit(1);
266. }
267. }
268.
269. /* Prepare a 16 bit dummy audio frame of 'frame_size' samples and
270. * 'nb_channels' channels. */
271. static AVFrame *get_audio_frame(OutputStream *ost)
272. {
273. AVFrame *frame = ost->tmp_frame;
274. int j, i, v;
275. int16_t *q = (int16_t*)frame->data[0];
276.
277. /* check if we want to generate more frames */
278. if (av_compare_ts(ost->next_pts, ost->enc->time_base,
279. STREAM_DURATION, (AVRational){ 1, 1 }) >= 0)
280. return NULL;
281.
282. for (j = 0; j <frame->nb_samples; j++) {
283. v = (int)(sin(ost->t) * 10000);
284. for (i = 0; i < ost->enc->channels; i++)
285. *q++ = v;
286. ost->t += ost->tincr;
287. ost->tincr += ost->tincr2;
288. }
289.
290. frame->pts = ost->next_pts;
291. ost->next_pts += frame->nb_samples;
292.
293. return frame;
294. }
295.
296. /*
297. * encode one audio frame and send it to the muxer
298. * return 1 when encoding is finished, 0 otherwise
299. */
300. static int write_audio_frame(AVFormatContext *oc, OutputStream *ost)
301. {
302. AVCodecContext *c;
303. AVPacket pkt = { 0 }; // data and size must be 0;
304. AVFrame *frame;
305. int ret;
306. int got_packet;
307. int dst_nb_samples;
308.
309. av_init_packet(&pkt);
310. c = ost->enc;
311.
312. frame = get_audio_frame(ost);
313.
314. if (frame) {
315. /* convert samples from native format to destination codec format, using the resampler */
316. /* compute destination number of samples */
317. dst_nb_samples = av_rescale_rnd(swr_get_delay(ost->swr_ctx, c->sample_rate) + frame->nb_samples,
318. c->sample_rate, c->sample_rate, AV_ROUND_UP);
319. av_assert0(dst_nb_samples == frame->nb_samples);
320.
321. /* when we pass a frame to the encoder, it may keep a reference to it
322. * internally;
323. * make sure we do not overwrite it here
324. */
325. ret = av_frame_make_writable(ost->frame);
326. if (ret < 0)
327. exit(1);
328.
329. /* convert to destination format */
330. ret = swr_convert(ost->swr_ctx,
331. ost->frame->data, dst_nb_samples,
332. (const uint8_t **)frame->data, frame->nb_samples);
333. if (ret < 0) {
334. fprintf(stderr, "Error while converting\n");
335. exit(1);
336. }
337.
338. frame = ost->frame;
339.
340. frame->pts = av_rescale_q(ost->samples_count, (AVRational){1, c->sample_rate}, c->time_base);
341. ost->samples_count += dst_nb_samples;
342. }
343.
344. ret = avcodec_encode_audio2(c, &pkt, frame, &got_packet);
345. if (ret < 0) {
346. fprintf(stderr, "Error encoding audio frame: %s\n", av_err2str(ret));
347. exit(1);
348. }
349.
350. if (got_packet) {
351. ret = write_frame(oc, &c->time_base, ost->st, &pkt);
352. if (ret < 0) {
353. fprintf(stderr, "Error while writing audio frame: %s\n",
354. av_err2str(ret));
355. exit(1);
356. }
357. }
358.
359. return (frame || got_packet) ? 0 : 1;
360. }
361.
362. /**************************************************************/
363. static AVFrame *get_audio_frame1(OutputStream *ost,IntputDev* input,int *got_pcm)
364. {
365. int j, i, v,ret,got_picture;
366. AVFrame *ret_frame=NULL;
367.
368. AVFrame *frame = ost->tmp_frame;
369.
370.
371. *got_pcm=1;
372. /* check if we want to generate more frames */
373. if (av_compare_ts(ost->next_pts, ost->enc->time_base,
374. STREAM_DURATION, (AVRational){ 1, 1 }) >= 0)
375. return NULL;
376.
377. if(av_read_frame(input->a_ifmtCtx, input->in_packet)>=0){
378. if(input->in_packet->stream_index==input->audioindex){
379. ret = avcodec_decode_audio4(input->pCodecCtx, input->pAudioFrame , &got_picture, input->in_packet);
380.
381. *got_pcm=got_picture;
382.
383. if(ret < 0){
384. printf("Decode Error.\n");
385. av_free_packet(input->in_packet);
386. return NULL;
387. }
388. if(got_picture){
389.
390. printf("src nb_samples %d dst nb-samples=%d out_buffer_size=%d\n",
391. input->pAudioFrame->nb_samples,frame->nb_samples,input->out_buffer_size);
392.
393. swr_convert(input->audio_convert_ctx, &input->dst_buffer, MAX_AUDIO_FRAME_SIZE, (const uint8_t **)input->pAudioFrame->data, input->pAudioFrame->nb_samples);
394.
395. frame->pts = ost->next_pts;
396. ost->next_pts += frame->nb_samples;
397.
398. if(frame->nb_samples*4==input->out_buffer_size)//16bit stereo
399. {
400. //memcpy(frame->data,input->dst_buffer,input->out_buffer_size);
401. memcpy(frame->data[0],input->dst_buffer,frame->nb_samples*4);
402. ret_frame= frame;
403. }
404. }
405. }
406. av_free_packet(input->in_packet);
407. }
408. return frame;
409. }
410.
411. static int write_audio_frame1(AVFormatContext *oc, OutputStream *ost,AVFrame *in_frame)
412. {
413. AVCodecContext *c;
414. AVPacket pkt = { 0 }; // data and size must be 0;
415. int ret;
416. int got_packet;
417. int dst_nb_samples;
418.
419. //if(in_frame==NULL)
420. // return 1;
421.
422.
423. av_init_packet(&pkt);
424.
425. AVFrame *frame=in_frame;
426.
427. c = ost->enc;
428.
429. if (frame) {
430. /* convert samples from native format to destination codec format, using the resampler */
431. /* compute destination number of samples */
432. dst_nb_samples = av_rescale_rnd(swr_get_delay(ost->swr_ctx, c->sample_rate) + frame->nb_samples,
433. c->sample_rate, c->sample_rate, AV_ROUND_UP);
434. av_assert0(dst_nb_samples == frame->nb_samples);
435.
436. /* when we pass a frame to the encoder, it may keep a reference to it
437. * internally;
438. * make sure we do not overwrite it here
439. */
440. ret = av_frame_make_writable(ost->frame);
441. if (ret < 0)
442. exit(1);
443.
444. /* convert to destination format */
445. ret = swr_convert(ost->swr_ctx,
446. ost->frame->data, dst_nb_samples,
447. (const uint8_t **)frame->data, frame->nb_samples);
448. if (ret < 0) {
449. fprintf(stderr, "Error while converting\n");
450. exit(1);
451. }
452. frame = ost->frame;
453.
454. frame->pts = av_rescale_q(ost->samples_count, (AVRational){1, c->sample_rate}, c->time_base);
455. ost->samples_count += dst_nb_samples;
456. }
457.
458. ret = avcodec_encode_audio2(c, &pkt, frame, &got_packet);
459. if (ret < 0) {
460. fprintf(stderr, "Error encoding audio frame: %s\n", av_err2str(ret));
461. exit(1);
462. }
463.
464. if (got_packet) {
465. ret = write_frame(oc, &c->time_base, ost->st, &pkt);
466. if (ret < 0) {
467. fprintf(stderr, "Error while writing audio frame: %s\n",
468. av_err2str(ret));
469. exit(1);
470. }
471. }
472.
473. return (frame || got_packet) ? 0 : 1;
474. }
475.
476.
477.
478. static void close_stream(AVFormatContext *oc, OutputStream *ost)
479. {
480. avcodec_free_context(&ost->enc);
481. av_frame_free(&ost->frame);
482. av_frame_free(&ost->tmp_frame);
483. sws_freeContext(ost->sws_ctx);
484. swr_free(&ost->swr_ctx);
485. }
486.
487. /**************************************************************/
488. /* media file output */
489.
490. static int flush_encoder(AVFormatContext *fmt_ctx,unsigned int stream_index)
491. {
492. int ret;
493. int got_frame;
494. AVPacket enc_pkt;
495. if (!(fmt_ctx->streams[stream_index]->codec->codec->capabilities &
496. AV_CODEC_CAP_DELAY))
497. return 0;
498. while (1) {
499. printf("Flushing stream #%u encoder\n", stream_index);
500. //ret = encode_write_frame(NULL, stream_index, &got_frame);
501. enc_pkt.data = NULL;
502. enc_pkt.size = 0;
503. av_init_packet(&enc_pkt);
504. ret = avcodec_encode_audio2 (fmt_ctx->streams[stream_index]->codec, &enc_pkt,
505. NULL, &got_frame);
506. av_frame_free(NULL);
507. if (ret < 0)
508. break;
509. if (!got_frame){
510. ret=0;
511. break;
512. }
513. printf("Succeed to encode 1 frame! 编码成功1帧！\n");
514. /* mux encoded frame */
515. ret = av_interleaved_write_frame(fmt_ctx, &enc_pkt);
516. if (ret < 0)
517. break;
518. }
519. return ret;
520. }
521.
522.
523.
524. int main(int argc, char **argv)
525. {
526. OutputStream video_st = { 0 }, audio_st = { 0 };
527. const char *filename;
528. AVOutputFormat *fmt;
529. AVFormatContext *oc;
530. AVCodec *audio_codec, *video_codec;
531. int ret;
532. int have_video = 0, have_audio = 0;
533. int encode_video = 0, encode_audio = 0;
534. AVDictionary *opt = NULL;
535. int i;
536.
537. if (argc < 2) {
538. printf("usage: %s output_file\n"
539. "API example program to output a media file with libavformat.\n"
540. "This program generates a synthetic audio and video stream, encodes and\n"
541. "muxes them into a file named output_file.\n"
542. "The output format is automatically guessed according to the file extension.\n"
543. "Raw images can also be output by using '%%d' in the filename.\n"
544. "\n", argv[0]);
545. return 1;
546. }
547.
548. filename = argv[1];
549. for (i = 2; i+1 < argc; i+=2) {
550. if (!strcmp(argv[i], "-flags") || !strcmp(argv[i], "-fflags"))
551. av_dict_set(&opt, argv[i]+1, argv[i+1], 0);
552. }
553.
554. /* allocate the output media context */
555. avformat_alloc_output_context2(&oc, NULL, NULL, filename);
556. if (!oc) {
557. printf("Could not deduce output format from file extension: using MPEG.\n");
558. avformat_alloc_output_context2(&oc, NULL, "mpeg", filename);
559. }
560. if (!oc)
561. return 1;
562.
563. fmt = oc->oformat;
564.
565. /* Add the audio and video streams using the default format codecs
566. * and initialize the codecs. */
567.
568. if (fmt->audio_codec != AV_CODEC_ID_NONE) {
569. add_stream(&audio_st, oc, &audio_codec, fmt->audio_codec);
570. have_audio = 1;
571. encode_audio = 1;
572. }
573.
574. /* Now that all the parameters are set, we can open the audio and
575. * video codecs and allocate the necessary encode buffers. */
576.
577. if (have_audio)
578. open_audio(oc, audio_codec, &audio_st, opt);
579.
580. av_dump_format(oc, 0, filename, 1);
581.
582. /* open the output file, if needed */
583. if (!(fmt->flags & AVFMT_NOFILE)) {
584. ret = avio_open(&oc->pb, filename, AVIO_FLAG_WRITE);
585. if (ret < 0) {
586. fprintf(stderr, "Could not open '%s': %s\n", filename,
587. av_err2str(ret));
588. return 1;
589. }
590. }
591.
592.
593. //********add alsa read***********//
594. #if ALSA_GET_ENABLE
595. IntputDev alsa_input = { 0 };
596. AVCodecContext *pCodecCtx;
597. AVCodec *pCodec;
598. AVFormatContext *a_ifmtCtx;
599.
600. //Register Device
601. avdevice_register_all();
602.
603. a_ifmtCtx = avformat_alloc_context();
604.
605.
606. //Linux
607. AVInputFormat *ifmt=av_find_input_format("alsa");
608. if(avformat_open_input(&a_ifmtCtx,"default",ifmt,NULL)!=0){
609. printf("Couldn't open input stream.default\n");
610. return -1;
611. }
612.
613.
614. if(avformat_find_stream_info(a_ifmtCtx,NULL)<0)
615. {
616. printf("Couldn't find stream information.\n");
617. return -1;
618. }
619.
620. int audioindex=-1;
621. for(i=0; i<a_ifmtCtx->nb_streams; i++)
622. if(a_ifmtCtx->streams[i]->codec->codec_type==AVMEDIA_TYPE_AUDIO)
623. {
624. audioindex=i;
625. break;
626. }
627. if(audioindex==-1)
628. {
629. printf("Couldn't find a video stream.\n");
630. return -1;
631. }
632.
633. pCodecCtx=a_ifmtCtx->streams[audioindex]->codec;
634. pCodec=avcodec_find_decoder(pCodecCtx->codec_id);
635. if(pCodec==NULL)
636. {
637. printf("Codec not found.\n");
638. return -1;
639. }
640. if(avcodec_open2(pCodecCtx, pCodec,NULL)<0)
641. {
642. printf("Could not open codec.\n");
643. return -1;
644. }
645.
646. AVPacket *in_packet=(AVPacket *)av_malloc(sizeof(AVPacket));
647.
648. AVFrame *pAudioFrame=av_frame_alloc();
649. if(NULL==pAudioFrame)
650. {
651. printf("could not alloc pAudioFrame\n");
652. return -1;
653. }
654.
655. //audio output paramter //resample
656. uint64_t out_channel_layout = AV_CH_LAYOUT_STEREO;
657. int out_sample_fmt = AV_SAMPLE_FMT_S16;
658. int out_nb_samples =1024; //pCodecCtx->frame_size;
659. int out_sample_rate = 48000;
660. int out_nb_channels = av_get_channel_layout_nb_channels(out_channel_layout);
661. int out_buffer_size = av_samples_get_buffer_size(NULL, out_nb_channels, out_nb_samples, out_sample_fmt, 1);
662. uint8_t *dst_buffer=NULL;
663. dst_buffer = (uint8_t *)av_malloc(MAX_AUDIO_FRAME_SIZE);
664. int64_t in_channel_layout = av_get_default_channel_layout(pCodecCtx->channels);
665.
666.
667. printf("audio sample_fmt=%d size=%d channel=%d sample_rate=%d in_channel_layout=%s\n",
668. pCodecCtx->sample_fmt, pCodecCtx->frame_size,
669. pCodecCtx->channels,pCodecCtx->sample_rate,av_ts2str(in_channel_layout));
670.
671. struct SwrContext *audio_convert_ctx = NULL;
672. audio_convert_ctx = swr_alloc();
673. if (audio_convert_ctx == NULL)
674. {
675. printf("Could not allocate SwrContext\n");
676. return -1;
677. }
678.
679. /* set options */
680. av_opt_set_int (audio_convert_ctx, "in_channel_count", pCodecCtx->channels, 0);
681. av_opt_set_int (audio_convert_ctx, "in_sample_rate", pCodecCtx->sample_rate, 0);
682. av_opt_set_sample_fmt(audio_convert_ctx, "in_sample_fmt", pCodecCtx->sample_fmt, 0);
683. av_opt_set_int (audio_convert_ctx, "out_channel_count", out_nb_channels, 0);
684. av_opt_set_int (audio_convert_ctx, "out_sample_rate", out_sample_rate, 0);
685. av_opt_set_sample_fmt(audio_convert_ctx, "out_sample_fmt", out_sample_fmt, 0);
686.
687. /* initialize the resampling context */
688. if ((ret = swr_init(audio_convert_ctx)) < 0) {
689. fprintf(stderr, "Failed to initialize the resampling context\n");
690. exit(1);
691. }
692.
693.
694. alsa_input.in_packet=in_packet;
695. alsa_input.pCodecCtx=pCodecCtx;
696. alsa_input.pCodec=pCodec;
697. alsa_input.a_ifmtCtx=a_ifmtCtx;
698. alsa_input.audioindex=audioindex;
699. alsa_input.pAudioFrame=pAudioFrame;
700. alsa_input.audio_convert_ctx=audio_convert_ctx;
701. alsa_input.dst_buffer=dst_buffer;
702. alsa_input.out_buffer_size=out_buffer_size;
703. #if OUTPUT_PCM
704. FILE *fp_pcm=fopen("output.pcm","wb+");
705. #endif
706.
707. #endif
708. //******************************//
709.
710.
711.
712. /* Write the stream header, if any. */
713. ret = avformat_write_header(oc, &opt);
714. if (ret < 0) {
715. fprintf(stderr, "Error occurred when opening output file: %s\n",
716. av_err2str(ret));
717. return 1;
718. }
719.
720. int got_pcm;
721.
722.
723. int frameCnt=3;
724. while (encode_audio) {
725. /* select the stream to encode */
726.
727. #if ALSA_GET_ENABLE
728.
729. AVFrame *frame=get_audio_frame1(&audio_st,&alsa_input,&got_pcm);
730. if(!got_pcm)
731. {
732. printf("get_audio_frame1 Error.\n");
733. usleep(10000);
734. continue;
735.
736. }
737.
738. encode_audio = !write_audio_frame1(oc, &audio_st,frame);
739. #if OUTPUT_PCM
740. if(frame!=NULL)
741. fwrite(frame->data[0],1,alsa_input.out_buffer_size,fp_pcm); //Y
742.
743. #endif
744.
745. #else
746.
747. encode_audio = !write_audio_frame(oc, &audio_st);
748.
749. #endif
750. }
751.
752. av_write_trailer(oc);
753. #if ALSA_GET_ENABLE
754.
755. #if OUTPUT_PCM
756. fclose(fp_pcm);
757. #endif
758.
759. swr_free(&alsa_input.audio_convert_ctx);
760. avcodec_close(alsa_input.pCodecCtx);
761. av_free(alsa_input.pAudioFrame);
762. av_free(alsa_input.dst_buffer);
763. avformat_close_input(&alsa_input.a_ifmtCtx);
764.
765. #endif
766.
767. /* Close each codec. */
768. if (have_audio)
769. close_stream(oc, &audio_st);
770.
771. if (!(fmt->flags & AVFMT_NOFILE))
772. /* Close the output file. */
773. avio_closep(&oc->pb);
774.
775.
776.
777. /* free the stream */
778. avformat_free_context(oc);
779.
780. return 0;
781. }
3.验证
3.1编译
1. #!/bin/sh
2. export PKG_CONFIG_PATH=/home/quange/ffmpeg_build/lib/pkgconfig/:$PKG_CONFIG_PATH
3. gcc ffmpeg_get_pcm_muxing.c -g -o ffmpeg_get_pcm_ muxing.out -lSDLmain -lSDL `pkg-config "libavcodec" --cflags --libs` `pkg-config "libavformat" --cflags --libs` `pkg-config "libavutil" --cflags --libs` `pkg-config "libswscale" --cflags --libs` `pkg-config "libavdevice" --cflags --libs`
3.2结果
使用软件mp4打开test.mp4,可以听到实时采集的音频
3.3存在的问题
1、本文档源码，采集ALSA获取的源数据是48K,stereo声道，1024采样大小，经转换成pcm的格式，同样是48K,stereo声道1024大小；不同的采样率会出现部分声音数据丢失，具体原因，在 《ffmpeg-音频实时采集保存》具体分析 ，涉及其他分辨率转换，不在本文讨论范围里。
2、由于本例程没有没有对ALSA音频进行重采样，（48000采样率，1024个采样点，双声道，16bit），故若保存成ts，flv等封装格式是，默认采用的音频编码格式分别是mpga,mp3,其采样点为1152。所以运行此程序会有问题，数据无法写入。而且目前暂不探讨音频重采样问题。故想到一种解决方法是，指定AAC音频编码格式（ts,mp4等都支持AAC音频编码，参考《ffmpeg-摄像头采集编码封装》的表格）。
1. if (fmt->audio_codec != AV_CODEC_ID_NONE) {
2. add_stream(&audio_st, oc, &audio_codec,AV_CODEC_ID_AAC);// fmt->audio_codec);
3. have_audio = 1;
4. encode_audio = 1;
5. }
4.附件
无
5.参考链接
[1] ffmpeg之PCM转AAC
https://blog.csdn.net/mengzhengjie/article/details/78919067
[2]官方Encode pcm file to aac
http://ffmpeg.org/pipermail/ffmpeg-user/2014-January/019268.html
[3]PCM编码AAC,参考其普通PCM格式与AAC转格式差异 https://blog.csdn.net/mengzhengjie/article/details/78919067
[4]https://cloud.tencent.com/developer/article/1194003