2 * AVFoundation input device
3 * Copyright (c) 2014 Thilo Borgmann <thilo.borgmann@mail.de>
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24 * AVFoundation input device
25 * @author Thilo Borgmann <thilo.borgmann@mail.de>
28 #import <AVFoundation/AVFoundation.h>
31 #include "libavutil/pixdesc.h"
32 #include "libavutil/opt.h"
33 #include "libavutil/avstring.h"
34 #include "libavformat/internal.h"
35 #include "libavutil/internal.h"
36 #include "libavutil/time.h"
39 static const int avf_time_base = 1000000;
41 static const AVRational avf_time_base_q = {
46 struct AVFPixelFormatSpec {
47 enum AVPixelFormat ff_id;
51 static const struct AVFPixelFormatSpec avf_pixel_formats[] = {
52 { AV_PIX_FMT_MONOBLACK, kCVPixelFormatType_1Monochrome },
53 { AV_PIX_FMT_RGB555BE, kCVPixelFormatType_16BE555 },
54 { AV_PIX_FMT_RGB555LE, kCVPixelFormatType_16LE555 },
55 { AV_PIX_FMT_RGB565BE, kCVPixelFormatType_16BE565 },
56 { AV_PIX_FMT_RGB565LE, kCVPixelFormatType_16LE565 },
57 { AV_PIX_FMT_RGB24, kCVPixelFormatType_24RGB },
58 { AV_PIX_FMT_BGR24, kCVPixelFormatType_24BGR },
59 { AV_PIX_FMT_0RGB, kCVPixelFormatType_32ARGB },
60 { AV_PIX_FMT_BGR0, kCVPixelFormatType_32BGRA },
61 { AV_PIX_FMT_0BGR, kCVPixelFormatType_32ABGR },
62 { AV_PIX_FMT_RGB0, kCVPixelFormatType_32RGBA },
63 { AV_PIX_FMT_BGR48BE, kCVPixelFormatType_48RGB },
64 { AV_PIX_FMT_UYVY422, kCVPixelFormatType_422YpCbCr8 },
65 { AV_PIX_FMT_YUVA444P, kCVPixelFormatType_4444YpCbCrA8R },
66 { AV_PIX_FMT_YUVA444P16LE, kCVPixelFormatType_4444AYpCbCr16 },
67 { AV_PIX_FMT_YUV444P, kCVPixelFormatType_444YpCbCr8 },
68 { AV_PIX_FMT_YUV422P16, kCVPixelFormatType_422YpCbCr16 },
69 { AV_PIX_FMT_YUV422P10, kCVPixelFormatType_422YpCbCr10 },
70 { AV_PIX_FMT_YUV444P10, kCVPixelFormatType_444YpCbCr10 },
71 { AV_PIX_FMT_YUV420P, kCVPixelFormatType_420YpCbCr8Planar },
72 { AV_PIX_FMT_NV12, kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange },
73 { AV_PIX_FMT_YUYV422, kCVPixelFormatType_422YpCbCr8_yuvs },
74 #if !TARGET_OS_IPHONE && __MAC_OS_X_VERSION_MIN_REQUIRED >= 1080
75 { AV_PIX_FMT_GRAY8, kCVPixelFormatType_OneComponent8 },
77 { AV_PIX_FMT_NONE, 0 }
85 int audio_frames_captured;
87 int64_t first_audio_pts;
88 pthread_mutex_t frame_lock;
89 pthread_cond_t frame_wait_cond;
91 id avf_audio_delegate;
94 int video_device_index;
95 int video_stream_index;
96 int audio_device_index;
97 int audio_stream_index;
100 char *audio_filename;
102 int num_video_devices;
105 int audio_bits_per_sample;
108 int audio_signed_integer;
110 int audio_non_interleaved;
112 int32_t *audio_buffer;
113 int audio_buffer_size;
115 enum AVPixelFormat pixel_format;
117 AVCaptureSession *capture_session;
118 AVCaptureVideoDataOutput *video_output;
119 AVCaptureAudioDataOutput *audio_output;
120 CMSampleBufferRef current_frame;
121 CMSampleBufferRef current_audio_frame;
124 static void lock_frames(AVFContext* ctx)
126 pthread_mutex_lock(&ctx->frame_lock);
129 static void unlock_frames(AVFContext* ctx)
131 pthread_mutex_unlock(&ctx->frame_lock);
134 /** FrameReciever class - delegate for AVCaptureSession
136 @interface AVFFrameReceiver : NSObject
138 AVFContext* _context;
141 - (id)initWithContext:(AVFContext*)context;
143 - (void) captureOutput:(AVCaptureOutput *)captureOutput
144 didOutputSampleBuffer:(CMSampleBufferRef)videoFrame
145 fromConnection:(AVCaptureConnection *)connection;
149 @implementation AVFFrameReceiver
151 - (id)initWithContext:(AVFContext*)context
153 if (self = [super init]) {
159 - (void) captureOutput:(AVCaptureOutput *)captureOutput
160 didOutputSampleBuffer:(CMSampleBufferRef)videoFrame
161 fromConnection:(AVCaptureConnection *)connection
163 lock_frames(_context);
165 if (_context->current_frame != nil) {
166 CFRelease(_context->current_frame);
169 _context->current_frame = (CMSampleBufferRef)CFRetain(videoFrame);
171 pthread_cond_signal(&_context->frame_wait_cond);
173 unlock_frames(_context);
175 ++_context->frames_captured;
180 /** AudioReciever class - delegate for AVCaptureSession
182 @interface AVFAudioReceiver : NSObject
184 AVFContext* _context;
187 - (id)initWithContext:(AVFContext*)context;
189 - (void) captureOutput:(AVCaptureOutput *)captureOutput
190 didOutputSampleBuffer:(CMSampleBufferRef)audioFrame
191 fromConnection:(AVCaptureConnection *)connection;
195 @implementation AVFAudioReceiver
197 - (id)initWithContext:(AVFContext*)context
199 if (self = [super init]) {
205 - (void) captureOutput:(AVCaptureOutput *)captureOutput
206 didOutputSampleBuffer:(CMSampleBufferRef)audioFrame
207 fromConnection:(AVCaptureConnection *)connection
209 lock_frames(_context);
211 if (_context->current_audio_frame != nil) {
212 CFRelease(_context->current_audio_frame);
215 _context->current_audio_frame = (CMSampleBufferRef)CFRetain(audioFrame);
217 pthread_cond_signal(&_context->frame_wait_cond);
219 unlock_frames(_context);
221 ++_context->audio_frames_captured;
226 static void destroy_context(AVFContext* ctx)
228 [ctx->capture_session stopRunning];
230 [ctx->capture_session release];
231 [ctx->video_output release];
232 [ctx->audio_output release];
233 [ctx->avf_delegate release];
234 [ctx->avf_audio_delegate release];
236 ctx->capture_session = NULL;
237 ctx->video_output = NULL;
238 ctx->audio_output = NULL;
239 ctx->avf_delegate = NULL;
240 ctx->avf_audio_delegate = NULL;
242 av_freep(&ctx->audio_buffer);
244 pthread_mutex_destroy(&ctx->frame_lock);
245 pthread_cond_destroy(&ctx->frame_wait_cond);
247 if (ctx->current_frame) {
248 CFRelease(ctx->current_frame);
252 static void parse_device_name(AVFormatContext *s)
254 AVFContext *ctx = (AVFContext*)s->priv_data;
255 char *tmp = av_strdup(s->filename);
259 ctx->video_filename = av_strtok(tmp, ":", &save);
260 ctx->audio_filename = av_strtok(NULL, ":", &save);
262 ctx->audio_filename = av_strtok(tmp, ":", &save);
266 static int add_video_device(AVFormatContext *s, AVCaptureDevice *video_device)
268 AVFContext *ctx = (AVFContext*)s->priv_data;
269 NSError *error = nil;
270 AVCaptureInput* capture_input = nil;
271 struct AVFPixelFormatSpec pxl_fmt_spec;
272 NSNumber *pixel_format;
273 NSDictionary *capture_dict;
274 dispatch_queue_t queue;
276 if (ctx->video_device_index < ctx->num_video_devices) {
277 capture_input = (AVCaptureInput*) [[[AVCaptureDeviceInput alloc] initWithDevice:video_device error:&error] autorelease];
279 capture_input = (AVCaptureInput*) video_device;
282 if (!capture_input) {
283 av_log(s, AV_LOG_ERROR, "Failed to create AV capture input device: %s\n",
284 [[error localizedDescription] UTF8String]);
288 if ([ctx->capture_session canAddInput:capture_input]) {
289 [ctx->capture_session addInput:capture_input];
291 av_log(s, AV_LOG_ERROR, "can't add video input to capture session\n");
296 ctx->video_output = [[AVCaptureVideoDataOutput alloc] init];
298 if (!ctx->video_output) {
299 av_log(s, AV_LOG_ERROR, "Failed to init AV video output\n");
303 // select pixel format
304 pxl_fmt_spec.ff_id = AV_PIX_FMT_NONE;
306 for (int i = 0; avf_pixel_formats[i].ff_id != AV_PIX_FMT_NONE; i++) {
307 if (ctx->pixel_format == avf_pixel_formats[i].ff_id) {
308 pxl_fmt_spec = avf_pixel_formats[i];
313 // check if selected pixel format is supported by AVFoundation
314 if (pxl_fmt_spec.ff_id == AV_PIX_FMT_NONE) {
315 av_log(s, AV_LOG_ERROR, "Selected pixel format (%s) is not supported by AVFoundation.\n",
316 av_get_pix_fmt_name(pxl_fmt_spec.ff_id));
320 // check if the pixel format is available for this device
321 if ([[ctx->video_output availableVideoCVPixelFormatTypes] indexOfObject:[NSNumber numberWithInt:pxl_fmt_spec.avf_id]] == NSNotFound) {
322 av_log(s, AV_LOG_ERROR, "Selected pixel format (%s) is not supported by the input device.\n",
323 av_get_pix_fmt_name(pxl_fmt_spec.ff_id));
325 pxl_fmt_spec.ff_id = AV_PIX_FMT_NONE;
327 av_log(s, AV_LOG_ERROR, "Supported pixel formats:\n");
328 for (NSNumber *pxl_fmt in [ctx->video_output availableVideoCVPixelFormatTypes]) {
329 struct AVFPixelFormatSpec pxl_fmt_dummy;
330 pxl_fmt_dummy.ff_id = AV_PIX_FMT_NONE;
331 for (int i = 0; avf_pixel_formats[i].ff_id != AV_PIX_FMT_NONE; i++) {
332 if ([pxl_fmt intValue] == avf_pixel_formats[i].avf_id) {
333 pxl_fmt_dummy = avf_pixel_formats[i];
338 if (pxl_fmt_dummy.ff_id != AV_PIX_FMT_NONE) {
339 av_log(s, AV_LOG_ERROR, " %s\n", av_get_pix_fmt_name(pxl_fmt_dummy.ff_id));
341 // select first supported pixel format instead of user selected (or default) pixel format
342 if (pxl_fmt_spec.ff_id == AV_PIX_FMT_NONE) {
343 pxl_fmt_spec = pxl_fmt_dummy;
348 // fail if there is no appropriate pixel format or print a warning about overriding the pixel format
349 if (pxl_fmt_spec.ff_id == AV_PIX_FMT_NONE) {
352 av_log(s, AV_LOG_WARNING, "Overriding selected pixel format to use %s instead.\n",
353 av_get_pix_fmt_name(pxl_fmt_spec.ff_id));
357 ctx->pixel_format = pxl_fmt_spec.ff_id;
358 pixel_format = [NSNumber numberWithUnsignedInt:pxl_fmt_spec.avf_id];
359 capture_dict = [NSDictionary dictionaryWithObject:pixel_format
360 forKey:(id)kCVPixelBufferPixelFormatTypeKey];
362 [ctx->video_output setVideoSettings:capture_dict];
363 [ctx->video_output setAlwaysDiscardsLateVideoFrames:YES];
365 ctx->avf_delegate = [[AVFFrameReceiver alloc] initWithContext:ctx];
367 queue = dispatch_queue_create("avf_queue", NULL);
368 [ctx->video_output setSampleBufferDelegate:ctx->avf_delegate queue:queue];
369 dispatch_release(queue);
371 if ([ctx->capture_session canAddOutput:ctx->video_output]) {
372 [ctx->capture_session addOutput:ctx->video_output];
374 av_log(s, AV_LOG_ERROR, "can't add video output to capture session\n");
381 static int add_audio_device(AVFormatContext *s, AVCaptureDevice *audio_device)
383 AVFContext *ctx = (AVFContext*)s->priv_data;
384 NSError *error = nil;
385 AVCaptureDeviceInput* audio_dev_input = [[[AVCaptureDeviceInput alloc] initWithDevice:audio_device error:&error] autorelease];
386 dispatch_queue_t queue;
388 if (!audio_dev_input) {
389 av_log(s, AV_LOG_ERROR, "Failed to create AV capture input device: %s\n",
390 [[error localizedDescription] UTF8String]);
394 if ([ctx->capture_session canAddInput:audio_dev_input]) {
395 [ctx->capture_session addInput:audio_dev_input];
397 av_log(s, AV_LOG_ERROR, "can't add audio input to capture session\n");
402 ctx->audio_output = [[AVCaptureAudioDataOutput alloc] init];
404 if (!ctx->audio_output) {
405 av_log(s, AV_LOG_ERROR, "Failed to init AV audio output\n");
409 ctx->avf_audio_delegate = [[AVFAudioReceiver alloc] initWithContext:ctx];
411 queue = dispatch_queue_create("avf_audio_queue", NULL);
412 [ctx->audio_output setSampleBufferDelegate:ctx->avf_audio_delegate queue:queue];
413 dispatch_release(queue);
415 if ([ctx->capture_session canAddOutput:ctx->audio_output]) {
416 [ctx->capture_session addOutput:ctx->audio_output];
418 av_log(s, AV_LOG_ERROR, "adding audio output to capture session failed\n");
425 static int get_video_config(AVFormatContext *s)
427 AVFContext *ctx = (AVFContext*)s->priv_data;
428 CVImageBufferRef image_buffer;
429 CGSize image_buffer_size;
430 AVStream* stream = avformat_new_stream(s, NULL);
436 // Take stream info from the first frame.
437 while (ctx->frames_captured < 1) {
438 CFRunLoopRunInMode(kCFRunLoopDefaultMode, 0.1, YES);
443 ctx->video_stream_index = stream->index;
445 avpriv_set_pts_info(stream, 64, 1, avf_time_base);
447 image_buffer = CMSampleBufferGetImageBuffer(ctx->current_frame);
448 image_buffer_size = CVImageBufferGetEncodedSize(image_buffer);
450 stream->codec->codec_id = AV_CODEC_ID_RAWVIDEO;
451 stream->codec->codec_type = AVMEDIA_TYPE_VIDEO;
452 stream->codec->width = (int)image_buffer_size.width;
453 stream->codec->height = (int)image_buffer_size.height;
454 stream->codec->pix_fmt = ctx->pixel_format;
456 CFRelease(ctx->current_frame);
457 ctx->current_frame = nil;
464 static int get_audio_config(AVFormatContext *s)
466 AVFContext *ctx = (AVFContext*)s->priv_data;
467 CMFormatDescriptionRef format_desc;
468 AVStream* stream = avformat_new_stream(s, NULL);
474 // Take stream info from the first frame.
475 while (ctx->audio_frames_captured < 1) {
476 CFRunLoopRunInMode(kCFRunLoopDefaultMode, 0.1, YES);
481 ctx->audio_stream_index = stream->index;
483 avpriv_set_pts_info(stream, 64, 1, avf_time_base);
485 format_desc = CMSampleBufferGetFormatDescription(ctx->current_audio_frame);
486 const AudioStreamBasicDescription *basic_desc = CMAudioFormatDescriptionGetStreamBasicDescription(format_desc);
489 av_log(s, AV_LOG_ERROR, "audio format not available\n");
493 stream->codec->codec_type = AVMEDIA_TYPE_AUDIO;
494 stream->codec->sample_rate = basic_desc->mSampleRate;
495 stream->codec->channels = basic_desc->mChannelsPerFrame;
496 stream->codec->channel_layout = av_get_default_channel_layout(stream->codec->channels);
498 ctx->audio_channels = basic_desc->mChannelsPerFrame;
499 ctx->audio_bits_per_sample = basic_desc->mBitsPerChannel;
500 ctx->audio_float = basic_desc->mFormatFlags & kAudioFormatFlagIsFloat;
501 ctx->audio_be = basic_desc->mFormatFlags & kAudioFormatFlagIsBigEndian;
502 ctx->audio_signed_integer = basic_desc->mFormatFlags & kAudioFormatFlagIsSignedInteger;
503 ctx->audio_packed = basic_desc->mFormatFlags & kAudioFormatFlagIsPacked;
504 ctx->audio_non_interleaved = basic_desc->mFormatFlags & kAudioFormatFlagIsNonInterleaved;
506 if (basic_desc->mFormatID == kAudioFormatLinearPCM &&
508 ctx->audio_bits_per_sample == 32 &&
510 stream->codec->codec_id = ctx->audio_be ? AV_CODEC_ID_PCM_F32BE : AV_CODEC_ID_PCM_F32LE;
511 } else if (basic_desc->mFormatID == kAudioFormatLinearPCM &&
512 ctx->audio_signed_integer &&
513 ctx->audio_bits_per_sample == 16 &&
515 stream->codec->codec_id = ctx->audio_be ? AV_CODEC_ID_PCM_S16BE : AV_CODEC_ID_PCM_S16LE;
516 } else if (basic_desc->mFormatID == kAudioFormatLinearPCM &&
517 ctx->audio_signed_integer &&
518 ctx->audio_bits_per_sample == 24 &&
520 stream->codec->codec_id = ctx->audio_be ? AV_CODEC_ID_PCM_S24BE : AV_CODEC_ID_PCM_S24LE;
521 } else if (basic_desc->mFormatID == kAudioFormatLinearPCM &&
522 ctx->audio_signed_integer &&
523 ctx->audio_bits_per_sample == 32 &&
525 stream->codec->codec_id = ctx->audio_be ? AV_CODEC_ID_PCM_S32BE : AV_CODEC_ID_PCM_S32LE;
527 av_log(s, AV_LOG_ERROR, "audio format is not supported\n");
531 if (ctx->audio_non_interleaved) {
532 CMBlockBufferRef block_buffer = CMSampleBufferGetDataBuffer(ctx->current_audio_frame);
533 ctx->audio_buffer_size = CMBlockBufferGetDataLength(block_buffer);
534 ctx->audio_buffer = av_malloc(ctx->audio_buffer_size);
535 if (!ctx->audio_buffer) {
536 av_log(s, AV_LOG_ERROR, "error allocating audio buffer\n");
541 CFRelease(ctx->current_audio_frame);
542 ctx->current_audio_frame = nil;
549 static int avf_read_header(AVFormatContext *s)
551 NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
552 uint32_t num_screens = 0;
553 AVFContext *ctx = (AVFContext*)s->priv_data;
554 AVCaptureDevice *video_device = nil;
555 AVCaptureDevice *audio_device = nil;
556 // Find capture device
557 NSArray *devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeVideo];
558 ctx->num_video_devices = [devices count];
560 ctx->first_pts = av_gettime();
561 ctx->first_audio_pts = av_gettime();
563 pthread_mutex_init(&ctx->frame_lock, NULL);
564 pthread_cond_init(&ctx->frame_wait_cond, NULL);
566 #if !TARGET_OS_IPHONE && __MAC_OS_X_VERSION_MIN_REQUIRED >= 1070
567 CGGetActiveDisplayList(0, NULL, &num_screens);
570 // List devices if requested
571 if (ctx->list_devices) {
573 av_log(ctx, AV_LOG_INFO, "AVFoundation video devices:\n");
574 for (AVCaptureDevice *device in devices) {
575 const char *name = [[device localizedName] UTF8String];
576 index = [devices indexOfObject:device];
577 av_log(ctx, AV_LOG_INFO, "[%d] %s\n", index, name);
580 #if !TARGET_OS_IPHONE && __MAC_OS_X_VERSION_MIN_REQUIRED >= 1070
581 if (num_screens > 0) {
582 CGDirectDisplayID screens[num_screens];
583 CGGetActiveDisplayList(num_screens, screens, &num_screens);
584 for (int i = 0; i < num_screens; i++) {
585 av_log(ctx, AV_LOG_INFO, "[%d] Capture screen %d\n", index + i, i);
590 av_log(ctx, AV_LOG_INFO, "AVFoundation audio devices:\n");
591 devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeAudio];
592 for (AVCaptureDevice *device in devices) {
593 const char *name = [[device localizedName] UTF8String];
594 int index = [devices indexOfObject:device];
595 av_log(ctx, AV_LOG_INFO, "[%d] %s\n", index, name);
600 // parse input filename for video and audio device
601 parse_device_name(s);
603 // check for device index given in filename
604 if (ctx->video_device_index == -1 && ctx->video_filename) {
605 sscanf(ctx->video_filename, "%d", &ctx->video_device_index);
607 if (ctx->audio_device_index == -1 && ctx->audio_filename) {
608 sscanf(ctx->audio_filename, "%d", &ctx->audio_device_index);
611 if (ctx->video_device_index >= 0) {
612 if (ctx->video_device_index < ctx->num_video_devices) {
613 video_device = [devices objectAtIndex:ctx->video_device_index];
614 } else if (ctx->video_device_index < ctx->num_video_devices + num_screens) {
615 #if !TARGET_OS_IPHONE && __MAC_OS_X_VERSION_MIN_REQUIRED >= 1070
616 CGDirectDisplayID screens[num_screens];
617 CGGetActiveDisplayList(num_screens, screens, &num_screens);
618 AVCaptureScreenInput* capture_screen_input = [[[AVCaptureScreenInput alloc] initWithDisplayID:screens[ctx->video_device_index - ctx->num_video_devices]] autorelease];
619 video_device = (AVCaptureDevice*) capture_screen_input;
622 av_log(ctx, AV_LOG_ERROR, "Invalid device index\n");
625 } else if (ctx->video_filename &&
626 strncmp(ctx->video_filename, "none", 4)) {
627 if (!strncmp(ctx->video_filename, "default", 7)) {
628 video_device = [AVCaptureDevice defaultDeviceWithMediaType:AVMediaTypeVideo];
630 // looking for video inputs
631 for (AVCaptureDevice *device in devices) {
632 if (!strncmp(ctx->video_filename, [[device localizedName] UTF8String], strlen(ctx->video_filename))) {
633 video_device = device;
638 #if !TARGET_OS_IPHONE && __MAC_OS_X_VERSION_MIN_REQUIRED >= 1070
639 // looking for screen inputs
642 if(sscanf(ctx->video_filename, "Capture screen %d", &idx) && idx < num_screens) {
643 CGDirectDisplayID screens[num_screens];
644 CGGetActiveDisplayList(num_screens, screens, &num_screens);
645 AVCaptureScreenInput* capture_screen_input = [[[AVCaptureScreenInput alloc] initWithDisplayID:screens[idx]] autorelease];
646 video_device = (AVCaptureDevice*) capture_screen_input;
647 ctx->video_device_index = ctx->num_video_devices + idx;
654 av_log(ctx, AV_LOG_ERROR, "Video device not found\n");
660 if (ctx->audio_device_index >= 0) {
661 NSArray *devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeAudio];
663 if (ctx->audio_device_index >= [devices count]) {
664 av_log(ctx, AV_LOG_ERROR, "Invalid audio device index\n");
668 audio_device = [devices objectAtIndex:ctx->audio_device_index];
669 } else if (ctx->audio_filename &&
670 strncmp(ctx->audio_filename, "none", 4)) {
671 if (!strncmp(ctx->audio_filename, "default", 7)) {
672 audio_device = [AVCaptureDevice defaultDeviceWithMediaType:AVMediaTypeAudio];
674 NSArray *devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeAudio];
676 for (AVCaptureDevice *device in devices) {
677 if (!strncmp(ctx->audio_filename, [[device localizedName] UTF8String], strlen(ctx->audio_filename))) {
678 audio_device = device;
685 av_log(ctx, AV_LOG_ERROR, "Audio device not found\n");
690 // Video nor Audio capture device not found, looking for AVMediaTypeVideo/Audio
691 if (!video_device && !audio_device) {
692 av_log(s, AV_LOG_ERROR, "No AV capture device found\n");
697 if (ctx->video_device_index < ctx->num_video_devices) {
698 av_log(s, AV_LOG_DEBUG, "'%s' opened\n", [[video_device localizedName] UTF8String]);
700 av_log(s, AV_LOG_DEBUG, "'%s' opened\n", [[video_device description] UTF8String]);
704 av_log(s, AV_LOG_DEBUG, "audio device '%s' opened\n", [[audio_device localizedName] UTF8String]);
707 // Initialize capture session
708 ctx->capture_session = [[AVCaptureSession alloc] init];
710 if (video_device && add_video_device(s, video_device)) {
713 if (audio_device && add_audio_device(s, audio_device)) {
716 [ctx->capture_session startRunning];
718 if (video_device && get_video_config(s)) {
723 if (audio_device && get_audio_config(s)) {
732 destroy_context(ctx);
736 static int avf_read_packet(AVFormatContext *s, AVPacket *pkt)
738 AVFContext* ctx = (AVFContext*)s->priv_data;
741 CVImageBufferRef image_buffer;
744 image_buffer = CMSampleBufferGetImageBuffer(ctx->current_frame);
746 if (ctx->current_frame != nil) {
748 if (av_new_packet(pkt, (int)CVPixelBufferGetDataSize(image_buffer)) < 0) {
752 pkt->pts = pkt->dts = av_rescale_q(av_gettime() - ctx->first_pts,
755 pkt->stream_index = ctx->video_stream_index;
756 pkt->flags |= AV_PKT_FLAG_KEY;
758 CVPixelBufferLockBaseAddress(image_buffer, 0);
760 data = CVPixelBufferGetBaseAddress(image_buffer);
761 memcpy(pkt->data, data, pkt->size);
763 CVPixelBufferUnlockBaseAddress(image_buffer, 0);
764 CFRelease(ctx->current_frame);
765 ctx->current_frame = nil;
766 } else if (ctx->current_audio_frame != nil) {
767 CMBlockBufferRef block_buffer = CMSampleBufferGetDataBuffer(ctx->current_audio_frame);
768 int block_buffer_size = CMBlockBufferGetDataLength(block_buffer);
770 if (!block_buffer || !block_buffer_size) {
774 if (ctx->audio_non_interleaved && block_buffer_size > ctx->audio_buffer_size) {
775 return AVERROR_BUFFER_TOO_SMALL;
778 if (av_new_packet(pkt, block_buffer_size) < 0) {
782 pkt->pts = pkt->dts = av_rescale_q(av_gettime() - ctx->first_audio_pts,
786 pkt->stream_index = ctx->audio_stream_index;
787 pkt->flags |= AV_PKT_FLAG_KEY;
789 if (ctx->audio_non_interleaved) {
790 int sample, c, shift, num_samples;
792 OSStatus ret = CMBlockBufferCopyDataBytes(block_buffer, 0, pkt->size, ctx->audio_buffer);
793 if (ret != kCMBlockBufferNoErr) {
797 num_samples = pkt->size / (ctx->audio_channels * (ctx->audio_bits_per_sample >> 3));
799 // transform decoded frame into output format
800 #define INTERLEAVE_OUTPUT(bps) \
802 int##bps##_t **src; \
803 int##bps##_t *dest; \
804 src = av_malloc(ctx->audio_channels * sizeof(int##bps##_t*)); \
805 if (!src) return AVERROR(EIO); \
806 for (c = 0; c < ctx->audio_channels; c++) { \
807 src[c] = ((int##bps##_t*)ctx->audio_buffer) + c * num_samples; \
809 dest = (int##bps##_t*)pkt->data; \
810 shift = bps - ctx->audio_bits_per_sample; \
811 for (sample = 0; sample < num_samples; sample++) \
812 for (c = 0; c < ctx->audio_channels; c++) \
813 *dest++ = src[c][sample] << shift; \
817 if (ctx->audio_bits_per_sample <= 16) {
818 INTERLEAVE_OUTPUT(16)
820 INTERLEAVE_OUTPUT(32)
823 OSStatus ret = CMBlockBufferCopyDataBytes(block_buffer, 0, pkt->size, pkt->data);
824 if (ret != kCMBlockBufferNoErr) {
829 CFRelease(ctx->current_audio_frame);
830 ctx->current_audio_frame = nil;
833 pthread_cond_wait(&ctx->frame_wait_cond, &ctx->frame_lock);
837 } while (!pkt->data);
842 static int avf_close(AVFormatContext *s)
844 AVFContext* ctx = (AVFContext*)s->priv_data;
845 destroy_context(ctx);
849 static const AVOption options[] = {
850 { "list_devices", "list available devices", offsetof(AVFContext, list_devices), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, AV_OPT_FLAG_DECODING_PARAM, "list_devices" },
851 { "true", "", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, AV_OPT_FLAG_DECODING_PARAM, "list_devices" },
852 { "false", "", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, AV_OPT_FLAG_DECODING_PARAM, "list_devices" },
853 { "video_device_index", "select video device by index for devices with same name (starts at 0)", offsetof(AVFContext, video_device_index), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, AV_OPT_FLAG_DECODING_PARAM },
854 { "audio_device_index", "select audio device by index for devices with same name (starts at 0)", offsetof(AVFContext, audio_device_index), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, AV_OPT_FLAG_DECODING_PARAM },
855 { "pixel_format", "set pixel format", offsetof(AVFContext, pixel_format), AV_OPT_TYPE_PIXEL_FMT, {.i64 = AV_PIX_FMT_YUV420P}, 0, INT_MAX, AV_OPT_FLAG_DECODING_PARAM},
859 static const AVClass avf_class = {
860 .class_name = "AVFoundation input device",
861 .item_name = av_default_item_name,
863 .version = LIBAVUTIL_VERSION_INT,
864 .category = AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT,
867 AVInputFormat ff_avfoundation_demuxer = {
868 .name = "avfoundation",
869 .long_name = NULL_IF_CONFIG_SMALL("AVFoundation input device"),
870 .priv_data_size = sizeof(AVFContext),
871 .read_header = avf_read_header,
872 .read_packet = avf_read_packet,
873 .read_close = avf_close,
874 .flags = AVFMT_NOFILE,
875 .priv_class = &avf_class,