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 "libavformat/internal.h"
34 #include "libavutil/internal.h"
35 #include "libavutil/time.h"
38 static const int avf_time_base = 1000000;
40 static const AVRational avf_time_base_q = {
45 struct AVFPixelFormatSpec {
46 enum AVPixelFormat ff_id;
50 static const struct AVFPixelFormatSpec avf_pixel_formats[] = {
51 { AV_PIX_FMT_MONOBLACK, kCVPixelFormatType_1Monochrome },
52 { AV_PIX_FMT_RGB555BE, kCVPixelFormatType_16BE555 },
53 { AV_PIX_FMT_RGB555LE, kCVPixelFormatType_16LE555 },
54 { AV_PIX_FMT_RGB565BE, kCVPixelFormatType_16BE565 },
55 { AV_PIX_FMT_RGB565LE, kCVPixelFormatType_16LE565 },
56 { AV_PIX_FMT_RGB24, kCVPixelFormatType_24RGB },
57 { AV_PIX_FMT_BGR24, kCVPixelFormatType_24BGR },
58 { AV_PIX_FMT_0RGB, kCVPixelFormatType_32ARGB },
59 { AV_PIX_FMT_BGR0, kCVPixelFormatType_32BGRA },
60 { AV_PIX_FMT_0BGR, kCVPixelFormatType_32ABGR },
61 { AV_PIX_FMT_RGB0, kCVPixelFormatType_32RGBA },
62 { AV_PIX_FMT_BGR48BE, kCVPixelFormatType_48RGB },
63 { AV_PIX_FMT_UYVY422, kCVPixelFormatType_422YpCbCr8 },
64 { AV_PIX_FMT_YUVA444P, kCVPixelFormatType_4444YpCbCrA8R },
65 { AV_PIX_FMT_YUVA444P16LE, kCVPixelFormatType_4444AYpCbCr16 },
66 { AV_PIX_FMT_YUV444P, kCVPixelFormatType_444YpCbCr8 },
67 { AV_PIX_FMT_YUV422P16, kCVPixelFormatType_422YpCbCr16 },
68 { AV_PIX_FMT_YUV422P10, kCVPixelFormatType_422YpCbCr10 },
69 { AV_PIX_FMT_YUV444P10, kCVPixelFormatType_444YpCbCr10 },
70 { AV_PIX_FMT_YUV420P, kCVPixelFormatType_420YpCbCr8Planar },
71 { AV_PIX_FMT_NV12, kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange },
72 { AV_PIX_FMT_YUYV422, kCVPixelFormatType_422YpCbCr8_yuvs },
73 #if __MAC_OS_X_VERSION_MIN_REQUIRED >= 1080
74 { AV_PIX_FMT_GRAY8, kCVPixelFormatType_OneComponent8 },
76 { 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);
258 ctx->video_filename = strtok(tmp, ":");
259 ctx->audio_filename = strtok(NULL, ":");
261 ctx->audio_filename = strtok(tmp, ":");
265 static int add_video_device(AVFormatContext *s, AVCaptureDevice *video_device)
267 AVFContext *ctx = (AVFContext*)s->priv_data;
268 NSError *error = nil;
269 AVCaptureInput* capture_input = nil;
271 if (ctx->video_device_index < ctx->num_video_devices) {
272 capture_input = (AVCaptureInput*) [[[AVCaptureDeviceInput alloc] initWithDevice:video_device error:&error] autorelease];
274 capture_input = (AVCaptureInput*) video_device;
277 if (!capture_input) {
278 av_log(s, AV_LOG_ERROR, "Failed to create AV capture input device: %s\n",
279 [[error localizedDescription] UTF8String]);
283 if ([ctx->capture_session canAddInput:capture_input]) {
284 [ctx->capture_session addInput:capture_input];
286 av_log(s, AV_LOG_ERROR, "can't add video input to capture session\n");
291 ctx->video_output = [[AVCaptureVideoDataOutput alloc] init];
293 if (!ctx->video_output) {
294 av_log(s, AV_LOG_ERROR, "Failed to init AV video output\n");
298 // select pixel format
299 struct AVFPixelFormatSpec pxl_fmt_spec;
300 pxl_fmt_spec.ff_id = AV_PIX_FMT_NONE;
302 for (int i = 0; avf_pixel_formats[i].ff_id != AV_PIX_FMT_NONE; i++) {
303 if (ctx->pixel_format == avf_pixel_formats[i].ff_id) {
304 pxl_fmt_spec = avf_pixel_formats[i];
309 // check if selected pixel format is supported by AVFoundation
310 if (pxl_fmt_spec.ff_id == AV_PIX_FMT_NONE) {
311 av_log(s, AV_LOG_ERROR, "Selected pixel format (%s) is not supported by AVFoundation.\n",
312 av_get_pix_fmt_name(pxl_fmt_spec.ff_id));
316 // check if the pixel format is available for this device
317 if ([[ctx->video_output availableVideoCVPixelFormatTypes] indexOfObject:[NSNumber numberWithInt:pxl_fmt_spec.avf_id]] == NSNotFound) {
318 av_log(s, AV_LOG_ERROR, "Selected pixel format (%s) is not supported by the input device.\n",
319 av_get_pix_fmt_name(pxl_fmt_spec.ff_id));
321 pxl_fmt_spec.ff_id = AV_PIX_FMT_NONE;
323 av_log(s, AV_LOG_ERROR, "Supported pixel formats:\n");
324 for (NSNumber *pxl_fmt in [ctx->video_output availableVideoCVPixelFormatTypes]) {
325 struct AVFPixelFormatSpec pxl_fmt_dummy;
326 pxl_fmt_dummy.ff_id = AV_PIX_FMT_NONE;
327 for (int i = 0; avf_pixel_formats[i].ff_id != AV_PIX_FMT_NONE; i++) {
328 if ([pxl_fmt intValue] == avf_pixel_formats[i].avf_id) {
329 pxl_fmt_dummy = avf_pixel_formats[i];
334 if (pxl_fmt_dummy.ff_id != AV_PIX_FMT_NONE) {
335 av_log(s, AV_LOG_ERROR, " %s\n", av_get_pix_fmt_name(pxl_fmt_dummy.ff_id));
337 // select first supported pixel format instead of user selected (or default) pixel format
338 if (pxl_fmt_spec.ff_id == AV_PIX_FMT_NONE) {
339 pxl_fmt_spec = pxl_fmt_dummy;
344 // fail if there is no appropriate pixel format or print a warning about overriding the pixel format
345 if (pxl_fmt_spec.ff_id == AV_PIX_FMT_NONE) {
348 av_log(s, AV_LOG_WARNING, "Overriding selected pixel format to use %s instead.\n",
349 av_get_pix_fmt_name(pxl_fmt_spec.ff_id));
353 ctx->pixel_format = pxl_fmt_spec.ff_id;
354 NSNumber *pixel_format = [NSNumber numberWithUnsignedInt:pxl_fmt_spec.avf_id];
355 NSDictionary *capture_dict = [NSDictionary dictionaryWithObject:pixel_format
356 forKey:(id)kCVPixelBufferPixelFormatTypeKey];
358 [ctx->video_output setVideoSettings:capture_dict];
359 [ctx->video_output setAlwaysDiscardsLateVideoFrames:YES];
361 ctx->avf_delegate = [[AVFFrameReceiver alloc] initWithContext:ctx];
363 dispatch_queue_t queue = dispatch_queue_create("avf_queue", NULL);
364 [ctx->video_output setSampleBufferDelegate:ctx->avf_delegate queue:queue];
365 dispatch_release(queue);
367 if ([ctx->capture_session canAddOutput:ctx->video_output]) {
368 [ctx->capture_session addOutput:ctx->video_output];
370 av_log(s, AV_LOG_ERROR, "can't add video output to capture session\n");
377 static int add_audio_device(AVFormatContext *s, AVCaptureDevice *audio_device)
379 AVFContext *ctx = (AVFContext*)s->priv_data;
380 NSError *error = nil;
381 AVCaptureDeviceInput* audio_dev_input = [[[AVCaptureDeviceInput alloc] initWithDevice:audio_device error:&error] autorelease];
383 if (!audio_dev_input) {
384 av_log(s, AV_LOG_ERROR, "Failed to create AV capture input device: %s\n",
385 [[error localizedDescription] UTF8String]);
389 if ([ctx->capture_session canAddInput:audio_dev_input]) {
390 [ctx->capture_session addInput:audio_dev_input];
392 av_log(s, AV_LOG_ERROR, "can't add audio input to capture session\n");
397 ctx->audio_output = [[AVCaptureAudioDataOutput alloc] init];
399 if (!ctx->audio_output) {
400 av_log(s, AV_LOG_ERROR, "Failed to init AV audio output\n");
404 ctx->avf_audio_delegate = [[AVFAudioReceiver alloc] initWithContext:ctx];
406 dispatch_queue_t queue = dispatch_queue_create("avf_audio_queue", NULL);
407 [ctx->audio_output setSampleBufferDelegate:ctx->avf_audio_delegate queue:queue];
408 dispatch_release(queue);
410 if ([ctx->capture_session canAddOutput:ctx->audio_output]) {
411 [ctx->capture_session addOutput:ctx->audio_output];
413 av_log(s, AV_LOG_ERROR, "adding audio output to capture session failed\n");
420 static int get_video_config(AVFormatContext *s)
422 AVFContext *ctx = (AVFContext*)s->priv_data;
424 // Take stream info from the first frame.
425 while (ctx->frames_captured < 1) {
426 CFRunLoopRunInMode(kCFRunLoopDefaultMode, 0.1, YES);
431 AVStream* stream = avformat_new_stream(s, NULL);
437 ctx->video_stream_index = stream->index;
439 avpriv_set_pts_info(stream, 64, 1, avf_time_base);
441 CVImageBufferRef image_buffer = CMSampleBufferGetImageBuffer(ctx->current_frame);
442 CGSize image_buffer_size = CVImageBufferGetEncodedSize(image_buffer);
444 stream->codec->codec_id = AV_CODEC_ID_RAWVIDEO;
445 stream->codec->codec_type = AVMEDIA_TYPE_VIDEO;
446 stream->codec->width = (int)image_buffer_size.width;
447 stream->codec->height = (int)image_buffer_size.height;
448 stream->codec->pix_fmt = ctx->pixel_format;
450 CFRelease(ctx->current_frame);
451 ctx->current_frame = nil;
458 static int get_audio_config(AVFormatContext *s)
460 AVFContext *ctx = (AVFContext*)s->priv_data;
462 // Take stream info from the first frame.
463 while (ctx->audio_frames_captured < 1) {
464 CFRunLoopRunInMode(kCFRunLoopDefaultMode, 0.1, YES);
469 AVStream* stream = avformat_new_stream(s, NULL);
475 ctx->audio_stream_index = stream->index;
477 avpriv_set_pts_info(stream, 64, 1, avf_time_base);
479 CMFormatDescriptionRef format_desc = CMSampleBufferGetFormatDescription(ctx->current_audio_frame);
480 const AudioStreamBasicDescription *basic_desc = CMAudioFormatDescriptionGetStreamBasicDescription(format_desc);
483 av_log(s, AV_LOG_ERROR, "audio format not available\n");
487 stream->codec->codec_type = AVMEDIA_TYPE_AUDIO;
488 stream->codec->sample_rate = basic_desc->mSampleRate;
489 stream->codec->channels = basic_desc->mChannelsPerFrame;
490 stream->codec->channel_layout = av_get_default_channel_layout(stream->codec->channels);
492 ctx->audio_channels = basic_desc->mChannelsPerFrame;
493 ctx->audio_bits_per_sample = basic_desc->mBitsPerChannel;
494 ctx->audio_float = basic_desc->mFormatFlags & kAudioFormatFlagIsFloat;
495 ctx->audio_be = basic_desc->mFormatFlags & kAudioFormatFlagIsBigEndian;
496 ctx->audio_signed_integer = basic_desc->mFormatFlags & kAudioFormatFlagIsSignedInteger;
497 ctx->audio_packed = basic_desc->mFormatFlags & kAudioFormatFlagIsPacked;
498 ctx->audio_non_interleaved = basic_desc->mFormatFlags & kAudioFormatFlagIsNonInterleaved;
500 if (basic_desc->mFormatID == kAudioFormatLinearPCM &&
503 stream->codec->codec_id = ctx->audio_be ? AV_CODEC_ID_PCM_F32BE : AV_CODEC_ID_PCM_F32LE;
505 av_log(s, AV_LOG_ERROR, "audio format is not supported\n");
509 if (ctx->audio_non_interleaved) {
510 CMBlockBufferRef block_buffer = CMSampleBufferGetDataBuffer(ctx->current_audio_frame);
511 ctx->audio_buffer_size = CMBlockBufferGetDataLength(block_buffer);
512 ctx->audio_buffer = av_malloc(ctx->audio_buffer_size);
513 if (!ctx->audio_buffer) {
514 av_log(s, AV_LOG_ERROR, "error allocating audio buffer\n");
519 CFRelease(ctx->current_audio_frame);
520 ctx->current_audio_frame = nil;
527 static int avf_read_header(AVFormatContext *s)
529 NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
530 AVFContext *ctx = (AVFContext*)s->priv_data;
531 ctx->first_pts = av_gettime();
532 ctx->first_audio_pts = av_gettime();
533 uint32_t num_screens = 0;
535 pthread_mutex_init(&ctx->frame_lock, NULL);
536 pthread_cond_init(&ctx->frame_wait_cond, NULL);
538 CGGetActiveDisplayList(0, NULL, &num_screens);
540 // List devices if requested
541 if (ctx->list_devices) {
542 av_log(ctx, AV_LOG_INFO, "AVFoundation video devices:\n");
543 NSArray *devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeVideo];
545 for (AVCaptureDevice *device in devices) {
546 const char *name = [[device localizedName] UTF8String];
547 index = [devices indexOfObject:device];
548 av_log(ctx, AV_LOG_INFO, "[%d] %s\n", index, name);
551 if (num_screens > 0) {
552 CGDirectDisplayID screens[num_screens];
553 CGGetActiveDisplayList(num_screens, screens, &num_screens);
554 for (int i = 0; i < num_screens; i++) {
555 av_log(ctx, AV_LOG_INFO, "[%d] Capture screen %d\n", index + i, i);
559 av_log(ctx, AV_LOG_INFO, "AVFoundation audio devices:\n");
560 devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeAudio];
561 for (AVCaptureDevice *device in devices) {
562 const char *name = [[device localizedName] UTF8String];
563 int index = [devices indexOfObject:device];
564 av_log(ctx, AV_LOG_INFO, "[%d] %s\n", index, name);
569 // Find capture device
570 AVCaptureDevice *video_device = nil;
571 AVCaptureDevice *audio_device = nil;
573 NSArray *video_devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeVideo];
574 ctx->num_video_devices = [video_devices count];
576 // parse input filename for video and audio device
577 parse_device_name(s);
579 // check for device index given in filename
580 if (ctx->video_device_index == -1 && ctx->video_filename) {
581 sscanf(ctx->video_filename, "%d", &ctx->video_device_index);
583 if (ctx->audio_device_index == -1 && ctx->audio_filename) {
584 sscanf(ctx->audio_filename, "%d", &ctx->audio_device_index);
587 if (ctx->video_device_index >= 0) {
588 if (ctx->video_device_index < ctx->num_video_devices) {
589 video_device = [video_devices objectAtIndex:ctx->video_device_index];
590 } else if (ctx->video_device_index < ctx->num_video_devices + num_screens) {
591 CGDirectDisplayID screens[num_screens];
592 CGGetActiveDisplayList(num_screens, screens, &num_screens);
593 AVCaptureScreenInput* capture_screen_input = [[[AVCaptureScreenInput alloc] initWithDisplayID:screens[ctx->video_device_index - ctx->num_video_devices]] autorelease];
594 video_device = (AVCaptureDevice*) capture_screen_input;
596 av_log(ctx, AV_LOG_ERROR, "Invalid device index\n");
599 } else if (ctx->video_filename &&
600 strncmp(ctx->video_filename, "default", 7)) {
601 // looking for video inputs
602 for (AVCaptureDevice *device in video_devices) {
603 if (!strncmp(ctx->video_filename, [[device localizedName] UTF8String], strlen(ctx->video_filename))) {
604 video_device = device;
609 // looking for screen inputs
612 if(sscanf(ctx->video_filename, "Capture screen %d", &idx) && idx < num_screens) {
613 CGDirectDisplayID screens[num_screens];
614 CGGetActiveDisplayList(num_screens, screens, &num_screens);
615 AVCaptureScreenInput* capture_screen_input = [[[AVCaptureScreenInput alloc] initWithDisplayID:screens[idx]] autorelease];
616 video_device = (AVCaptureDevice*) capture_screen_input;
617 ctx->video_device_index = ctx->num_video_devices + idx;
622 av_log(ctx, AV_LOG_ERROR, "Video device not found\n");
626 video_device = [AVCaptureDevice defaultDeviceWithMediaType:AVMediaTypeVideo];
630 if (ctx->audio_device_index >= 0) {
631 NSArray *devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeAudio];
633 if (ctx->audio_device_index >= [devices count]) {
634 av_log(ctx, AV_LOG_ERROR, "Invalid audio device index\n");
638 audio_device = [devices objectAtIndex:ctx->audio_device_index];
639 } else if (ctx->audio_filename &&
640 strncmp(ctx->audio_filename, "default", 7)) {
641 NSArray *devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeAudio];
643 for (AVCaptureDevice *device in devices) {
644 if (!strncmp(ctx->audio_filename, [[device localizedName] UTF8String], strlen(ctx->audio_filename))) {
645 audio_device = device;
651 av_log(ctx, AV_LOG_ERROR, "Audio device not found\n");
655 audio_device = [AVCaptureDevice defaultDeviceWithMediaType:AVMediaTypeAudio];
658 // Video nor Audio capture device not found, looking for AVMediaTypeVideo/Audio
659 if (!video_device && !audio_device) {
660 av_log(s, AV_LOG_ERROR, "No AV capture device found\n");
665 if (ctx->video_device_index < ctx->num_video_devices) {
666 av_log(s, AV_LOG_DEBUG, "'%s' opened\n", [[video_device localizedName] UTF8String]);
668 av_log(s, AV_LOG_DEBUG, "'%s' opened\n", [[video_device description] UTF8String]);
672 av_log(s, AV_LOG_DEBUG, "audio device '%s' opened\n", [[audio_device localizedName] UTF8String]);
675 // Initialize capture session
676 ctx->capture_session = [[AVCaptureSession alloc] init];
678 if (video_device && add_video_device(s, video_device)) {
681 if (audio_device && add_audio_device(s, audio_device)) {
684 [ctx->capture_session startRunning];
686 if (video_device && get_video_config(s)) {
691 if (audio_device && get_audio_config(s)) {
700 destroy_context(ctx);
704 static int avf_read_packet(AVFormatContext *s, AVPacket *pkt)
706 AVFContext* ctx = (AVFContext*)s->priv_data;
711 CVImageBufferRef image_buffer = CMSampleBufferGetImageBuffer(ctx->current_frame);
713 if (ctx->current_frame != nil) {
714 if (av_new_packet(pkt, (int)CVPixelBufferGetDataSize(image_buffer)) < 0) {
718 pkt->pts = pkt->dts = av_rescale_q(av_gettime() - ctx->first_pts,
721 pkt->stream_index = ctx->video_stream_index;
722 pkt->flags |= AV_PKT_FLAG_KEY;
724 CVPixelBufferLockBaseAddress(image_buffer, 0);
726 void* data = CVPixelBufferGetBaseAddress(image_buffer);
727 memcpy(pkt->data, data, pkt->size);
729 CVPixelBufferUnlockBaseAddress(image_buffer, 0);
730 CFRelease(ctx->current_frame);
731 ctx->current_frame = nil;
732 } else if (ctx->current_audio_frame != nil) {
733 CMBlockBufferRef block_buffer = CMSampleBufferGetDataBuffer(ctx->current_audio_frame);
734 int block_buffer_size = CMBlockBufferGetDataLength(block_buffer);
736 if (!block_buffer || !block_buffer_size) {
740 if (ctx->audio_non_interleaved && block_buffer_size > ctx->audio_buffer_size) {
741 return AVERROR_BUFFER_TOO_SMALL;
744 if (av_new_packet(pkt, block_buffer_size) < 0) {
748 pkt->pts = pkt->dts = av_rescale_q(av_gettime() - ctx->first_audio_pts,
752 pkt->stream_index = ctx->audio_stream_index;
753 pkt->flags |= AV_PKT_FLAG_KEY;
755 if (ctx->audio_non_interleaved) {
756 int sample, c, shift;
758 OSStatus ret = CMBlockBufferCopyDataBytes(block_buffer, 0, pkt->size, ctx->audio_buffer);
759 if (ret != kCMBlockBufferNoErr) {
763 int num_samples = pkt->size / (ctx->audio_channels * (ctx->audio_bits_per_sample >> 3));
765 // transform decoded frame into output format
766 #define INTERLEAVE_OUTPUT(bps) \
768 int##bps##_t **src; \
769 int##bps##_t *dest; \
770 src = av_malloc(ctx->audio_channels * sizeof(int##bps##_t*)); \
771 if (!src) return AVERROR(EIO); \
772 for (c = 0; c < ctx->audio_channels; c++) { \
773 src[c] = ((int##bps##_t*)ctx->audio_buffer) + c * num_samples; \
775 dest = (int##bps##_t*)pkt->data; \
776 shift = bps - ctx->audio_bits_per_sample; \
777 for (sample = 0; sample < num_samples; sample++) \
778 for (c = 0; c < ctx->audio_channels; c++) \
779 *dest++ = src[c][sample] << shift; \
783 if (ctx->audio_bits_per_sample <= 16) {
784 INTERLEAVE_OUTPUT(16)
786 INTERLEAVE_OUTPUT(32)
789 OSStatus ret = CMBlockBufferCopyDataBytes(block_buffer, 0, pkt->size, pkt->data);
790 if (ret != kCMBlockBufferNoErr) {
795 CFRelease(ctx->current_audio_frame);
796 ctx->current_audio_frame = nil;
799 pthread_cond_wait(&ctx->frame_wait_cond, &ctx->frame_lock);
803 } while (!pkt->data);
808 static int avf_close(AVFormatContext *s)
810 AVFContext* ctx = (AVFContext*)s->priv_data;
811 destroy_context(ctx);
815 static const AVOption options[] = {
816 { "frame_rate", "set frame rate", offsetof(AVFContext, frame_rate), AV_OPT_TYPE_FLOAT, { .dbl = 30.0 }, 0.1, 30.0, AV_OPT_TYPE_VIDEO_RATE, NULL },
817 { "list_devices", "list available devices", offsetof(AVFContext, list_devices), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, AV_OPT_FLAG_DECODING_PARAM, "list_devices" },
818 { "true", "", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, AV_OPT_FLAG_DECODING_PARAM, "list_devices" },
819 { "false", "", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, AV_OPT_FLAG_DECODING_PARAM, "list_devices" },
820 { "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 },
821 { "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 },
822 { "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},
826 static const AVClass avf_class = {
827 .class_name = "AVFoundation input device",
828 .item_name = av_default_item_name,
830 .version = LIBAVUTIL_VERSION_INT,
831 .category = AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT,
834 AVInputFormat ff_avfoundation_demuxer = {
835 .name = "avfoundation",
836 .long_name = NULL_IF_CONFIG_SMALL("AVFoundation input device"),
837 .priv_data_size = sizeof(AVFContext),
838 .read_header = avf_read_header,
839 .read_packet = avf_read_packet,
840 .read_close = avf_close,
841 .flags = AVFMT_NOFILE,
842 .priv_class = &avf_class,