一、需求描述
在和固件設(shè)備進(jìn)行P2P通話時(shí),想錄制某一段音視頻,但是在官方的webrtc中,只能拿到一幀一幀視頻原始數(shù)據(jù),拿不到音頻數(shù)據(jù),故需要實(shí)現(xiàn)音視頻錄制必須修改webrtc源碼回調(diào)音視頻原始數(shù)據(jù).
二、webrtc源碼新增音頻回調(diào)
代碼在sdk/objc/native/src/audio/audio_device_ios.mm中,找到OnGetPlayoutData方法,增加代碼:
AudioBuffer* audio_buffer = &io_data->mBuffers[0];
??RTC_OBJC_TYPE(RTCTAudioData) *audioData = [[RTC_OBJC_TYPE(RTCTAudioData) alloc] init];
??audioData.audioBufferList = *io_data;
??audioData.timeStamp = *time_stamp;
??audioData.busNumber = bus_number;
??audioData.numFrames = num_frames;
??RTC_OBJC_TYPE(RTCAudioSession)* session = [RTC_OBJC_TYPE(RTCAudioSession) sharedInstance];
??[session lockForConfiguration];
??if?(session) {
?session.audioData = audioData;
??}
??[session unlockForConfiguration];
??RTC_DCHECK_EQ(1, audio_buffer->mNumberChannels);
音頻就通過RTCAudioSession回調(diào)到業(yè)務(wù)層去了,再使用逐幀編碼寫入到文件.
三、音視頻幀寫入文件
業(yè)務(wù)層監(jiān)聽webrtc音頻流及視頻流
[[RTCAudioSession sharedInstance] addObserver:self forKeyPath:@"audioData"?options:NSKeyValueObservingOptionNew context:@"RTCAudioData"];
[remoteRederer addObserver:self forKeyPath:@"RTCFrame"?options:NSKeyValueObservingOptionNew context:@"RTCVideoFrame"];
監(jiān)聽到流處理
if?(context == @"RTCAudioData") {
????RTCTAudioData *audioData = change[@"new"];
????dispatch_async(dispatch_get_main_queue(), ^{
????????if?(self.delegate && [self.delegate respondsToSelector:@selector(webRTCClientAudioStream:audioData:)]) {
????????????[self.delegate webRTCClientAudioStream:self audioData:audioData];
????????}
????});
}?else?if?(context == @"RTCVideoFrame") {
????RTCVideoFrame *frame = change[@"new"];
????dispatch_async(dispatch_get_main_queue(), ^{
????????if?(frame && self.delegate && [self.delegate respondsToSelector:@selector(webRTCClientVideoStream:videoFrame:)]) {
????????????[self.delegate webRTCClientVideoStream:self videoFrame:frame];
????????}
????});
}
視頻寫入? ?
if?(videoFrame) {
?RTCCVPixelBuffer *buff = videoFrame.buffer;
?if?(buff) {
?CVPixelBufferRef buffRef = buff.pixelBuffer;
?if?(buffRef) {
?if?(self.startRecordPts ==?0) {
?self.startRecordPts = videoFrame.timeStampNs;
?}
?if?(self.writeManager) {
?if?(!self.writeManager.canWrite) {
?self.writeManager.outputSize = CGSizeMake(buff.width, buff.height);
?[self.writeManager startWrite];
?}
?int64_t pts = videoFrame.timeStampNs - self.startRecordPts;
?[self.writeManager appendCVPixelBuffer:buffRef pts:pts /?1000];
?}
?}
?}
?}
音頻寫入
int?sampleRate = [RTCAudioSession sharedInstance].sampleRate;
AudioBufferList buffList = audioData.audioBufferList;
int?channels = buffList.mBuffers[0].mNumberChannels;
int?len = buffList.mBuffers[0].mDataByteSize;
NSLog(@"%d", channels);
NSLog(@"%d", len);
AudioStreamBasicDescription asbd;
asbd.mSampleRate = sampleRate;
asbd.mFormatID = kAudioFormatLinearPCM;
asbd.mFormatFlags = kLinearPCMFormatFlagIsPacked | kLinearPCMFormatFlagIsSignedInteger;
asbd.mChannelsPerFrame = channels;
asbd.mBitsPerChannel =?16;
asbd.mFramesPerPacket =?1;
asbd.mBytesPerFrame = asbd.mBitsPerChannel /?8?* asbd.mChannelsPerFrame;
asbd.mBytesPerPacket = asbd.mBytesPerFrame * asbd.mFramesPerPacket;
OSStatus error =?0;
static?CMFormatDescriptionRef format = NULL;
error = CMAudioFormatDescriptionCreate(kCFAllocatorDefault, &asbd,?0, NULL,?0, NULL, NULL, &format);
if?(error) {
????NSLog(@"CMAudioFormatDescriptionCreate returned error: %ld", (long)error);
????return;
}
CMSampleBufferRef buff = NULL;
CMSampleTimingInfo timing = {CMTimeMake(1, sampleRate), kCMTimeZero, kCMTimeInvalid};
error = CMSampleBufferCreate(kCFAllocatorDefault, NULL,?false, NULL, NULL, format, (CMItemCount)len / (2?* channels),?1, &timing,?0, NULL, &buff);
if?(error) {
????NSLog(@"CMSampleBufferCreate returned error: %ld", (long)error);
????return;
}
CFRelease(format);
error = CMSampleBufferSetDataBufferFromAudioBufferList(buff, kCFAllocatorDefault, kCFAllocatorDefault,?0, &buffList);
if?(error) {
????NSLog(@"CMSampleBufferSetDataBufferFromAudioBufferList returned error: %ld", (long)error);
????return;
}
[self.writeManager appendSampleBuffer:buff ofMediaType:AVMediaTypeAudio];
CFRelease(buff);
