I know that if you want background audio from AVPlayer you need to detatch your AVPlayer from either your AVPlayerViewController or your AVPlayerLayer in addition to having your AVAudioSession configured correctly. I have that all squared away and background audio is fine until we introduce AVPictureInPictureController or use the PiP behavior baked into AVPlayerViewController. If you want PiP to behave as expected when you put your app into the background by switching to another app or going to the homescreen you can't perform the detachment operation otherwise the PiP display fails. On an iPad if PiP is active and you lock your device you continue to get background audio playback. However on an iPhone if PiP is active and you lock the device the audio pauses. However if PiP is inactive and you lock the device the audio will pause and you have to manually tap play on the lockscreen controls. This is the same between iPad and iPhone devices. My questions are: Is there a way to keep background-audio playback going when PiP is inactive and the device is locked (iPhone and iPad) Is there a way to keep background-audio playback going when PiP is active and the device is locked? (iPhone)

I work on a video conferencing application, which makes use of AVAudioEngine and the videoChat AVAudioSession.Mode This past Friday, an internal user reported an "audio cutting in and out" issue with their new iPhone 14 Pro, and I was able to reproduce the issue later that day on my iPhone 14 Pro Max. No other iOS devices running iOS 16 are exhibiting this issue. I have narrowed down the root cause to the videoChat AVAudioSession.Mode after changing line 53 of the ViewController.swift file in Apple's "Using Voice Processing" sample project (https://developer.apple.com/documentation/avfaudio/audio_engine/audio_units/using_voice_processing) from: try session.setCategory(.playAndRecord, options: .defaultToSpeaker) to try session.setCategory(.playAndRecord, mode: .videoChat, options: .defaultToSpeaker) This only causes issues on my iPhone 14 Pro Max device, not on my iPhone 13 Pro Max, so it seems specific to the new iPhones only. I am also seeing the following logged to the console using either device, which appears to be specific to iOS 16, but am not sure if it is related to the videoChat issue or not: 2022-09-19 08:23:20.087578-0700 AVEchoTouch[2388:1474002] [as] ATAudioSessionPropertyManager.mm:71  Invalid input size for property 1684431725 2022-09-19 08:23:20.087605-0700 AVEchoTouch[2388:1474002] [as] ATAudioSessionPropertyManager.mm:225  Invalid input size for property 1684431725 I am assuming 1684431725 is 'dfcm' but I am not sure what Audio Session Property that might be.

I am getting an error in iOS 16. This error doesn't appear in previous iOS versions. I am using RemoteIO to playback live audio at 4000 hz. The error is the following: Input data proc returned inconsistent 2 packets for 186 bytes; at 2 bytes per packet, that is actually 93 packets This is how the audio format and the callback is set: // Set the Audio format AudioStreamBasicDescription audioFormat; audioFormat.mSampleRate = 4000; audioFormat.mFormatID = kAudioFormatLinearPCM; audioFormat.mFormatFlags = kAudioFormatFlagIsSignedInteger | kAudioFormatFlagIsPacked; audioFormat.mFramesPerPacket = 1; audioFormat.mChannelsPerFrame = 1; audioFormat.mBitsPerChannel = 16; audioFormat.mBytesPerPacket = 2; audioFormat.mBytesPerFrame = 2; AURenderCallbackStruct callbackStruct; // Set output callback callbackStruct.inputProc = playbackCallback; callbackStruct.inputProcRefCon = (__bridge void * _Nullable)(self); status = AudioUnitSetProperty(audioUnit, kAudioUnitProperty_SetRenderCallback, kAudioUnitScope_Global, kOutputBus, &callbackStruct, sizeof(callbackStruct)); Note that the mSampleRate I set is 4000 Hz. In iOS 15 I get 0.02322 seconds of buffer duration (IOBufferDuration) and 93 frames in each callback. This is expected, because: number of frames * buffer duration = sampling rate 93 * 0.02322 = 4000 Hz However, in iOS 16 I am getting the aforementioned error in the callback. Input data proc returned inconsistent 2 packets for 186 bytes; at 2 bytes per packet, that is actually 93 packets Since the number of frames is equal to the number of packets, I am getting 1 or 2 frames in the callback and the buffer duration is of 0.02322 seconds. This didn't affect the playback of the "raw" signal, but it did affect the playback of the "processed" signal. number of frames * buffer duration = sampling rate 2 * 0.02322 = 0.046 Hz That doesn't make any sense. This error appears for different sampling rates (8000, 16000, 32000), but not for 44100. However I would like to keep 4000 as my sampling rate. I have also tried to set the sampling rate by using the setPreferredSampleRate(_:) function of AVAudioSession, but the attempt didn't succeed. The sampling rate was still 44100 after calling that function. Any help on this issue would be appreciated.

Hi Folks, I'm currently working on video conferencing app and use AVRoutePickerView for output device selection. Since iOS 16 release it started to display incorrect names for ear-piece and speaker options. For both of them the name is iPhone (before it was iPhone/Speaker) Output changes works correctly but display names confuse. Here is my audio session configuration:     private func configureSession() {         let configuration = RTCAudioSessionConfiguration.webRTC()         configuration.category = AVAudioSession.Category.playAndRecord.rawValue         configuration.categoryOptions = [.allowBluetooth, .allowBluetoothA2DP, .duckOthers, .mixWithOthers]         let session = RTCAudioSession.sharedInstance()         session.lockForConfiguration()         do {             try session.setConfiguration(configuration)         } catch let error as NSError {             logError("[AudioSessionManager] Unable to configure RTCAudioSession with error: \(error.localizedDescription)")         }         session.unlockForConfiguration()     }          private func overrideOutputPortIfNeeded() {         DispatchQueue.main.async {             guard let currentOutputType = self.session.currentRoute.outputs.first?.portType else { return }                      self.session.lockForConfiguration()             let shouldOverride = [.builtInReceiver, .builtInSpeaker].contains(currentOutputType)             logDebug("[AudioSessionManager] Should override output to speaker? \(shouldOverride)")             if shouldOverride {                 do {                     try self.session.overrideOutputAudioPort(.speaker)                 } catch let error as NSError {                     logError("[AudioSessionManager] Unable to override output to Speaker: \(error.localizedDescription)")                 }             }             self.session.unlockForConfiguration()         }     } Any help appreciated, Thansk!

I tried to press system Talk button to talk in channel when my APP in background, I see the channelManager(_ : PTChannelManager, _: UUID, didBeginTransmittingFrom _: PTChannelTransmitRequestSource) is called, but the function func channelManager(_ : PTChannelManager, didActivate _: AVAudioSession) never called. When my APP in foreground, everything works fine. I've set the APP's capability Background Modes -> Push to Talk. Anyone can help?

Turn on address sanitizer on Xcode and use a real device and put a Test.mp3 file in the Xcode project. Then it will crash when you initialise a AVAudioPlayer with a mp3 file (with a wav file it works fine). I have made an entry in feedback assistent -> FB12425453. var player : AVAudioPlayer? func playSound() { if let url = Bundle.main.url(forResource: "Test", withExtension: "mp3") { self.player = try? AVAudioPlayer(contentsOf: url) // --> deallocation of non allocated memory problem --> with a "wav" file it works .... } }

I'd like to allow the speech synthesizer to play on the device speaker while simultaneously mixing with a phone call. I've worked with a number of different configurations but am unable to find a configuration that achieves the functionality I am trying to achieve - or allows mixing with a phone call at all. There is a flag: mixToTelephonyUplink that seems to suggest that at least some mixing with a phone call is possible using the speech synthesizer, but I'm currently unable to find almost any documentation about this flag besides basic API docs. I've had some some luck at least getting the synthesizer to always play to the speaker with the following audio session configuration - but the sound never is mixed with a phone call. Instead, it is ducked and muted while the phone call takes place. I've tried quite a few configuration combinations for the category and overrides, but nothings seems to work quite as I'd expect it to. synthesizer.mixToTelephonyUplink = true try? audioSession.setCategory(.playback, mode: .voicePrompt, options: [.mixWithOthers, .defaultToSpeaker]) try? audioSession.setActive(true, options: []) try? audioSession.overrideOutputAudioPort(.speaker) Is there some kind of documentation for this that's off the beaten path that I'm somehow missing? I'm going to continue with guess and check, but I'm starting to think this flag - and the functionality it implies, actually wasn't ever fully implemented.

I was play a pcm(24kHz 16bit) file with AudioQueue, and it mix with other sound( 192kHz 24bit) named sound2. Setting for AVAudioSession as: category (AVAudioSessionCategoryPlayback), options (AVAudioSessionCategoryOptionMixWithOthers|AVAudioSessionCategoryOptionDuckOthers) when playing pcm the sound2 should pushed volume lower as setting. BUT, there has a absolutly mute keep 0.5 second when the sound2 become low, and after a 0.5s mute the pushed lower sound came out. It only become in sound2 situation(192kHz, 24bit). if the sound2's quality lower everything is OK.

The loop plays smoothly in audacity but when I run it in the device or simulator it clicks each loop at different intensities. I config the session at App level: let audioSession = AVAudioSession.sharedInstance() do { try audioSession.setCategory(.playback, mode: .default, options: [.mixWithOthers]) try audioSession.setActive(true) } catch { print("Setting category session for AVAudioSession Failed") } And then I made my method on my class: func playSound(soundId: Int) { let sound = ModelData.shared.sounds[soundId] if let bundle = Bundle.main.path(forResource: sound.filename, ofType: "flac") { let backgroundMusic = NSURL(fileURLWithPath: bundle) do { audioPlayer = try AVAudioPlayer(contentsOf:backgroundMusic as URL) audioPlayer?.prepareToPlay() audioPlayer?.numberOfLoops = -1 // for infinite times audioPlayer?.play() isPlayingSounds = true } catch { print(error) } } } Does anyone have any clue? Thanks! PS: If I use AVQueuePlayer and repeat the item the click noise disappear (but its no use, because I would need to repeat it indefinitely without wasting memory), if I use AVLooper I get a silence between loops. All with the same sound. Idk :/ PS2: The same happens with ALAC files.

AVSpeechSynthesizer was not working. it was working perfect before. below is my code objective - c. -(void)playVoiceMemoforMessageEVO:(NSString*)msg { [[AVAudioSession sharedInstance] overrideOutputAudioPort:AVAudioSessionPortOverrideSpeaker error:nil]; AVSpeechSynthesizer *synthesizer = [[AVSpeechSynthesizer alloc]init]; AVSpeechUtterance *speechutt = [AVSpeechUtterance speechUtteranceWithString:msg]; speechutt.volume=90.0f; speechutt.rate=0.50f; speechutt.pitchMultiplier=0.80f; [speechutt setRate:0.3f]; speechutt.voice = [AVSpeechSynthesisVoice voiceWithLanguage:@"en-us"]; [synthesizer speakUtterance:speechutt]; } please help me to solve this issue.

When recording audio over bluetooth from AirPods to iPhone using the AVAudioRecorder the Bluetooth audio codec used is always AAC-ELD independent of the codec to store which is selected in the AVAudioRecorder instance. As far as I know must every Bluetooth device support SBC, hence, it should be possible for the AirPods to transmit the recorded audio using the SBC codec instead of AAC-ELD. However, I could not find any resource on how the request this codec using the AVAudioRecorder or AVAudioEngine. Is it possible to request SBC at all and if yes how?

AVSpeechSynthesisVoice.speechVoices() returns voices that are no longer available after upgrading from iOS 16 to iOS 17 (although this has been an issue for a long time, I think). To reproduce: On iOS 16 download 1 or more enhanced voices under “Accessibility > Spoken Content > Voices”. Upgrade to iOS 17 Call AVSpeechSynthesisVoice.speechVoices() and note that the voices installed in step (1) are still present, yet they are no longer downloaded, therefore they don’t work. And there is no property on AVSpeechSynthesisVoice to indicate if the voice is still available or not. This is a problem for apps that allow users to choose among the available system voices. I receive many support emails surrounding iOS upgrades about this issue. I have to tell them to re-download the voices which is not obvious to them. I've created a feedback item for this as well (FB12994908).

Hi! I am working on an audio application on iOS. This is how I retreive the workgroup from the remoteIO audiounit (ioUnit). The unit is initialized and is working fine (meaning that it is regularly called by the system). os_workgroup_t os_workgroup{nullptr}; uint32_t os_workgroup_index_size; if (status = AudioUnitGetProperty(ioUnit, kAudioOutputUnitProperty_OSWorkgroup, kAudioUnitScope_Global, 0, &os_workgroup, &os_workgroup_index_size); status != noErr) { throw runtime_error("AudioUnitSetProperty kAudioOutputUnitProperty_OSWorkgroup - Failed with OSStatus: " + to_string(status)); } However the resulting os_workgroup's value is 0x40. Which seems not correct. No wonder I cannot join any other realtime threads to the workgroup as well. The returned status however is a solid 0. Can anyone help?

From an app that reads audio from the built-in microphone, I'm receiving many crash logs where the AVAudioEngine fails to start again after the app was suspended. Basically, I'm calling these two methods in the app delegate's applicationDidBecomeActive and applicationDidEnterBackground methods respectively: let audioSession = AVAudioSession.sharedInstance() func startAudio() throws { self.audioEngine = AVAudioEngine() try self.audioSession.setCategory(.record, mode: .measurement)} try audioSession.setActive(true) self.audioEngine!.inputNode.installTap(onBus: 0, bufferSize: 4096, format: nil, block: { ... }) self.audioEngine!.prepare() try self.audioEngine!.start() } func stopAudio() throws { self.audioEngine?.stop() self.audioEngine?.inputNode.removeTap(onBus: 0) self.audioEngine = nil try self.audioSession.setActive(false, options: [.notifyOthersOnDeactivation]) } In the crash logs (iOS 16.6) I'm seeing that this works fine several times as the app is opened and closed, but suddenly the audioEngine.start() call fails with the error Error Domain=com.apple.coreaudio.avfaudio Code=-10851 "(null)" UserInfo={failed call=err = AUGraphParser::InitializeActiveNodesInInputChain(ThisGraph, *GetInputNode())} and the audioEngine!.inputNode.outputFormat(forBus: 0) is something like <AVAudioFormat 0x282301c70: 2 ch, 0 Hz, Float32, deinterleaved> . Also, right before installing the tap, audioSession.availableInputs contains an entry of type MicrophoneBuiltIn but audioSession.currentRoute lists no inputs at all. I was not able to reproduce this situation on my own devices yet. Does anyone have an idea why this is happening?

when i found other sounds by using '[AVAudioSession sharedInstance].otherAudioPlaying', if it return true, i will get currentRoute. but i meet some issues, it's outputs & inputs number all > 0 . i want to judge the sound is by inputs devices or output devices? please help!!!!!!!

in iOS 17 (21A5326A) audioSession.setCategory(.playAndRecord, mode: .default,options: .allowBluetooth) does not set input to bluetooth. In iOS 16 it does. Here the steps to reproduce: Create project with storyboard. in info.plist add NSMicrophoneUsageDescription Your microphone will be used to record your speech when you press the "Start Recording" button. put in ViewController: import UIKit import Speech class ViewController: UIViewController { override func viewDidLoad() { super.viewDidLoad() // Do any additional setup after loading the view. } override public func viewDidAppear(_ animated: Bool) { super.viewDidAppear(animated) AVAudioSession.sharedInstance().requestRecordPermission { granted in } } func startAudioSession(){ let audioSession = AVAudioSession.sharedInstance() do { try audioSession.setCategory(.playAndRecord, mode: .default,options: .allowBluetooth) try audioSession.setActive(true,options: .notifyOthersOnDeactivation) print(audioSession.currentRoute.description) } catch { } } @IBAction func btnTap(_ sender: UIButton) { startAudioSession() } } put button on the Main.storyboard and link it to btnTap Connect bluetooth headset to iphone, start the App and tap button. in iOS 16 see the current route - bluetooth. in iOS 17 see the current route - speaker

I've integrated MPVolumeView into my view, and it correctly responds to hardware volume changes as expected. However, once I initiate audio streaming using AVAudioEngine to capture microphone audio and AudioUnit for decoding, the MPVolumeView ceases to reflect changes made using the hardware volume buttons. Additionally, even when I adjust the volume using the slider on MPVolumeView, it doesn't change the system volume. Has anyone else encountered this issue? What might be causing MPVolumeView to stop responding to hardware volume changes once streaming starts? For the AVAudioSession.Mode, I use the default setting because using .voiceChat prevents MPVolumeView update from device volume changes permanently. let session = AVAudioSession.sharedInstance() do { try session.setCategory(.playAndRecord, options: [.allowBluetooth]) try session.setActive(true) } catch { print(error.localizedDescription) }

Hello, I have struggled to resolve issue above question. I could speak utterance when I turn on my iPhone, but when my iPhone goes to background mode(turn off iPhone), It doesn't speak any more. I think it is possible to play audio or speak utterance because I can play music on background status in youtube. Any help please??

I want the audio session to always use the built-in microphone. However, when using the setPreferredInput() method like in this example private func enableBuiltInMic() { // Get the shared audio session. let session = AVAudioSession.sharedInstance() // Find the built-in microphone input. guard let availableInputs = session.availableInputs, let builtInMicInput = availableInputs.first(where: { $0.portType == .builtInMic }) else { print("The device must have a built-in microphone.") return } // Make the built-in microphone input the preferred input. do { try session.setPreferredInput(builtInMicInput) } catch { print("Unable to set the built-in mic as the preferred input.") } } and calling that function once in the initializer, the audio session still switches to the external microphone once one is plugged in. The session's preferredInput is nil again at that point, even if the built-in microphone is still listed in availableInputs. So, why is the preferredInput suddenly reset? when would be the appropriate time to set the preferredInput again? Observing the session’s availableInputs did not work and setting the preferredInput again in the routeChangeNotification handler seems a bad choice as it’s already a bit too late then.

Hey, I've been trying to start an audio recording from an intent started from a widget in iOS 17. However every time I run audioRecorder.record() it just returns false. Is it generally possible to start an audio recording from a widget's intent? In the simulator this actually worked, but on device it would not start the recording. Would love to get any pointers to what might be required to make this work.