I'm getting Crashlytics crashes from some my users, deep in the Apple code: Crashed: AXSpeech EXC_BAD_ACCESS KERN_INVALID_ADDRESS 0x00000007ec54b360 0 libobjc.A.dylib 0x3c9c objc_retain_x8 + 16 1 AudioToolboxCore 0x99580 auoop::RenderPipeUser::~RenderPipeUser() + 112 2 AudioToolboxCore 0xe6090 -[AUAudioUnit_XPC internalDeallocateRenderResources] + 92 3 AVFAudio 0x90a0 AUInterfaceBaseV3::Uninitialize() + 60 4 AVFAudio 0x4cbe0 AVAudioEngineGraph::PerformCommand(AUGraphNodeBaseV3&, AVAudioEngineGraph::ENodeCommand, void*, unsigned int) const + 768 5 AVFAudio 0x56b0c AVAudioEngineGraph::_Uninitialize(NSError**) + 132 6 AVFAudio 0x7834 AVAudioEngineImpl::Stop(NSError**) + 388 7 AVFAudio 0x636c -[AVAudioEngine dealloc] + 52 8 TextToSpeech 0x30674 _TTSNameForVoiceInformation + 20864 9 libobjc.A.dylib 0x20a4 object_cxxDestructFromClass(objc_object*, objc_class*) + 116 10 libobjc.A.dylib 0x6e00 objc_destructInstance + 80 11 libobjc.A.dylib 0x104fc _objc_rootDealloc + 80 12 TextToSpeech 0x2d2f4 _TTSNameForVoiceInformation + 7680 13 TextToSpeech 0x496c TTSVocalizerCopyURLForFallbackResource + 8540 14 TextToSpeech 0x26094 TTSSpeechUnitTestingMode + 5548 15 libAXSpeechManager.dylib 0x108b0 -[AXSpeechManager .cxx_destruct] + 192 16 libobjc.A.dylib 0x20a4 object_cxxDestructFromClass(objc_object*, objc_class*) + 116 17 libobjc.A.dylib 0x6e00 objc_destructInstance + 80 18 libobjc.A.dylib 0x104fc _objc_rootDealloc + 80 19 libAXSpeechManager.dylib 0x5298 -[AXSpeechManager dealloc] + 268 20 Foundation 0x3b8a4 __NSThreadPerformPerform + 272 21 CoreFoundation 0xd3208 __CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE0_PERFORM_FUNCTION__ + 28 22 CoreFoundation 0xdf864 __CFRunLoopDoSource0 + 176 23 CoreFoundation 0x646c8 __CFRunLoopDoSources0 + 244 24 CoreFoundation 0x7a1c4 __CFRunLoopRun + 828 25 CoreFoundation 0x7f4dc CFRunLoopRunSpecific + 612 26 Foundation 0x420c4 -[NSRunLoop(NSRunLoop) runMode:beforeDate:] + 212 27 libAXSpeechManager.dylib 0x13390 -[AXSpeechThread main] + 552 28 Foundation 0x5b634 __NSThread__start__ + 716 29 libsystem_pthread.dylib 0x16b8 _pthread_start + 148 30 libsystem_pthread.dylib 0xb88 thread_start + 8 It's most likely related to my use of AVSpeechSynthesizer. I do change some of the utterance fields, including the voice that's being used (which is set to a value from speechVoices()). UtilAudioIos_tts = AVSpeechSynthesizer() let utterance = AVSpeechUtterance utterance.voice = AVSpeechSynthesisVoice(identifier: voice.voiceCode) utterance.volume = volume utterance.pitchMultiplier = pitch utterance.rate = rate UtilAudioIos_tts!.speak(utterance) By coincidence or not, the following sometimes appears in the device log: 2023-05-30 20:35:29.948078+0100 <appname>[466:12882] [catalog] Unable to list voice folder and also, sometimes: 2023-05-30 20:37:35.345933+0100 <appname>[466:13298] [catalog] Query for com.apple.MobileAsset.VoiceServices.VoiceResources failed: 2 2023-05-30 20:37:35.360854+0100 rehearserfree[466:13433] [AXTTSCommon] MauiVocalizer: 11006 (Can't compile rule): regularExpression=\Oviedo(?=, (\x1b\\pause=\d+\\)?Florida)\b, message=unrecognized character follows \, characterPosition=1 2023-05-30 20:37:35.363163+0100 <appname>[466:13433] [AXTTSCommon] MauiVocalizer: 16038 (Resource load failed): component=ttt/re, uri=, contentType=application/x-vocalizer-rettt+text, lhError=88602000 2023-05-30 20:37:35.363182+0100 <appname>[466:13433] [AXTTSCommon] Error loading rules: 2147483648 All of these crashes have been on the various versions of iOS 16. Edit: I can't reproduce the crash myself - it's just some (not all) app users. The log entries above appear locally on my device (with no crash) but I can't see the logs of the users who have the crashes. Any idea what this might be caused by, or how to go about tracking the problem down?

My app uses the AVFoundation to pronounce some words. Running the app from Xcode, either to a simulator or device, I frequently get this crash at start-up: AXSpeech (13): EXC_BAD_ACCESS (code=EXC_I386_GPFLT). It seems to occur randomly, maybe 20%-30% of the time I launch the app. When it does not crash, using audio works as expected. When launched from the device, it never crashes (so far, at least). Here's the code that outputs speech: Declared at the top level of the View struct: @State var synth = AVSpeechSynthesizer() In the View, as part of a Button's action closure: let utterance = AVSpeechUtterance(string: answer) utterance.voice = AVSpeechSynthesisVoice(language: "en_US") synth.speak(utterance) Any idea on how to stop this? It's annoying having to launch the app multiple times to test on a simulator or device.

Hi everyone, I've encountered a rare and strange crash in my app that I can't consistently reproduce. The crash seems to occur deep within Apple's internal frameworks, and I can't pinpoint which line of my own code is causing it. Here's the crash stack trace: #44 AXSpeech SIGSEGV SEGV_ACCERR 0 CoreFoundation ___CFCheckCFInfoPACSignature + 4 1 CoreFoundation _CFRunLoopSourceSignal + 28 2 Foundation _performQueueDequeue + 492 3 Foundation ___NSThreadPerformPerform + 88 4 CoreFoundation ___CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE0_PERFORM_FUNCTION__ + 28 5 CoreFoundation ___CFRunLoopDoSource0 + 176 6 CoreFoundation ___CFRunLoopDoSources0 + 340 7 CoreFoundation ___CFRunLoopRun + 828 8 CoreFoundation _CFRunLoopRunSpecific + 608 9 Foundation -[NSRunLoop(NSRunLoop) runMode:beforeDate:] + 212 10 TextToSpeech _TTSCFAttributedStringCreateStringByBracketingAttributeWithString + 776 11 Foundation ___NSThread__start__ + 732 12 libsystem_pthread.dylib __pthread_start + 136 Sometimes, instead of line 10 referencing _TTSCFAttributedStringCreateStringByBracketingAttributeWithString, it shows: 10 TextToSpeech LogWarning(char const*, ...) + 7288 Has anyone experienced a similar issue or know what might be triggering this crash? Any guidance on how to investigate or resolve this would be greatly appreciated. Thank you!

Hi everyone, I’m seeing recurring internal AVFoundation camera logs on iOS 26.2 and I’m trying to understand whether this is expected behavior or a regression in the capture pipeline. These logs appear shortly after starting an AVCaptureSession, while video frames are being delivered, and also when the camera is stopped or the capture session is torn down. <<<< FigXPCUtilities >>>> signalled err=-17281 at <>:302 <<<< FigCaptureSourceRemote >>>> Fig assert: "err == 0 " at bail (FigCaptureSourceRemote.m:569) - (err=-17281) Even in this clean, minimal setup, the same logs appear on iOS 26.2 The exact same logic did not produce these logs on iOS 18.x. To rule out issues caused by my own code, GPT created a minimal SwiftUI example from scratch. My primary interest is to perform real-time processing on the video frames delivered by the camera (via AVCaptureVideoDataOutput), for tasks such as analysis, computer vision, or custom frame handling, while simultaneously displaying the live preview. Thanks in advance for any insight. Example Code

I have a question regarding the behavior of AVAudioSession.sharedInstance().outputVolume. Observed behavior: When the app is in the foreground, I read audioSession.outputVolume (for example, 0.1). The app is then moved to the background. While the app is in the background, the user changes the system volume using the hardware buttons (for example, to 0.5). When the app returns to the foreground, audioSession.outputVolume still reports the previous value (0.1). From my testing, outputVolume only seems to update when the system volume is changed while the app is in the foreground. Volume changes made while the app is in the background are not reflected when the app returns to the foreground. Questions: According to Apple’s documentation for AVAudioSession.outputVolume: “The systemwide output volume set by the user.” https://developer.apple.com/documentation/avfaudio/avaudiosession/outputvolume However, based on our testing on iOS 18.6.2 and iOS 18.1, the observed behavior seems to differ from this description. Questions: The documentation states that outputVolume represents the system-wide volume set by the user. In our testing, the value does not reflect volume changes made while the app is in the background and only updates when the app is in the foreground.Is this the expected behavior of AVAudioSession.outputVolume? Is there any other recommended way in Swift to retrieve the current system volume that reflects user changes made both while the app is in the foreground and while it is in the background? Any clarification on the intended behavior or recommended handling would be greatly appreciated.

I’m implementing a custom playback pipeline using AVSampleBufferAudioRenderer together with AVSampleBufferRenderSynchronizer. hasSufficientMediaDataForReliablePlaybackStart appears to be the intended signal for determining when enough media has been queued to start playback. For local playback, this works well in practice — the property becomes true after a reasonable amount of media is enqueued. However, when the output route is AirPlay, using this property becomes difficult: AirPlay requires significantly more buffered media before the renderer reports sufficient data. The required preroll amount is much larger than for local playback. For short assets, it is possible to enqueue the entire audio track and still never observe hasSufficientMediaDataForReliablePlaybackStart == true. In that situation there is no more media data to enqueue, but the renderer still reports that playback is not ready. Given this behavior, what is the recommended way to determine playback readiness when using AVSampleBufferAudioRenderer with AirPlay?

Hello, I’m building an iOS video player using AVPlayer and a custom playback queue. I implemented remote controls using MPRemoteCommandCenter and enabled: nextTrackCommand previousTrackCommand playCommand pauseCommand I disabled: skipForwardCommand skipBackwardCommand seekForwardCommand seekBackwardCommand I also set queue metadata in MPNowPlayingInfoCenter: MPNowPlayingInfoPropertyPlaybackQueueIndex MPNowPlayingInfoPropertyPlaybackQueueCount Even with these commands enabled and the queue count greater than 1, the iOS lock screen continues to display the 10-second skip buttons instead of the previous/next track buttons. The commands themselves work correctly when triggered externally (Control Center, headphones, etc.), but the UI still shows the skip controls. Is there a way to force the lock screen UI to display previous / next track buttons for a video playlist? Or is this behavior expected when using AVPlayer with video content? Thanks.

Hello, I've been very familiar with the UVC Support in iPadOS ever since it launched in iOS 17. There are a number of people that use the software I've developed built around UVC and there are often queries about 8-Bit vs. 10-Bit. My understanding is that the newest UVC Spec is 1.5 which was standardised in 2012 and almost every UVC Capture Card runs at 8-Bit. The only 10-Bit Capture Card that is on my radar is the AJA U-Tap SDI, however it looks like this is 10-Bit up until the UVC Part where the 10-Bit Input is downsampled to 8-Bit. Though I have read in certain places that it works as a 10-Bit Capture Card on macOS but not on iPadOS. I was just wondering if 10-Bit via UVC is even possible on iPadOS? If there was indeed a true 10-Bit Source being passed into an iPad, would iPadOS allow it or would it be downsampled by AVFoundation so it can show up as a valid external video input? All USB Capture Cards that I have encountered use one of the following formats: kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange kCVPixelFormatType_420YpCbCr8BiPlanarFullRange kCVPixelFormatType_32BGRA So if a UVC Device delivered a 10-Bit Format, would that be accessible by iPadOS or would it fallback to these 8-Bit Formats by default? Thanks!

Hello everyone, I’m looking for more detailed information regarding UVC (USB Video Class) over MFi within the Apple ecosystem and would appreciate some clarification. I’m interested in developing (or interfacing with) an accessory that transmits video over USB using the UVC standard, and I’d like to better understand how this works within the MFi (Made for iPhone) program. Here are my main questions: 1. Do iOS devices provide native support for UVC over USB-C or Lightning within the MFi framework? 2. Are there any specific firmware or authentication requirements when the accessory is MFi-certified? 3. Does UVC support depend solely on the hardware interface (USB-C vs Lightning), or are there additional software-level requirements? 4. Is there any official documentation outlining the recommended flow for implementing UVC-based video capture accessories on iOS? From what I understand, USB-C iPads appear to offer more direct support for standard UVC devices, but it’s not entirely clear how this integrates with the MFi ecosystem with iOS, especially for commercial product development. If anyone has gone through this process or can point me to relevant technical documentation, I would greatly appreciate the guidance. Thank you!

Hi, I’m trying to better understand how AVAssetDownloadConfiguration selects video variants when downloading HLS content for offline playback. Suppose I have an HLS master playlist (.m3u8) that contains several video variants defined with #EXT-X-STREAM-INF. For example, the master playlist may contain multiple video streams like this: Same resolution, different BANDWIDTH Or different resolutions (for example 720p, 1080p, etc.) My question is: How many video variants are actually downloaded when using AVAssetDownloadConfiguration without specifying any variantQualifiers? In other words: If the master playlist contains multiple video variants, will the download task fetch only one variant, or multiple variants? Does the behavior differ depending on whether the variants differ only by BANDWIDTH or also by RESOLUTION? What I observed in testing In my tests, I always end up with only one video variant downloaded, specifically the one with the highest BANDWIDTH parameter. In the m3u8 files I tested, all video variants had identical parameters (resolution, codec, frame rate, etc.) and differed only by the BANDWIDTH attribute in the master playlist. However, when inspecting the downloaded .movpkg, I noticed something interesting in boot.xml. It lists two video streams: one with complete="true" (the one with highest bandwidth) another with complete="no" (the one with lowest bandwidth) I actually had 3 video streams listed in m3u8, but the one with middle bandwidth wasn't listed in boot.xml file at all. There are also additional streams for audio and subtitles in boot.xml file. This made me wonder whether the system initially attempts to download another video variant (possibly a lower bitrate one), but then switches to the highest-quality variant and only completes that one. Additional question about variantQualifiers If I provide a predicate such as: NSPredicate(format: "peakBitRate > 0") which should theoretically match all variants, will the download task attempt to download all matching video variants, or will it still select only one? Summary So the main questions are: Without variantQualifiers, does AVAssetDownloadConfiguration always download a single video variant, and if so, how is it chosen? Does the behavior differ if variants have different resolutions vs only different bitrates? When a predicate matches multiple variants, can multiple video variants actually be downloaded in a single .movpkg? Why might boot.xml list multiple video streams when only one appears to be fully downloaded? Any clarification on the intended behavior would be greatly appreciated. Thanks!

In Instruments, I'm seeing "Zero Time Stamp" events in the "Audio Server" lane. What does that mean?

Hello, I’m experiencing a severe performance degradation when running CoreML models on a live AVFoundation video feed compared to offline or synthetic inference. This happens across multiple models I've converted (including SCI, RTMPose, and RTMW) and affects multiple devices. The Environment OS: macOS 26.3, iOS 26.3, iPadOS 26.3 Hardware: Mac14,6 (M2 Max), iPad Pro 11 M1, iPhone 13 mini Compute Units: cpuAndNeuralEngine The Numbers When testing my SCI_output_image_int8.mlpackage model, the inference timings are drastically different: Synthetic/Offline Inference: ~1.34 ms Live Camera Inference: ~15.96 ms Preprocessing is completely ruled out as the bottleneck. My profiling shows total preprocessing (nearest-neighbor resize + feature provider creation) takes only ~0.4 ms in camera mode. Furthermore, no frames are being dropped. What I've Tried I am building a latency-critical app and have implemented almost every recommended optimization to try and fix this, but the camera-feed penalty remains: Matched the AVFoundation camera output format exactly to the model input (640x480 at 30/60fps). Used IOSurface-backed pixel buffers for everything (camera output, synthetic buffer, and resize buffer). Enabled outputBackings. Loaded the model once and reused it for all predictions. Configured MLModelConfiguration with reshapeFrequency = .frequent and specializationStrategy = .fastPrediction. Wrapped inference in ProcessInfo.processInfo.beginActivity(options: .latencyCritical, reason: "CoreML_Inference"). Set DispatchQueue to qos: .userInteractive. Disabled the idle timer and enabled iOS Game Mode. Exported models using coremltools 9.0 (deployment target iOS 26) with ImageType inputs/outputs and INT8 quantization. Reproduction To completely rule out UI or rendering overhead, I wrote a standalone Swift CLI script that isolates the AVFoundation and CoreML pipeline. The script clearly demonstrates the ~15ms latency on live camera frames versus the ~1ms latency on synthetic buffers. (I have attached camera_coreml_benchmark.swift and coreml model (very light low light enghancement model) to this repo on github https://github.com/pzoltowski/apple-coreml-camera-latency-repro). My Question: Is this massive overhead expected behavior for AVFoundation + Core ML on live feeds, or is this a framework/runtime bug? If expected, what is the Apple-recommended pattern to bypass this camera-only inference slowdown? One think found interesting when running in debug model was faster (not as fast as in performance benchmark but faster than 16ms. Also somehow if I did some dummy calculation on on different DispatchQueue also seems like model got slightly faster. So maybe its related to ANE Power State issues (Jitter/SoC Wake) and going to fast to sleep and taking a long time to wakeup? Doing dummy calculation in background thought is probably not a solution. Thanks in advance for any insights!

In iOS 26, AVSpeechSynthesizer read Mandarin into Cantonese pronunciation. No matter how you set the language, and change the settings of my phone system, it doesn't work. let utterance = AVSpeechUtterance(string: "你好啊") //let voice = AVSpeechSynthesisVoice(language: "zh-CN") // not work let voice = AVSpeechSynthesisVoice(language: "zh-Hans") // not work too utterance.voice = voice et synth = AVSpeechSynthesizer() synth.speak(utterance)

Please include the line below in follow-up emails for this request. Case-ID: 11089799 When using AVSpeechUtterance and setting it to play in Mandarin, if Siri is set to Cantonese on iOS 18, it will be played in Cantonese. There is no such issue on iOS 17 and 16. 1.let utterance = AVSpeechUtterance(string: textView.text) let voice = AVSpeechSynthesisVoice(language: "zh-CN") utterance.voice = voice 2.In the phone settings, Siri is set to Cantonese

My code that streams buffers into AVAudioPlayerNode is stuttering when the buffer is finished and before the next one is played. while engine.isRunning { let framesToCopy = min(buffer.frameLength - framePosition, Self.BufferSize) let srcRaw = UnsafeRawPointer(srcPtr) let playbackBuffer = AVAudioPCMBuffer(pcmFormat: buffer.format, frameCapacity: Self.BufferSize)! let playbackPtr = playbackBuffer.floatChannelData![0] let destRaw = UnsafeMutableRawPointer(mutating: playbackPtr) memcpy(destRaw, srcRaw, Int(framesToCopy) * MemoryLayout<Float>.stride) srcPtr = srcPtr.advanced(by: Int(framesToCopy)) playbackBuffer.frameLength = framesToCopy await player.scheduleBuffer(playbackBuffer, at: nil, options: [], completionCallbackType: .dataRendered) } I've tried to schedule multiple buffers at once using a combination of both the synchronous and async versions of scheduleBuffer because I thought the delay might be but it still stutters and the data copied into the playbackBuffer matches the source buffer. I've tried all combinations of options and completionCallbackType but no luck. I've tried increasing the buffer size but that just spaces out the stutters because the buffer is larger. What am I missing about this API?

Hi everyone, We’re encountering an issue where AudioQueueNewOutput blocks indefinitely and never returns, and we’re hoping to get some insight or confirmation if this is a known behavior/regression on newer iOS versions. Issue Description When triggering audio playback, we create an output AudioQueue using AudioQueueNewOutput. On some devices, the call hangs inside AudioQueueNewOutput and never returns, with no OSStatus error and no subsequent logs. This behavior is reproducible mainly on iOS 18.3. Earlier iOS versions do not show this issue under the same code path. if (audioDes) { mAudioDes.mSampleRate = audioDes->mSampleRate; mAudioDes.mBitsPerChannel = audioDes->mBitsPerChannel; mAudioDes.mChannelsPerFrame = audioDes->mChannelsPerFrame; mAudioDes.mFormatID = audioDes->mFormatID; mAudioDes.mFormatFlags = audioDes->mFormatFlags; mAudioDes.mFramesPerPacket = audioDes->mFramesPerPacket; mAudioDes.mBytesPerFrame = audioDes->mBytesPerFrame; mAudioDes.mBytesPerPacket = audioDes->mBytesPerFrame; mAudioDes.mReserved = 0; } // Create AudioQueue for output OSStatus status = AudioQueueNewOutput( &mAudioDes, AQOutputCallback, this, NULL, NULL, 0, &audioQueue ); code-block The thread blocks inside AudioQueueNewOutput, and execution never reaches the next line. Additional Notes / Observations ASBD is confirmed to be valid Standard PCM output Sample rate, channels, bytes per frame/packet all consistent Same ASBD works correctly on earlier iOS versions AudioQueue is created on a background thread Not on the main thread Not inside the AudioQueue callback On first creation, AVAudioSession may not yet be active setCategory and setActive:YES may be called shortly before creating the AudioQueue There may be a timing window where the session is still activating Issue is reported mainly on iOS 18.3 Multiple user reports point to iOS 18.3 devices Same code path works on iOS 17.x and earlier No OSStatus error is returned — the call simply never returns. Questions Is it expected that AudioQueueNewOutput can block indefinitely while waiting for AVAudioSession / audio route / HAL readiness? Have there been any behavior changes in iOS 18.3 regarding AudioQueue creation or AudioSession synchronization? Is it unsafe to call AudioQueueNewOutput before AVAudioSession is fully active on recent iOS versions? Are there recommended patterns (or delays / callbacks) to ensure AudioQueue creation does not hang? Any insight or confirmation would be greatly appreciated. Thanks in advance!

I read somewhere that the frames are returned in decode order instead of presentation order when using AVAssetReader. The documentation seems sparse on the subject. I have so far failed to find a video file where the frames are not returned in presentation order. Can anyone confirm the frames are actually returned in decode order?

Hello, I am currently developing a live streaming application using AVPlayer to play LL-HLS (Low-Latency HLS) content. During our testing phase, we consistently encountered the following error in the logs: CoreMediaErrorDomain Code=-15517 The challenge we are facing is that the error description is quite vague. It only provides cryptic messages such as "Key not found" or "No value information," which makes it extremely difficult to identify the root cause or perform a deep-dive analysis. I have searched through the official Apple Developer documentation and technical notes, but I couldn’t find any specific reference to what Code -15517 signifies in the context of LL-HLS or CoreMedia. Regarding this issue, I have the following questions: What is the specific meaning of this error code (-15517)? Does it relate to missing tags in the HLS manifest, or is it an internal state issue within the AVPlayer stack? Specifically, I would like to know if this is a critical error that disrupts playback, or if it is just a warning that can be safely ignored. Is there any additional logging or debugging tool you would recommend to further investigate "Key not found" issues in LL-HLS? Any insights or guidance from the community or Apple engineers would be greatly appreciated. Thank you in advance for your help.

I just bought a monitor S2725QC from Dell Technologies and isn't fully-integrated with MacOS even though it says on the website it is compatible with MacOS. https://www.dell.com/en-us/shop/all-monitors/sac/monitors/all-monitors/macos-compatible?appliedRefinements=51765 The screen brightness and volume control buttons don't work with the monitors (I have two). What can I do in terms of writing code with Dell Monitor SDK and MacOS Frameworks/Technologies?

I want to: Run ARKit on the main rear camera, and while it's running shoot high resolution pictures on the wide camera, without disturbing the AR tracking. Is this possible?