Fundamentally, my questions are: is there a known transform I can apply onto a given (pixel) position (passed into a Metal Fragment Function) to correctly sample a texture provided by the main cameras + processed by a Vision request. If so, what is it? If not, how can I accurately sample my masks? My goal is to highlight people in a Vision Pro app using Compositor Services. To start, I asynchronously receive camera frames for the main left and right cameras. This is the breakdown of the specific CameraVideoFormat I pass along to the CameraFrameProvider: minFrameDuration: 0.03 maxFrameDuration: 0.033333335 frameSize: (1920.0, 1080.0) pixelFormat: 875704422 cameraType: main cameraPositions: [left, right] cameraRectification: mono From each camera frame sample, I extract the left and right buffers (CVReadOnlyPixelBuffer.withUnsafebuffer ==> CVPixelBuffer). I asynchronously process the extracted buffers by performing a VNGeneratePersonSegmentationRequest on both of them: // NOTE: This block of code and all following code blocks contain simplified representations of my code for clarity's sake. var request = VNGeneratePersonSegmentationRequest() request.qualityLevel = .balanced request.outputPixelFormat = kCVPixelFormatType_OneComponent8 ... let lHandler = VNSequenceRequestHandler() let rHandler = VNSequenceRequestHandler() ... func processBuffers() async { try lHandler.perform([request], on: lBuffer) guard let lMask = request.results?.first?.pixelBuffer else {...} try rHandler.perform([request], on: rBuffer) guard let rMask = request.results?.first?.pixelBuffer else {...} appModel.latestPersonMasks = (lMask, rMask) } I store the two resulting CVPixelBuffers in my appModel. For both of these buffers aka grayscale masks: width (in pixels) = 512 height (in pixels) = 384 byters per row = 512 plane count = 0 pixel format type = 1278226488 I am using Compositor Services to render my content in Immersive Space. My implementation of Compositor Services is based off of the same code from Interacting with virtual content blended with passthrough. Within the Shaders.metal, the tint's Fragment Shader is now passed the grayscale masks (converted from CVPixelBuffer to MTLTexture via CVMetalTextureCacheCreateTextureFromImage() at the beginning of the main render pipeline). fragment float4 tintFragmentShader( TintInOut in [[stage_in]], ushort amp_id [[amplification_id]], texture2d<uint> leftMask [[texture(0)]], texture2d<uint> rightMask [[texture(1)]] ) { if (in.color.a <= 0.0) { discard_fragment(); } float2 uv; if (amp_id == 0) { // LEFT uv = ??????????????????????; } else { // RIGHT uv = ??????????????????????; } constexpr sampler linearSampler (mip_filter::linear, mag_filter::linear, min_filter::linear); // Sample the PersonSegmentation grayscale mask float maskValue = 0.0; if (amp_id == 0) { // LEFT if (leftMask.get_width() > 0) { maskValue = rightMask.sample(linearSampler, uv).r; } } else { // RIGHT if (rightMask.get_width() > 0) { maskValue = rightMask.sample(linearSampler, uv).r; } } if (maskValue > 250) { return (1.0, 1.0, 1.0, 0.5) } return in.color; } I need to correctly sample the masks for a given fragment. The LayerRenderer.Layout is set to .layered. From Developer Documentation. A layout that specifies each view’s content as a slice of a single texture. Using the Metal debugger, I know that the final render target texture for each view / eye is 1888 x 1792 pixels, giving an aspect ratio of 59:56. The initial CVPixelBuffer provided by the main left and right cameras is 1920x1080 (16:9). The grayscale CVPixelBuffer returned by the VNPersonSegmentationRequest is 512x384 (4:3). All of these aspect ratios are different. My questions come down to: is there a known transform I can apply onto a given (pixel) position to correctly sample a texture provided by the main cameras + processed by a Vision request. If so, what is it? If not, how can I accurately sample my masks? Within the tint's Vertex Shader, after applying the modelViewProjectionMatrix, I have tried every version I have been able to find that takes the pixel space position (= vertices[vertexID].position.xy) and the viewport size (1888x1792) to compute the correct clip space position (maybe = pixel space position.xy / (viewport size * 0.5)???) of the grayscale masks but nothing has worked. The "highlight" of the person segmentations is off: scaled a little too big, offset little to far up and off to the side.

Device: iPhone 16 Pro Max OS: iOS 26.0.1 AVCaptureDevice.DeviceType: builtInUltraWideCamera avtiveFormat: <AVCaptureDeviceFormat: 0x10ffb9ac0 'vide'/'420v' 1440x1080, { 1- 60 fps}, photo dims:{1440x1080,2016x1512}, fov:101.022, gdc fov:103.625, binned, max zoom:94.50 (upscales @1.40), system zoom range:1.0-3.0, AF System:1, ISO:15.0-3600.0, SS:0.000023-1.000000, system exposure bias range:-2.0-2.0, supports multicam, supports CS RoI, supports Smart Style, supports Smudge Detection> API: device.isFocusModeSupported(.continuousAutoFocus) == true setting: device.focusMode = .continuousAutoFocus setting is ok, but it's not working actually with continuousAutoFocus

I'm writing a program to control a PTZ camera connected via USB. I can get access to target camera's unique_id, and also other infos provided by AVFoundation. But I don't know how to locate my target USB device to send a UVC ControlRequest. There's many Cameras with same VendorID and ProductID connected at a time, so I need a more exact way to find out which device is my target. It looks that the unique_id provided is (locationID<<32|VendorID<<16|ProductID) as hex string, but I'm not sure if I can always assume this behavior won't change. Is there's a document declares how AVFoundation generate the unique_id for USB camera, so I can assume this convert will always work? Or is there's a way to send a PTZ control request to AVCaptureDevice? https://stackoverflow.com/questions/40006908/usb-interface-of-an-avcapturedevice I have seen this similar question. But I'm worrying that Exacting LocationID+VendorID+ProductID from unique_id seems like programming to implementation instead of interface. So, if there's any other better way to control my camera? here's my example code for getting unique_id: // // camera_unique_id_test.mm // // 测试代码：使用C++获取当前系统摄像头的AVCaptureDevice unique_id // // 编译命令: // clang++ -framework AVFoundation -framework CoreMedia -framework Foundation // camera_unique_id_test.mm -o camera_unique_id_test // #include <iostream> #include <string> #include <vector> #import <AVFoundation/AVFoundation.h> #import <Foundation/Foundation.h> struct CameraInfo { std::string uniqueId; }; std::vector<CameraInfo> getAllCameraDevices() { std::vector<CameraInfo> cameras; @autoreleasepool { NSArray<AVCaptureDevice*>* devices = [AVCaptureDevice devicesWithMediaType:AVMediaTypeVideo]; AVCaptureDevice* defaultDevice = [AVCaptureDevice defaultDeviceWithMediaType:AVMediaTypeVideo]; // 遍历所有设备 for (AVCaptureDevice* device in devices) { CameraInfo info; // 获取unique_id info.uniqueId = std::string([device.uniqueID UTF8String]); cameras.push_back(info); } } return cameras; } int main(int argc, char* argv[]) { std::vector<CameraInfo> cameras = getAllCameraDevices(); for (size_t i = 0; i < cameras.size(); i++) { const CameraInfo& camera = cameras[i]; std::cout << " 设备 " << (i + 1) << ":" << std::endl; std::cout << " unique_id: " << camera.uniqueId << std::endl; } return 0; } and here's my code for UVC control: // clang++ -framework Foundation -framework IOKit uvc_test.cpp -o uvc_test #include <iostream> #include <CoreFoundation/CoreFoundation.h> #include <IOKit/IOCFPlugIn.h> #include <IOKit/IOKitLib.h> #include <IOKit/IOMessage.h> #include <IOKit/usb/IOUSBLib.h> #include <IOKit/usb/USB.h> CFStringRef CreateCFStringFromIORegistryKey(io_service_t ioService, const char* key) { CFStringRef keyString = CFStringCreateWithCString(kCFAllocatorDefault, key, kCFStringEncodingUTF8); if (!keyString) return nullptr; CFStringRef result = static_cast<CFStringRef>( IORegistryEntryCreateCFProperty(ioService, keyString, kCFAllocatorDefault, kIORegistryIterateRecursively)); CFRelease(keyString); return result; } std::string GetStringFromIORegistry(io_service_t ioService, const char* key) { CFStringRef cfString = CreateCFStringFromIORegistryKey(ioService, key); if (!cfString) return ""; char buffer[256]; Boolean success = CFStringGetCString(cfString, buffer, sizeof(buffer), kCFStringEncodingUTF8); CFRelease(cfString); return success ? std::string(buffer) : std::string(""); } uint32_t GetUInt32FromIORegistry(io_service_t ioService, const char* key) { CFStringRef keyString = CFStringCreateWithCString(kCFAllocatorDefault, key, kCFStringEncodingUTF8); if (!keyString) return 0; CFNumberRef number = static_cast<CFNumberRef>( IORegistryEntryCreateCFProperty(ioService, keyString, kCFAllocatorDefault, kIORegistryIterateRecursively)); CFRelease(keyString); if (!number) return 0; uint32_t value = 0; CFNumberGetValue(number, kCFNumberSInt32Type, &value); CFRelease(number); return value; } int main() { // Get matching dictionary for USB devices CFMutableDictionaryRef matchingDict = IOServiceMatching(kIOUSBDeviceClassName); // Get iterator for matching services io_iterator_t serviceIterator; IOServiceGetMatchingServices(kIOMasterPortDefault, matchingDict, &serviceIterator); // Iterate through matching devices io_service_t usbService; while ((usbService = IOIteratorNext(serviceIterator))) { uint32_t locationId = GetUInt32FromIORegistry(usbService, "locationID"); uint32_t vendorId = GetUInt32FromIORegistry(usbService, "idVendor"); uint32_t productId = GetUInt32FromIORegistry(usbService, "idProduct"); IOCFPlugInInterface** plugInInterface = nullptr; IOUSBDeviceInterface** deviceInterface = nullptr; SInt32 score; // Get device plugin interface IOCreatePlugInInterfaceForService(usbService, kIOUSBDeviceUserClientTypeID, kIOCFPlugInInterfaceID, &plugInInterface, &score); // Get device interface (*plugInInterface) ->QueryInterface(plugInInterface, CFUUIDGetUUIDBytes(kIOUSBDeviceInterfaceID), (LPVOID*)&deviceInterface); (*plugInInterface)->Release(plugInInterface); // Try to find UVC control interface using CreateInterfaceIterator io_iterator_t interfaceIterator; IOUSBFindInterfaceRequest interfaceRequest; interfaceRequest.bInterfaceClass = kUSBVideoInterfaceClass; // 14 interfaceRequest.bInterfaceSubClass = kUSBVideoControlSubClass; // 1 interfaceRequest.bInterfaceProtocol = kIOUSBFindInterfaceDontCare; interfaceRequest.bAlternateSetting = kIOUSBFindInterfaceDontCare; (*deviceInterface) ->CreateInterfaceIterator(deviceInterface, &interfaceRequest, &interfaceIterator); (*deviceInterface)->Release(deviceInterface); io_service_t usbInterface = IOIteratorNext(interfaceIterator); IOObjectRelease(interfaceIterator); if (usbInterface) { std::cout << "Get UVC device with:" << std::endl; std::cout << "locationId: " << std::hex << locationId << std::endl; std::cout << "vendorId: " << std::hex << vendorId << std::endl; std::cout << "productId: " << std::hex << productId << std::endl << std::endl; IOObjectRelease(usbInterface); } IOObjectRelease(usbService); } IOObjectRelease(serviceIterator); }

My implementation of LockedCameraCapture does not launch my app when tapped from locked screen. But when the same widget is in the Control Center, it launches the app successfully. Standard Xcode target template: Lock_Screen_Capture.swift @main struct Lock_Screen_Capture: LockedCameraCaptureExtension { var body: some LockedCameraCaptureExtensionScene { LockedCameraCaptureUIScene { session in Lock_Screen_CaptureViewFinder(session: session) } } } Lock_Screen_CaptureViewFinder.swift: import SwiftUI import UIKit import UniformTypeIdentifiers import LockedCameraCapture struct Lock_Screen_CaptureViewFinder: UIViewControllerRepresentable { let session: LockedCameraCaptureSession var sourceType: UIImagePickerController.SourceType = .camera init(session: LockedCameraCaptureSession) { self.session = session } func makeUIViewController(context: Self.Context) -> UIImagePickerController { let imagePicker = UIImagePickerController() imagePicker.sourceType = sourceType imagePicker.mediaTypes = [UTType.image.identifier, UTType.movie.identifier] imagePicker.cameraDevice = .rear return imagePicker } func updateUIViewController(_ uiViewController: UIImagePickerController, context: Self.Context) { } } Then I have my widget: struct CameraWidgetControl: ControlWidget { var body: some ControlWidgetConfiguration { StaticControlConfiguration( kind: "com.myCompany.myAppName.lock-screen") { ControlWidgetButton(action: MyAppCaptureIntent()) { Label("Capture", systemImage: "camera.shutter.button.fill") } } } } My AppIntent: struct MyAppContext: Codable {} struct MyAppCaptureIntent: CameraCaptureIntent { typealias AppContext = MyAppContext static let title: LocalizedStringResource = "MyAppCaptureIntent" static let description = IntentDescription("Capture photos and videos with MyApp.") @MainActor func perform() async throws -> some IntentResult { .result() } } The Issue LockedCameraCapture Widget does not launch my app when tapped from locked screen. You get the Face ID prompt and takes you to just Home Screen. But when the same widget is in the Control Center, it launches the app successfully. Error Message When tapped on Lock Screen, I get the following error code: LaunchServices: store ‹private > or url ‹private > was nil: Error Domain=NSOSStatusErrorDomain Code=-54 "process may not map database" UserInfo=&NSDebugDescription=process may not map database, _LSLine=72, _LSFunction=_LSServer_GetServerStoreForConnectionWithCompletionHandler} Attempt to map database failed: permission was denied. This attempt will not be retried. Failed to initialize client context with error Error Domain=NSOSStatusErrorDomain Code=-54 "process may not map database" UserInfo=&NSDebugDescription=process may not map database, _LSLine=72, _LSFunction=_LSServer_GetServerStoreForConnectionWithCompletionHandler} Things I tried Widget image displays correctly App ID and the Provisioning Profile seem to be fine since they work fine when the same code injected in to AVCam sample app and when used the same App ID's. AppIntent file contains the target memberships of the Lock Screen capture and Widget Apple compiles without errors or warnings.

Just downloaded iOS 26.1 and my phone keeps ringing after the call has been answered. Any fixes for this?

I discovered when editing photos with the PhotoKit API, PHContentEditingOutput's renderedContentURL is a file in the app container's tmp directory with a filename that seems to follow the format render.<uuid>.JPG, and that file does not get deleted if the edit does not complete successfully (the user cancels the edit request, an error occurs, the app crashes, etc). I understand the system is supposed to automatically delete tmp files every once in a while, but some users are noticing my app's Documents & Data inflates, so I'm considering deleting these render files each time the app is launched. But I don't want to delete everything in the tmp directory as there could possibly be other data in there. What's the best way to remove those temporary files? Does the filename always start with render. no matter the device language? I thought I'd delete files in NSTemporaryDirectory() with that prefix but then I discovered in Mac Catalyst the location is not the tmp directory directly, they're in tmp/TemporaryItems/<bundleid>. Thanks!

Hi, We have created an app which allows recording 4K 60 fps videos in the app. We have noted that some time the recording switched to 20 fps (the value 20 is constant) even though the resolution settings is at 4K 60fps. We are using AVCaptureDevice to invoke the camera. has anyone experienced this problem before ? What is unique to 20 fps? Why does it resort to 20 fps from 60 fps and why not to other numbers ?

Area ImageCaptureCore / ICDeviceBrowser Description On iOS 26.1 beta, calling requestControlAuthorization() requestContentsAuthorization() always returns .notDetermined and never transitions to .authorized or .denied. This prevents apps from properly accessing device control or contents authorization. The issue occurs regardless of device state or prior requests. Steps to Reproduce 1. Create and start an ICDeviceBrowser instance. 2. Call requestControlAuthorization() or requestContentsAuthorization(). 3. Inspect the returned ICAuthorizationStatus. Expected Result • The system should prompt the user if necessary. • A final status of either .authorized or .denied should be returned. Actual Result • The completion handler always reports .notDetermined. • No user prompt appears and the status does not change. Version / Build • iOS 26.1 beta • Xcode Hardware • [iPhone 15 Pro, iPad Pro (M2)] Impact This regression blocks development and testing of features relying on ImageCaptureCore. Applications depending on device browsing and content access cannot proceed, which significantly affects workflows involving external device integration. Notes This appears to be a regression compared to earlier iOS releases.

When attempting to access a PHAsset that is in the hidden folder of iOS26, the PHFetchResult always returns no items, even when the user has granted full access to photos and even when includeHiddenAssets is true. This is the code suggested by ChatGPT; it always fails: public func fetchAsset(withLocalIdentifier identifier: String) -> PSSPHAssetImplementing? { // First try the direct fetch by identifier (fast path) let directResult = PHAsset.fetchAssets(withLocalIdentifiers: [identifier], options: nil) if let asset = directResult.firstObject { return build(from: asset) } // Fallback: fetch all assets including hidden, then filter manually let options = PHFetchOptions() options.includeHiddenAssets = true let allAssets = PHAsset.fetchAssets(with: options) for index in 0..<allAssets.count { let asset = allAssets.object(at: index) if asset.localIdentifier == identifier { return build(from: asset) } } return nil } Is it no longer possible to retrieve a hidden photo in iOS 26?

I want to fully support the new iPhone models in my app, and ideally need to know the available lenses. However, I can't find information about this on the web and they're not reported in the simulators. The closest thing I found was this, but it's very out of date. https://developer.apple.com/library/archive/documentation/DeviceInformation/Reference/iOSDeviceCompatibility/Cameras/Cameras.html My only other option is to buy each device, which isn't really feasible, or to log the data from real users via an analytics tool which isn't ideal either. Thanks, Alex

Hey, I'm having a very strange issue on my iPhone 17 Pro. I'm trying to capture a 12MP ProRAW image using the Telephoto Lens with the photoQualityPrioritization set to .Quality. Unfortunately I receive this error when trying to capture the image: Error Domain=AVFoundationErrorDomain Code=-11800 "The operation could not be completed" UserInfo={NSUnderlyingError=0x134f7a1f0 {Error Domain=NSOSStatusErrorDomain Code=-16802 "(null)"}, NSLocalizedFailureReason=An unknown error occurred (-16802), AVErrorRecordingFailureDomainKey=4, NSLocalizedDescription=The operation could not be completed} The photo captures correctly at 7.9x zoom, it's only a problem when the zoom goes over 8x. Also, it's only this particular configuration of settings which causes the issue. I'm able to capture an image if I either: Set quality to ".balanced" Set max dimensions to 48MP Capture a JPEG image instead of a ProRAW image Use the TripleCamera fusion lens Any help would be greatly appreciated. Alex

I am able to capture 48mp photos using .builtInWideAngleCamera, but it seems like .builtInTripleCamera is capped at 12mp? Is there a way to capture 48mp photos using .builtInTripleCamera? Because .builtInTripleCamera provides smooth transition between cameras during zooming, and I'd like to keep this behavior. New iPhone 17 Pro have all their cameras at 48mp. Is there a chance that their .builtInTripleCamera is capable of capturing 48mp? Or is this an API limitation?

I'm experiencing an issue with my app when saving images to the camera roll. This is intermittent, but it happens several times a day. The error I receive is the following: Connection to assetsd was interrupted - assetsd exited, died, or closed the photo library Error getting remote object proxy for -[PLNonBindingAssetsdPhotoKitClient sendChangesRequest:reply:]_block_invoke: Error Domain=NSCocoaErrorDomain Code=4097 "connection to service named com.apple.photos.service" UserInfo={NSDebugDescription=connection to service named com.apple.photos.service} PhotoKit XPC proxy is invalid. Dropping request on the floor and returning an error: Error Domain=PHPhotosErrorDomain Code=3301 "(null)" (underlying error Error Domain=NSCocoaErrorDomain Code=4097 "connection to service named com.apple.photos.service" UserInfo={NSDebugDescription=connection to service named com.apple.photos.service}) CoreData: error: XPC: synchronousRemoteObjectProxyWithErrorHandler: store 'file:///var/mobile/Media/PhotoData/Photos.sqlite' encountered error: Error Domain=NSCocoaErrorDomain Code=4099 "The connection to service created from an endpoint was invalidated: failed to check-in, peer may have been unloaded: mach_error=10000003." UserInfo={NSDebugDescription=The connection to service created from an endpoint was invalidated: failed to check-in, peer may have been unloaded: mach_error=10000003.} CoreData: error: XPC: synchronousRemoteObjectProxyWithErrorHandler: store 'file:///var/mobile/Media/PhotoData/Photos.sqlite' encountered error: Error Domain=NSCocoaErrorDomain Code=4099 "The connection to service created from an endpoint was invalidated: failed to check-in, peer may have been unloaded: mach_error=10000003." UserInfo={NSDebugDescription=The connection to service created from an endpoint was invalidated: failed to check-in, peer may have been unloaded: mach_error=10000003.} My code is unchanged from using my app daily on an iPhone 16 Pro with iOS 26. I never saw the issue on this device. Here is an excerpt from my code for saving the image: var localIdentifier = String() PHPhotoLibrary.shared().performChanges({ let albumChangeRequest = PHAssetCollectionChangeRequest(for: album) let assetCreationRequest = PHAssetCreationRequest.forAsset() let options = PHAssetResourceCreationOptions() assetCreationRequest.addResource(with: .photo, data: imageData, options: options) assetCreationRequest.creationDate = Date.now let placeHolder = assetCreationRequest.placeholderForCreatedAsset albumChangeRequest?.addAssets([placeHolder!] as NSArray) if placeHolder != nil { localIdentifier = (placeHolder?.localIdentifier)! } }) { (didSucceed, error) in OperationQueue.main.addOperation({ didSucceed ? success(localIdentifier) : failure(error) }) } I'm not sure why this would be device specific but I have had users with iPhone 17 Pro and iPhone Air reporting the issue. Alex

I want to use both front UW and TrueDepth cameras in iPad which has front UW camera. Firstly, I have used only front builtInDualCamera by AVFoundation and tried all the formats that can be used with builtInDualCamera, but there was no format that could capture UW. Secondly, I have tried to both front builtInDualCamera and builtInUltraWideCamera, but there was no combination that could use builtInUltraWideCamera and builtInDualCamera. Is there any way ?

I'm receiving output from avcapturesession and capturing an image using Vision, but the image is output in landscape orientation instead of portrait. Even when I set the orientation to up in ciimage, cgimage, and uiimage, the image is still output in landscape orientation. On iPhones 16 and below, the image is output in portrait orientation. But on iPhones 17 and above, the image is output in landscape orientation. Please help.

I tried to modify the AVCam sample code by copying the code here https://developer.apple.com/documentation/avfoundation/adopting-smart-framing-in-your-camera-app#Configure-the-smart-framing-monitor smart framing monitors I can ensure the activeformat supports smart framing, but the supported frames in monitor is always nil. In my another project it has supported value, but the observation has never been triggered, then I tried to keep printing the recommended frame, it's always nil. Could the engineer embed the code into AVCam rather than posting a few code pieces?

I'm getting an error writing a ciImage as a heif image: // Create CIImage directly from pixel buffer let ciImage = CIImage(cvPixelBuffer: pixelBuffer, options: [CIImageOption.properties: combinedMetadata]) // Write HEIC synchronously do { try ciContext.writeHEIFRepresentation(of: ciImage, to: url, format: .RGBA8, colorSpace: colorSpace) The error I'm getting is: Error Domain=CINonLocalizedDescriptionKey Code=3 "(null)" UserInfo={CINonLocalizedDescriptionKey=failed to write HEIC data to file., NSUnderlyingError=0x11b1a1ec0 {Error Domain=CINonLocalizedDescriptionKey Code=10 "(null)" UserInfo={CINonLocalizedDescriptionKey=failed to add image to the PhotoCompressionSession.}}} Both try ciContext.writeJPEGRepresentation(of: copiedCIImage, to: url, colorSpace: colorSpace, options: options) and try ciContext.writePNGRepresentation(of: copiedCIImage, to: url, format: .RGBA8, colorSpace: colorSpace) work. I also verified that the code works with iOS 18. Is there something new we need to do for Heif images? Thanks in advance

I tested the accuracy of the depth map on iPhone 12, 13, 14, 15, and 16, and found that the variance of the depth map after iPhone 12 is significantly greater than that of iPhone 12. Enabling depth filtering will cause the depth data to be affected by the previous frame, adding more unnecessary noise, especially when the phone is moving. This is not friendly for high-precision reconstruction. I tried to add depth map smoothing in post-processing to solve the problem of large depth map deviation, but the performance is still poor. Is there any depth map smoothing solutions already announced by Apple?

What options do I have if I don't want to use Blackmagic's Camera ProDock as the external Sync Hardware, but instead I want to create my own USB-C hardware accessory which would show up as an AVExternalSyncDevice on the iPhone 17 Pro? Which protocol does my USB-C device have to implement to show up as an eligible clock device in AVExternalSyncDevice.DiscoverySession?

Where can I find the documentation of the Genlock feature of the iPhone 17 Pro? How does it work and how can I use it in my app?