Hi Apple Team and Developers, First of all, I’d like to express my appreciation for the incredible results achieved using PhotogrammetrySession. I’ve been developing a portrait scanning app using Object Capture, and in many tests—especially with human models—I’ve found the reconstructed body surfaces are remarkably smooth and clean, often outperforming tools like Metashape and RealityCapture in terms of aesthetic results. However, I’ve encountered some challenges when working with complex areas like long hair overlapping the face. For instance, with female models where strands of hair partially occlude the face, the resulting mesh tends to merge the hair and facial geometry. This leads to distorted or “melted” facial features, likely due to ambiguity in the geometry estimation phase. Feature Suggestion: Would it be possible to allow developers to supply two versions of the input images: • One version (original) for texture generation • A pre-processed version (e.g., contrast-enhanced or CLAHE filtered) to guide mesh reconstruction only This would give us the flexibility to enhance edge features or shadow detail without affecting the final texture appearance. In other photogrammetry pipelines, applying image enhancement selectively before dense reconstruction improves geometry quality in low-contrast areas. Question: Is there any plan to support this kind of two-path workflow in future versions of PhotogrammetrySession? Or perhaps expose more intermediate stages or tunable parameters to developers? Also, any hints on what we can expect from WWDC 2025 regarding improvements to Object Capture or related vision/3D technologies? Thanks again for this powerful API. Looking forward to hearing insights from the team and other developers. Warm regards, KitCheng

Hi guys, In visionOS, when using a ZStack decorated with .glassBackgroundEffect(), you can see the 3D glass background from the front, but when viewed from the side, the view appears to have no thickness. However, I noticed that in an app built by Apple, when viewing a glass background view from the side, it appears to have thickness. I tried adding .frame(depth:) to a glass background view, but it appears as two separate layers spaced by the depth value. My question is: Is there a view modifier that adds visual thickness to a glass background view, as shown in the picture? Or, if not, how should I write a custom view modifier to achieve this effect? Thanks!

In Reality Composer Pro, why is the Sky Sphere so much larger than the Sky Dome? By my estimate, the Sky Sphere has a radius of 100m, while the Sky only has a radius of only 12m.

Hi, I would like to train Gaussian splats from my object captures. So I need a pointcloud and camera positions together with the original photos taken to train GS In an app like postShot. I could do this with Reality Capture, which supports exporting pointclouds and camera position but it does not do well with turntable photogrammetry. While the Apple object capture API does produce really solid results with turntable images. so my question is, can I export camera data from my object captures to use in another application? Or is there may be a plan to at this feature in the future? It would be really helpful in creating ultra realistic, 3-D objects in Gaussian splat format. Thanks for any isuggestions…

I'm placing sphere at finger tip and updating its position as hand move. Finger joint tracking functions correctly, but I’ve observed noticeable latency in hand tracking updates whenever a UITextView becomes active. This lag happens intermittently during app usage, lasting about 5–10 seconds, after which the latency disappears and the sphere starts following the finger joints immediately. When I open the immersive space for the first time, the profiler shows a large performance spike upto 328%. After that, it stabilizes and runs smoothly. Note: I don’t observe any lag when CPU usage spikes to 300% (upon immersive view load) yet the lag still occurs even when CPU usage remains below 100%. I’m using the following code for hand tracking: private func processHandTrackingUpdates() async { for await update in handTracking.anchorUpdates { let handAnchor = update.anchor if handAnchor.isTracked { switch handAnchor.chirality { case .left: leftHandAnchor = handAnchor updateHandJoints(for: handAnchor, with: leftHandJointEntities) case .right: rightHandAnchor = handAnchor updateHandJoints(for: handAnchor, with: rightHandJointEntities) } } else { switch handAnchor.chirality { case .left: leftHandAnchor = nil hideAllJoints(in: leftHandJointEntities) case .right: rightHandAnchor = nil hideAllJoints(in: rightHandJointEntities) } } await MainActor.run { handTrackingData.processNewHandAnchors( leftHand: self.leftHandAnchor, rightHand: self.rightHandAnchor ) } } } And here’s the function I’m using to update the joint positions: private func updateHandJoints( for handAnchor: HandAnchor, with jointEntities: [HandSkeleton.JointName: Entity] ) { guard handAnchor.isTracked else { hideAllJoints(in: jointEntities) return } // Check if the little finger tip and intermediate base are both tracked. if let tipJoint = handAnchor.handSkeleton?.joint(.littleFingerTip), let intermediateBaseJoint = handAnchor.handSkeleton?.joint(.littleFingerIntermediateTip), tipJoint.isTracked, intermediateBaseJoint.isTracked, let pinkySphere = jointEntities[.littleFingerTip] { // Convert joint transforms to world space. let tipTransform = handAnchor.originFromAnchorTransform * tipJoint.anchorFromJointTransform let intermediateBaseTransform = handAnchor.originFromAnchorTransform * intermediateBaseJoint.anchorFromJointTransform // Extract positions from the transforms. let tipPosition = SIMD3<Float>(tipTransform.columns.3.x, tipTransform.columns.3.y, tipTransform.columns.3.z) let intermediateBasePosition = SIMD3<Float>(intermediateBaseTransform.columns.3.x, intermediateBaseTransform.columns.3.y, intermediateBaseTransform.columns.3.z) // Calculate the midpoint. let midpointPosition = (tipPosition + intermediateBasePosition) / 2.0 // Position the sphere at the midpoint and make it visible. pinkySphere.isEnabled = true pinkySphere.transform.translation = midpointPosition } else { // If either joint is not tracked, hide the sphere. jointEntities[.littleFingerTip]?.isEnabled = false } // Update the positions of all other hand joint spheres. for (jointName, entity) in jointEntities { if jointName == .littleFingerTip { // Already handled the pinky above. continue } guard let joint = handAnchor.handSkeleton?.joint(jointName), joint.isTracked else { entity.isEnabled = false continue } entity.isEnabled = true let jointTransform = handAnchor.originFromAnchorTransform * joint.anchorFromJointTransform entity.transform.translation = SIMD3<Float>(jointTransform.columns.3.x, jointTransform.columns.3.y, jointTransform.columns.3.z) } } I’ve attached both a profiler trace and a video recording from Vision Pro that clearly demonstrate the issue. Profiler: https://drive.google.com/file/d/1fDWyGj_fgxud2ngkGH_IVmuH_kO-z0XZ Vision Pro Recordings: https://drive.google.com/file/d/17qo3U9ivwYBsbaSm26fjaOokkJApbkz- https://drive.google.com/file/d/1LxTxgudMvWDhOqKVuhc3QaHfY_1x8iA0 Has anyone else experienced this behavior? My thought is that there might be some background calculations happening at the OS level causing this latency. Any guidance would be greatly appreciated. Thanks!

I am working on an app that will allow a user to load and share their model files (usdz, usda, usdc). I'm looking at security options to prevent bad actors. Are there security or validation methods built into ARKit/RealityKit/CloudKit when loading models or saving them on the cloud? I want to ensure no one can inject any sort of exploit through these file types.

Visionos26 Enterprise api has the new feature: Shared Coordinate Space, participants exchange their coordinate data by SharedCoordinateSpaceProvider through their own network, when shared coordinate space established with nearby participants, the event: connectedParticipantIdentifiers(participants: [UUID]) will be received. But the Event.participantIdentifier still be an invalid default value(00000000-0000-0000-FFFF-FFFFFFFF) in this time, I wonder when or how I can get a valid event.participantIdentifier, or is there some other way to get the local participantIdentifier? Or If it's a bug, please fix it in later beta release version, thank you.

In a simple test, I'm observing ~30% higher CPU usage with the ARWorldTrackingConfiguration compared to the ARBodyTrackingConfiguration when both configurations have AREnvironmentTexturing enabled. In Instruments, I observe Recon3D consuming ~5.5 seconds of CPU time with the ARWorldTrackingConfiguration vs <0.3 second with the ARBodyTrackingConfiguration in two separate 30 seconds samples. This is on an iPhone 12 Pro equipped with lidar. Is there a reason why two separate configurations, both having the same features enabled would have a different CPU overhead?

I noticed that when I drag the menu window in an Immersive View, the entities behind it becomes semi-transparent, and the boundary between virtual and real-world objects is very pronounced. May I ask how does VisionOS implement this effect? Is there any API or technique I can use in my own code to enable the same semi-transparent overlay - even when I am not dragging the menu window?

The AR based app I am working on right now is experiencing an issue. Sometimes, the AR session fails with a call to my ARSessionObserver's session(_ session: ARSession, didFailWithError error: Error) with the following error: Error Domain=com.apple.arkit.error Code=102 "Required sensor failed." NSLocalizedFailureReason="A sensor failed to deliver the required input.," NSLocalizedRecoverySuggestion="Make sure that the application has the required privacy settings." The underlying error seems to point to the CoreMotion framework: Domain=CMErrorDomain Code=102 "(null) Some people seem to have experienced this issue and solved it by making sure that the Compass Calibration switch is ON in Settings > Privacy > Location Services > System Services. For context, the ARWorldTrackingConfiguration.worldAlignment is set to .gravity The thing is it is already ON when I experience this issue. I also noticed that this issue happens way more often on the iPhone 16e than in any other device. Has anyone had similar experiences? I am looking for a way to prevent this error from happening (ideally) or handling in a way that does not affect the user. Any help is appreciated

Can I apply .scrollInputBehavior(.enabled, for: .look) to a WebView (wrapped UIViewRepresentable) in a visionOS 26 app? I tried it myself, but I couldn't do it, so I would like to know if there is any way to do this. Best regards.

Currently I am using mixed style immersive view to place both my WindowView(plain style) and ImmersiveView content together. The issue is that the rendering depth testing may always let the virtual content block my normal WindowView. Is it possible to manually set windowedVIew always displays in the front of my virtual view in mixed style immersion? (I know modelSortGroup but it doesn't quite fits here) Or if I can dynamically change the .progressive value when the immersive space is open (set the value to zero means .mixed itself right?)

Hello Community, I’m currently working with the sample code “CapturingDepthUsingTheLiDARCamera” and using it to capture the depth map of an image taken with the iPhone 14 Pro. From this depth map, I generate a point cloud using the intrinsic camera parameters. I've noticed that objects not facing the camera directly appear distorted in the resulting point cloud. For example: An object with surfaces that are perpendicular to each other appears with a sharper angle in the point cloud — around 60° instead of 90°. My question is: Is this due to the general accuracy limitations of the LiDAR sensor? Or could it be related to the sample code? To obtain the depth map, I’m using: AVCapturePhoto.depthData.converting(toDepthDataType: kCVPixelFormatType_DepthFloat32) Thanks in advance for your help!

Hello all, I saw this interesting VisionOS app: https://apps.apple.com/us/app/splitscreen-multi-display/id6478007837 I was wondering if there was any documentation on the Swift APIs that were used to create this app.

I want to select a sub model under a large model in a mixed space, and when I select this sub model, I will add a stroke to it, similar to the effect of selecting a model in Reality Composer Pro ，How to create entity strokes similar to this effect

I work on motion capture systems for VTubing. I can't seem to find any information on gaining access to the Face Tracking features on iOS while developing for Vision OS. I would love to bring VStreamer Live to Vision OS

How do I convert a blend shape/morphed 3D lip-synced model into a usdz that will play in AR on an iPhone?

Is there any interest in this forum for those developing for the spatial web and safari. I can't seem to find any posts that are relevant here.

当我进入混合空间时，出现一个模型，但模型后面有一个 windowGroup，无法完全查看。如果我想点击进入 mix 空间，我需要使用代码将 windowGroup 移动到另一个位置，而不是手动移动 ![]（“https://developer.apple.com/forums/content/attachment/0471ead0-4c74-43a7-9ecc-12e67e81cec6” “title=WechatIMG31.jpg;宽度=915;高度=777”）

Hi team, I believe I’ve found a registration issue between ARFrame.sceneDepth and ARFrame.capturedImage when using high-resolution frame capture on a 2022 iPad Pro (6th gen). When enabling high-resolution capture: if let highResFormat = ARWorldTrackingConfiguration.recommendedVideoFormatForHighResolutionFrameCapturing { config.videoFormat = highResFormat } … arView.session.captureHighResolutionFrame { ... } the depth map provided by ARFrame.sceneDepth no longer aligns correctly with the corresponding high-resolution capturedImage. This misalignment results in consistently over-estimated distance measurements in my app (which relies on mapping depth to 2D pixel coordinates). iPad Pro (6th gen): misalignment occurs only when capturing high-resolution frames. iPhone 16 Pro: depth is correctly registered for both standard and high-resolution captures. It appears the camera intrinsics, specifically the FOV, change between the “regular” resolution stream and the high-resolution capture on the iPad. My suspicion is that the depth data continues using the intrinsics of the lower resolution stream, resulting in an unregistered depth-to-RGB mapping. Once I have the iPad in hand again, I will confirm whether camera.intrinsics or FOV differ between the low-res and high-res frames. Is this a known issue with high-resolution frame capture on the 2022 iPad Pro? If not, I’m happy to provide some more thorough sample code. Thanks for your time!