This paper presents VR Hands Recording, a Unity-based system designed to simplify the recording of hand motion animations in virtual reality environments. Built on the OpenXR standard, the system ensures broad compatibility with various hand-tracking-enabled headsets, including Meta Quest and Pico devices. Users can intuitively record hand gestures through a preconfigured Unity scene and export animations in FBX format for seamless integration into 3D graphics software such as Blender. Compared to existing solutions, VR Hands Recording ofers a flexible, cost-efective, and cross-platform alternative suitable for creative production, human-computer interaction research, and XR prototyping. The system has been evaluated both quantitatively and qualitatively, demonstrating high usability, precision, and ease of integration into post-production pipelines.
Several solutions have been proposed for recording hand movements in virtual reality, each ofering diferent trade-ofs in terms of compatibility, licensing, and integration flexibility. Among the most notable are the following. This open-source project 2, available on GitHub, allows users to record hand tracking data using Meta Quest headsets (Quest 2, Quest 3, and Quest Pro). The system supports FBX export and is designed specifically for Unity environments.
However, the main limitation of this solution lies in its dependency on Meta’s proprietary SDK and lack of support for non-Meta devices. Furthermore, it is not based on the OpenXR standard, reducing its portability across XR ecosystems.
Glycon3D 3 is a commercial hand and body motion capture solution compatible with multiple platforms, including Meta Quest, SteamVR, and HTC Vive. It supports FBX and BVH export formats and ofers a professional-grade recording environment.
While highly capable and flexible, Glycon3D requires a commercial license (starting at $99), which may not be accessible for small teams, educational use, or prototyping scenarios.
The VR Hands Recording system is designed to provide a seamless and modular framework for capturing hand movements in virtual reality. It relies on widely adopted tools and standards to ensure ease of use, cross-device compatibility, and integration with post-production pipelines.
The system is implemented using the following software components: • Unity: version 2022.3.49f1 (LTS), serving as the development and execution environment. • XR Framework: OpenXR (via the Mixed Reality Toolkit), ensuring broad compatibility across XR devices. • Unity Packages. Recorder: used for capturing animations directly within the Unity Editor. FBX Exporter: enables export of recorded animations to FBX format, compatible with 3D graphics software. UnityGLTF: enables export FBX file for a more wide compatible GLB/GLTF file https://github.com/KhronosGroup/UnityGLTF
This software stack allows users to record animations with minimal setup while maintaining flexibility for later customization.
The system supports a variety of headsets equipped with hand tracking capabilities, including: • Meta Quest 2 • Meta Quest 3 / 3S • Meta Quest Pro • Pico (any model with hand tracking support)
All headsets must be connected to a PC (e.g., via Meta Quest Link or equivalent methods if used on another headset) to run the Unity-based recording environment.
This architecture allows the system to operate across multiple hardware configurations while maintaining consistent performance and accuracy in gesture capture. 2https://github.com/richdrummer33/Meta-Quest-Hand-Tracking-Motion-Recording 3https://www.glycon3d.com
This section describes the general workflow available on the github repository 4 and the architectural choices behind the VR Hands Recording system. The goal is to provide a lightweight and flexible tool for capturing hand motion data in virtual reality environments, while ensuring compatibility with industry-standard 3D tools and diverse hardware platforms. We first present the operational workflow designed for end users, followed by the rationale behind the selection of technologies. We then detail the post-processing pipeline, and conclude by discussing technical challenges encountered during development and the corresponding solutions.
The system was designed to minimize user intervention while maintaining flexibility. Upon launching the Unity scene, the user is presented with a preconfigured XR rig with hand tracking and a visual mirror for live feedback (see Fig. 1). The recording process involves selecting the rig in the Recorder panel, pressing Play, performing hand movements using a compatible headset, and stopping the recording to export the animation in FBX format.
At project startup (see Fig. 2), the Unity environment presents the user with a preconfigured scene that includes : • an XR rig with hand tracking enabled; • a virtual mirror to visualize hand movements in real time; • the Recorder panel for managing the capture process.
• Select the FBX option in the Recorder Panel (see Figure 4). 4https://github.com/theTMO/VR-Hands-recording • select the XR rig in the Recorder panel (see Figure 5). • Press Play to start recording. • Move your hands while wearing the headset. • Press Stop to complete the capture.
• The generated file will be exported in .fbx format using the FBX Exporter package.
To ensure compatibility with software such as Blender, Maya, and 3ds Max, the use of UnityGLTF is recommended: • it converts Unity .fbx files to formats such as .GLB, .GLTF; • it preserves animations, materials, and skeletal hierarchy; • it is free and available at: https://github.com/KhronosGroup/UnityGLTF
The following settings are recommended when importing the animation into Blender: • Shading: Use Normal Data • Lighting Mode: Standard • Merge Material Slot: Enabled • Pack Images: Enabled • Bone Dir: Fortune (may look better, less accurate) • Import Scene as Collection: Enabled • Select Imported Objects: Enabled • Import Scene Extras: Enabled
These parameters help prevent issues related to scale, rotation, or data loss (see Fig. 6).
Unity was selected for its broad compatibility with major VR headsets on the market, particularly those based on the Qualcomm XR2 chipset (such as Meta Quest 2/3 and Pico Neo 3/4). It also provides native integration with recording and exporting plugins.
OpenXR is an open standard supported by all major VR hardware manufacturers. This ensures compatibility not only with Meta Quest, but also with other headsets that ofer hand tracking support.
FBX was chosen as the export format due to its versatility and wide compatibility with most 3D animation software (including Blender, Maya, Unity, and Unreal). The subsequent conversion to glTF/glb reduces file size and ensures modern, web-friendly interoperability.
Minor technical challenges were encountered during development. In particular, unnecessary controllers and skeleton structures were sometimes included in the exported animation files. This was resolved by simplifying the Unity rig and disabling non-essential components prior to export.
Additionally, hand tracking quality was found to vary depending on the headset used (see Fig. 7). Devices such as the Meta Quest Pro and Pico 4 demonstrated high tracking accuracy, thanks to their support for the most recent version of the OpenXR Hand Tracking Interaction.
Exported hand motion data in FBX format are post-processed using Noesis or imported into Blender using a specific configuration to preserve skeletal hierarchy and animation quality. Recommended import settings include enabling bone merging, normal data shading, and scene extras. This ensures visual consistency and reduces errors in scale and rotation.
During the development of the system, several minor issues were encountered: • Unnecessary controllers or skeletons in exported files. This was addressed by simplifying the Unity rig and disabling non-essential components before exporting the animation. • Noisy hand tracking. The quality of hand tracking heavily depends on the headset used. Devices such as the Meta Quest Pro and Pico 4 produced highly accurate results thanks to support for the latest version of the OpenXR Hand Tracking Interaction Profile 5.
Tracking Accuracy. The system’s accuracy depends on both the headset used and the environmental conditions. With the latest implementation of OpenXR (version 1.0.27+), the system is capable of tracking up to 26 joints per hand with millimeter-level precision, supporting pinching, grasping, and coherent 3D positioning 6.
Recording Time. Recording is immediate and real-time. Users can start and stop the recording process using the Unity Recorder panel with a single click.
Export/Import Performance.
• FBX file generation: 1–3 seconds, depending on animation length and hardware performance. • Conversion to .glb: Instantaneous using Noesis or Blender.
User Testing. The system was evaluated by digital creatives, designers, and XR developers. Feedback
highlighted the tool’s ease of use, smooth integration with Blender, and its versatility for generating
animated content. Its efectiveness is further supported by previous XR-based training applications
employing similar hardware and frameworks [
Usability. The familiar Unity interface, combined with the in-scene virtual mirror, provides users with direct visual feedback of their hand movements, efectively reducing the learning curve. 5https://registry.khronos.org/OpenXR/specs/1.0/man/html/XrHandTrackerEXT.html 6https://registry.khronos.org/OpenXR/specs/1.0/man/html/XrHandTrackerEXT.html
• Creators of animated content for web, virtual reality, and interactive storytelling • XR game designers and immersive application developers • Researchers in the field of human-computer interaction • Educational or artistic teams aiming to prototype realistic hand movements
VR Hands Recording provides an accessible, flexible, and cost-efective solution for capturing hand movement animations in virtual reality environments. Built on Unity and the OpenXR framework, the system supports a wide range of headsets, including Meta Quest and Pico devices, ofering broad compatibility and future-proof integration.
Its intuitive workflow allows users to quickly record hand gestures, export them in FBX format, and import them into popular 3D graphics tools such as Blender for further editing and rendering. Compared to existing solutions, VR Hands Recording stands out for its: • Cross-platform compatibility enabled by OpenXR; • Ease of use through pre-configured Unity scenes; • Direct FBX export with minimal post-processing; Free availability, making it suitable for small teams, educational settings, and research environments.
This system is particularly well-suited for creative industries, multimedia production,
humancomputer interaction studies, and AI-driven XR applications suggesting future directions such as
co-creation of therapeutic XR spaces via AI integration [