We are starting a new project that needs to run on Meta Quest 3, Vive, and potentially PSVR2. Is OpenXR finally at a point where we can ditch vendor-specific SDKs like Oculus Integration? I’m worried about losing access to specific hardware features like hand tracking or eye tracking if we go with a more generic standard. What’s the consensus for 2024?
3 answers
OpenXR has become the industry standard for a reason, and I strongly recommend it for any new project in 2024. Most major hardware features, including hand tracking and passthrough, are now supported through OpenXR extensions. By using the XR Interaction Toolkit (XRI) in Unity or the OpenXR plugin in Unreal, you write your input logic once and it maps correctly across devices. The "vendor lock-in" of the past is fading. While you might still need a small shim for a very specific Meta-only feature, the time you save on maintenance for a multi-platform build is massive. It allows your team to focus on the actual user experience rather than rewriting input wrappers for every new headset that hits the market.
Have you checked the latest compatibility layers for Eye Tracking in OpenXR? Last I checked, some PSVR2 specific haptics still required a bit of native code.
We switched to OpenXR mid-development and it cut our QA time in half. No more debugging three different input systems for the same grab mechanic.
I agree with Jessica. Input Mapping is the biggest headache in VR, and OpenXR solves it by focusing on "actions" rather than specific buttons.
Steven, you're right about the haptics. Mark, for the haptic triggers on PSVR2, you’ll likely need the specific Sony wrapper. However, you can still keep 90% of your codebase in OpenXR and just use a conditional compile for those specific sensory features. This "hybrid" approach is what we used for our last enterprise training sim, and it worked flawlessly across Quest and Vive Pro.