PCVR Development Dilemma - Seeking Community Guidance

Optimisation is quite an art and requires a good understanding of the render pipeline. You don't mention which engine you're developing in which makes it tricky to give specific advice. In broad terms, you should be looking at

• reducing the complexity of your shaders,
• atlas your textures so you can fewer materials and share them across multiple meshes (allows them to be batched),
• bake your lights if in forward rendering (dynamic lights and shadows are expensive),
• use occlusion culling (in Unity this has to be baked, in Unreal it comes by default),
• mesh combine materials with the same material but do so in a way that still allows occlusion culling to work its magic (if you combine everything into a huge mesh, then if only part of that mesh is visible to the camera then all of its verts are computed),
• keep vert counts sensible (use LODs, optimise meshes so they're game-ready. This includes good modelling best practice like not having non-manifold geometry, having good lightmap UVs, etc. If you're buying off the shelf environment assets you'll find they can often be quite badly modelled!)
• use GPU instancing and static batching as appropriate
• skinned meshes (character avatars) are expensive
• post processing (although doable on PCVR) are expensive. Choose those full screen effects carefully
• Be careful with overdraw (transparencies and render order) Draw calls are expensive even on PCVR

For compatibility across different PCVR hardware... You have no control over your end user's PCs so decide on your minimum tagret platform and test on that. For headsets, this also requires a lot of testing (different Inputs can be especially difficult) but use OpenXR as your backend to make things more widely compatible. And don't use features from platform specific SDKs.

In terms of community resources, I'd highly recommend the Oculus developer docs. Start here for PCVR: https://developer.oculus.com/documentation/unity/unity-perf/ but do also have a look at the standalone (Quest) recommendations.

The above articles mention some profiling tools. Generally your engine's built-in tools should be adequate if you don't have that much experience. If you want to go deeper then RenderDoc is also incredibly useful for frame debugging. I can't overstate this enough - learn to profile! Find where your bottle neck lies and optimise for that. Are you CPU or GPU bound? Optimising for one can be detrimental to the other. This is for Unity mobile platforms but is still incredibly valuable advice: https://blog.unity.com/technology/optimize-your-mobile-game-performance-tips-on-profiling-memory-and-code-architecture

For application debugging, that's just game dev stuff. In VR it's useful to have an in-app console or some for of on-screen logging so you can see what's going on without having to take off the headset to look at the engine's output console. The Meta XR Simulator looks useful for this too, but I've never used it (only just heard about it at the recent Connect event).

Re. input options to make the experience more immersive. Floating UI panels are generally considered to be immersion breaking. (think laser pointer being used to highlight and click on buttons) You might instead consider diegetic UI, where buttons and interface elements exist within the world and are part of that world. (this is a brief intro to this concept https://uxdesign.cc/vr-diegetic-interfaces-dont-break-the-experience-554f210b6e46 )

Good luck on your journey. Hopefully you can now appreciate what you were paying that consultancy firm for 🙂