Oh My Galaxy VR/MR Development Analysis: nDreams' Tech Deep Dive

Most VR developers treat mixed reality like a side quest. They build for virtual worlds, then awkwardly shoehorn in some passthrough features and call it spatial computing. nDreams took the opposite approach with Oh My Galaxy: they built the transition itself as the main feature. The result is a Meta Quest title that doesn't just switch between VR and MR modes but makes that switch feel intentional, smooth, and technically impressive enough to make other developers take notes.

The Reality-Switching Problem Nobody Talks About

Switching between virtual and mixed reality isn't just flipping a passthrough toggle. Your spatial understanding changes, your performance budget shifts, your lighting model breaks, and suddenly players are staring at floating UI elements in their actual kitchen. Most games handle this transition like a loading screen with extra steps. Oh My Galaxy treats it like a core mechanic that needs to work flawlessly every single time.

The technical challenge starts with spatial mapping. VR games can fake physics and collision because players can't see the real world anyway. MR games need to understand actual room geometry, furniture placement, and lighting conditions that change throughout the day. nDreams solved this by implementing what they call "dual-space rendering," where the game maintains both virtual and real-world spatial maps simultaneously, updating the real-world data in the background even while players are in full VR mode.

This approach demands serious optimization. Running spatial mapping, object tracking, and full game logic simultaneously would normally crater performance on mobile VR hardware. The development team had to build custom occlusion systems that selectively render only the virtual objects that wouldn't be blocked by real-world geometry, reducing overdraw by up to 40% during MR segments.

Spatial Mapping That Actually Works in Real Rooms

Here's where most MR implementations fall apart: they assume players have empty rooms with perfect lighting and white walls. Real people have cluttered spaces, weird furniture arrangements, and lighting that ranges from "dungeon" to "dentist office." Oh My Galaxy's spatial mapping system accounts for these variables through what nDreams calls "adaptive mesh refinement."

The system starts with Meta's standard room mapping but continuously refines the spatial mesh during gameplay. Instead of relying on the initial room scan, the game uses computer vision to identify surfaces, obstacles, and play areas in real-time. When players move between VR and MR modes, the transition system already knows exactly where virtual objects can safely appear without clipping through real furniture or floating awkwardly in mid-air.

The lighting integration deserves special attention. VR games typically use baked lighting or simple dynamic systems optimized for consistent virtual environments. MR games need to match real-world lighting conditions or risk breaking immersion completely. nDreams implemented a hybrid lighting system that samples real-world illumination through the headset cameras and adjusts virtual object shading to match. The result is virtual game elements that look like they actually belong in your physical space instead of obviously floating computer graphics.

Developers can apply these techniques beyond games. Any MR application that needs to blend virtual content with real environments faces the same core challenges: spatial understanding, performance optimization, and visual coherence. The dual-space rendering approach could work for productivity apps, training simulations, or collaborative tools that need to switch between focused virtual work and real-world context.

Performance Optimization for Dual-Reality Rendering

Running both VR and MR rendering pipelines simultaneously sounds like a recipe for thermal throttling and dropped frames. The Quest's mobile processor wasn't designed to handle dual-space rendering, computer vision processing, and game logic at 90fps. nDreams had to get creative with their optimization strategies, and the solutions they developed offer practical lessons for any spatial computing project.

The key insight is selective processing. Instead of running full spatial mapping continuously, the system identifies "stable" areas of the room that don't need constant updates and focuses computational resources on dynamic regions where changes are likely. This reduces the spatial mapping overhead by roughly 60% compared to continuous full-room scanning.

For rendering optimization, the team implemented what they call "contextual LOD" (level of detail). Virtual objects use high-detail models when players are in VR mode but automatically switch to optimized versions during MR transitions. The LOD system accounts for real-world occlusion, so objects that would be partially hidden behind furniture use simplified geometry for the hidden portions.

The audio system required similar dual-space thinking. VR audio can be completely synthetic, but MR audio needs to blend with real-world acoustics. Oh My Galaxy implements spatial audio that adapts to actual room acoustics measured through the headset microphones, adjusting reverb and occlusion in real-time as players transition between modes.

Development Methodology for Mixed Reality Games

Building games that work equally well in VR and MR requires rethinking fundamental development practices. nDreams discovered that traditional VR development workflows break down when you add real-world constraints. They had to develop new testing methodologies, design validation processes, and debugging tools specifically for dual-reality games.

The testing challenge is immediately obvious: VR games can be tested in controlled virtual environments, but MR games need to work in countless real-world scenarios. The team built a "synthetic room" testing system that simulates different room layouts, lighting conditions, and furniture arrangements digitally. This allows developers to test MR functionality without physically setting up dozens of different real-world environments.

Design validation becomes more complex when virtual game elements need to coexist with unpredictable real-world spaces. nDreams developed design guidelines that account for common furniture heights, typical room dimensions, and standard lighting scenarios. Game mechanics that work perfectly in virtual space might be impossible in a small apartment or break completely in rooms with unusual layouts.

The debugging tools had to evolve as well. Traditional VR debugging shows you what's happening in virtual space, but MR debugging needs to visualize the interaction between virtual and real elements. The team created visualization tools that overlay spatial mesh data, lighting samples, and occlusion calculations directly in the headset, allowing developers to see exactly how the dual-space systems are interpreting real-world environments.

What This Means

for Spatial Computing Development Oh My Galaxy proves that seamless VR-to-MR transitions aren't just technically possible but can be implemented efficiently enough for consumer hardware. The techniques nDreams developed solve real problems that every spatial computing developer faces: spatial understanding, performance optimization, and visual coherence between virtual and real elements.

The dual-space rendering approach offers a practical framework for any application that needs to work across VR and MR modes. Instead of treating these as separate experiences that happen to run on the same hardware, developers can design for seamless transitions as a core feature. The optimization strategies become essential as more applications need to handle complex spatial computing tasks on mobile processors.

For developers entering spatial computing, Oh My Galaxy demonstrates that the technical challenges are solvable with current hardware and development tools. The game's success suggests we're moving beyond the experimental phase of MR development into practical implementation strategies that work in real consumer environments. That's the foundation the entire spatial computing ecosystem needs to build on.