At the Vehicle Display Interface (VDI) 26 conference, Chris Chinnock interviewed Thomas Labardens, Chief Commercial Officer at UVR (United VR), covering the company’s spectral/physically-based rendering simulation software for automotive cockpit and display design:

Company background

• UVR was founded about 10 years ago by two researchers, with a mission to build 3D collaboration software bridging design and engineering teams — combining visually compelling renders with measurable, trustworthy precision rather than just aesthetics.

Core technology: spectral, polarized GPU ray tracing

• Built on standard ray tracing but enhanced to run on GPU and carry full spectral and polarized light information through the scene, rather than the typical approach of most render engines (including game engines), which convert light sources to RGB values early on and lose the ability to accurately simulate wavelength-specific interactions between materials and lighting.
• Not real-time, but highly optimized and considered state-of-the-art for physically accurate rendering; RGB-based approaches aren’t “wrong,” just optimized for different (typically real-time) use cases.
• Spectral resolution is user-configurable (e.g., 1nm, 10nm, 15nm bands) using a stochastic ray-sampling approach rather than computing every nanometer individually; results can be exported as EXR files containing per-pixel spectral data.

End-to-end, measured input chain

• The system aims to address the full value chain rather than leaving gaps that bias results:
  • Ambient lighting: captured via a patented 360° spectral HDRI-capture device.
  • Materials: measured using commercial gonio-spectrophotometers (e.g., from X-Rite) to capture angle-dependent reflectance of leather, plastic, films, etc.
  • Displays: incorporates hemispheric measurement data from device makers like Eldim and Konica Minolta, capturing how viewing angle and coatings affect display appearance.
  • Observer/sensor: supports presets or imported camera data (e.g., an iPhone’s characteristics); human perceptual factors like color blindness are handled as a post-process using a companion tool called PredictLab, since modeling full human visual/neural perception directly in the render is considered too complex.

Output and visualization

• Since the final image must display in RGB, UVR anticipates the target display’s color space to minimize the gap between real-world and rendered perception. HDR displays and HDR export (via EXR) are recommended for high-dynamic-range use cases like simulating headlamp glare.
• Exports available as EXR or PNG; EXR retains full HDR/spectral-derived data for further analysis in PredictLab.
• The software is GPU-based and scalable via render farms (cloud or on-premise); Faraday notes a customer use case achieving real-time rendering even with spectral, polarized, and measured data. A new software version (beta releasing end of September, for Unreal Engine) is described as faster, more interactive, and less power-intensive.

Integration and interactivity

• Not a standalone tool — it’s embedded as a plugin/enrichment layer within existing platforms already used in automotive visualization, including Unity, Unreal Engine, and Blender.
• Interaction today is typically mouse-based navigation within these editors. UVR has also experimented with VR-style experiences (e.g., using a virtual iPad within Unity to “capture” and display their higher-fidelity render), and pre-calculated/preset-view navigation experiences, since true real-time interactivity isn’t achievable with their full spectral computation on typical (1–2 GPU) hardware — though convergence is described as fast even for complex automotive use cases.

Business model

• Revenue comes from software licensing, plus services — particularly measurement services (materials, displays, environment) and sale of their patented spectral capture device.
• A validation service called isophotography is also offered: photographing a subject in a fully characterized, measured light booth and replicating the same setup virtually, then comparing the real and rendered images pixel-by-pixel to validate that virtual materials accurately match real-world appearance — used by some customers to validate their simulated materials.