Advanced Rendering

Mental Ray
Rendering Imagination Visible
Image courtesy of Amaru Zeas
Render Your Imagination.
We're taking big steps to evolve the mental ray rendering technology and workflows within Autodesk Maya and 3ds Max.

In 2007, Mental Ray was the first commercial solution to provide a physically based rendering workflow with the new materials and lights added to Maya and 3ds Max. In 2016, that tradition continues in Mental Ray 3.14 with support for the NVIDIA Material Definition Language (MDL).

Just as important, Mental Ray 3.14 adds GI-Next - a new global illumination method that is far easier to control, 2X as fast as previous methods of similar quality, and GPU-accelerated for 2-20X faster again from a single GPU.

Mental Ray 3.14 Features

The latest features shipping with Mental Ray 3.14 and integrated into Autodesk 3dsMax 2018 and Autodesk Maya 2016 and 2017 include:


New GI Next Global Illumination Engine

New Global Illumination solution that is easy to use, fast and physically accurate

  • New GI Next engine with full support of Mental Ray shaders, motion blur, depth of field, hair, volumes and particles
  • Easily controlled by a single "quality" parameter.
  • Advanced quality settings for fine control
  • Up to 2X faster than Final Gather Force GI
  • Over 10X faster again by adding a GPU vs. a 14-core CPU
  • Only geometry is loaded into GPU memory, allowing texture-heavy scenes to be GPU accelerated.
  • Easy to adopt for any scene set-up up to use GI by choosing the new method.

Performance Scaling

NVIDIA® Mental Ray® supports any number or combination of GPUs and/or CPUs running within the same machine, and gives you control over which are being used for rendering. Load balancing between GPUs of different sizes, types, and generations are automatic – Even older systems transform into powerful, cost effective rendering nodes by adding NVIDIA GPUs.

Performance Scaling

NVIDIA Material Definition Language

Render physically based materials in NVIDIA's MDL format

  • Load MDL materials directly for use in mental ray
  • Freely access +1,000 materials from the NVIDIA vMaterials library for accurate starting points to your own custom needs.
  • MDL material capabilities vary, with many having the power and flexibility of traditional Mental Ray shaders.
  • Supports all Mental Ray effects like a traditional shader
  • Reuse MDL materials with other MDL-compliant renderers like NVIDIA® Iray®

Light Importance Sampling by Default

Modern lighting setups rendered far more quickly by default

  • Light Importance Sampling (LIS) optimization is now enabled by default
  • Efficient sampling of area lights, light cards, emissive geometry, image-based lighting allows lights to be added without degrading performance
  • Lighting quality is achieved faster and with less noise
  • Very feasible to use with thousands of traditional lights
  • Built-in logic and heuristics apply LIS based on benefit and quality balance.

New Alembic Features

Fuller support for production pipelines

  • Alembic "curves" can now be loaded as Mental Ray "hair"
  • New subdivision control added for higher fidelity of tessellation
  • Per-face materials assignments now stored as Alembic facesets

OpenEXR 2.2

OpenEXR 2.2 for better performance and more features

  • Now in Mental Ray core as a primary multi-channel image format
  • Render layers can be stored as "multi-part" .exr for faster loading in comp
  • Backwards compatible to OpenEXR 1.x when required
  • OpenEXR "deep" pixel data generated by Mental Ray for 3D compositing
OpenEXR 2.0 OpenEXR 2.0 OpenEXR 2.0 OpenEXR 2.0 OpenEXR 2.0

Ambient Occlusion Pass, GPU Accelerated

Ambient Occlusion computed on GPU “for free” while rendering beauty on CPU

  • Ambient occlusion render pass works with any scene setup and shaders
  • Easy to enable without touching the scene content
  • Automatic fallback to CPU produces pixel identical results

Layering Shaders (MILA) Improved

The layering shaders, feature enhanced and optimized

  • New: Support for new Light Path Expressions (LPE) machinery in Mental Ray
  • LPE render pass improved workflow

Native UV Tiling

Common UV Tiling schemes directly supported for best efficiency

  • Mental Ray UV tiled textures at runtime
  • Supports most common UV tiling conventions, e.g. UDIM
  • Takes advantage of Mental Ray's texture loading and caching pipeline
  • Guarantees optimal performance and memory consumption

"Deep" Data

Compute "deep" information during rendering, rasterizer or ray tracing

  • Deep data computation enabled per render pass / frame buffer
  • Deep information saved in OpenEXR "DeepTile" format
  • Works in rasterizer and ray tracing rendering modes
  • Supports existing shaders in third-party applications

Custom Color Profiles

Easily customizable Color Transformations

  • New type of "color profile" shader introduced
  • Fits into existing color management pipeline and scene description
  • Perfectly suited for integrated Mental Ray color managements applications
  • Support color pre- and post operations with least impact on performance

Mental Ray Core Features

  • Mental Ray core now accessed as a complete individual dynamic library
  • CPU systems of any core count now fully supported with hyper threading
  • Allows Mental Ray to be updated independently of the host application
  • Provides full support large, modern multi-CPU systems