Here is a summary of some of the new features and feature improvements in version 3.11 of mental ray. Please refer to the release notes for more details and for other changes which are not mentioned here.
The iray version 3.0 is a major update of this rendering mode built into mental ray. It adds support for the latest NVIDIA GPU architecture called Kepler, while retaining compatibility to previous generations of GPUs.
This version of mental ray offers a NVIDIA GPU accelerated version of brute-force ambient occlusion computation. Once enabled, a final ambient occlusion pass is computed on the GPU if present, without using any approximation like a point cache. In addition to the general speed up expected by using the massively parallel compute architecture of a GPU, this GPU pass comes almost for free since it can be calculated in parallel to any regular rendering done on the CPU. The AO GPU pass can be enabled and controlled with string options and on the command line of standalone mental ray.
Modern image-based lighting workflows employ area lights with emission controlled by high-dynamic-range textures, like light cards. The new light importance sampling technique significantly improves the rendering quality in such situations. Once enabled it forces to sample "planar" area lights according to the local intensity of the light area surface, in other words, driven by the texture map. The "volumetric" area light types like sphere will also benefit from this mode since their shape is taken into account when determining light "importance". This mode can be controlled with string options. It is disabled by default.
Different importance sampling techniques can be combined so that rendering may converge even faster. The multiple importance sampling mode can take advantage of materials that have BSDFs attached and can couple it with light importance sampling to achieve final quality in less, sometimes much less render time, depending on the lighting situation. This mode can be enabled with a string option. It is disabled by default.
mental ray adds support for multiple motion transformations, to handle common non-linear motion blur cases more efficiently that do not involve deformations (multiple motion vectors). For each motion step within the shutter interval a separate motion transformation matrix can be specified to define the transformation at that sub-frame time. During rendering, the kernel applies a final transformation to geometry and camera that is interpolated between the two nearest final transformations defined in the shutter interval for that scene element.
A new shader package layering comes with mental ray. It introduces a new way of material compositing by layering base components, like illumination models and reflection properties. Commonly used mental ray shaders like architectural material or subsurface scattering are also provided as standard base layers in this package.
The following changes were made in the .mi scene description syntax:
"ambient occlusion gpu" on|off "ambient occlusion framebuffer" "fb_name" "ambient occlusion min distance" min_dist "ambient occlusion max distance" max_dist "ambient occlusion falloff" falloff "ambient occlusion rays" rays "ambient occlusion gpu passes" passesThe object/instance flags
visible
, shadow
and
transparency
are respected to produce intuitive AO results
that are similar to the behavior of "shadows". If a capable GPU is not
detected then a CPU fallback is used that renders according to mental ray's
sampling pattern that is close but not pixel identical to the GPU one.
"light importance sampling" "on"|"off"|"always" "light importance sampling variance" tolerance "light importance sampling precomp" on|offIf
"on"
then importance sampling is enabled for rectangular and
disc "planar" area lights if high emission variance was detected,
and for all spherical, cylindrical and object "volumetric" lights
in any case. If set to "always"
then importance sampling is
always enabled on all area lights. The tolerance to variance can be adjusted
with an option. The default tolerance is 0.001,
so that "planar" lights without texture attached are not considered for
importance sampling because no quality gain is expected and computational
efforts can be avoided. The "precomp"
option controls if the
pre-computed light emission data is also used during rendering. This can
gain performance by reducing light shader calls, but works correctly only
with well-behaving light shaders with low variance and without side effects,
like directional emission dependencies.
"multiple importance sampling" on|off
Copyright © 1986, 2015 NVIDIA ARC GmbH. All rights reserved.