Mitsuba软件

这就是一个渲染软件，下载网址是http://www.mitsuba-renderer.org/download.html

Mitsuba is aresearch-oriented rendering system in the style of PBRT,from which it derives much inspiration. It is written in portable C++,implements unbiased as well as biased techniques, and contains heavyoptimizations targeted towards current CPU architectures. Mitsuba is extremelymodular: it consists of a small set of core libraries and over 100 differentplugins that implement functionality ranging from materials and light sourcesto complete rendering algorithms.

Mitsuba是以PBRT风格为研究对象的渲染系统，从中获得了很多灵感。 它是用可移植的C ++编写的，实现了无偏差和偏向的技术，并包含针对当前CPU架构的重优化。 Mitsuba是极其模块化的：它由一小组核心库和100多种不同的插件组成，这些插件实现了从材料和光源到完整渲染算法的功能。

In comparison toother open source renderers, Mitsuba places a strong emphasis on experimentalrendering techniques, such as path-based formulations of Metropolis LightTransport and volumetric modeling approaches. Thus, it may be of genuineinterest to those who would like to experiment with such techniques thathaven't yet found their way into mainstream renderers, and it also provides asolid foundation for research in this domain.

与其他开源渲染器相比，Mitsuba非常重视实验渲染技术，例如Metropolis LightTransport的路径配方和体积建模方法。 因此，对于那些想要尝试这种尚未进入主流渲染器的技术的人来说，它可能是真正感兴趣的，它也为这一领域的研究提供了坚实的基础。

The renderercurrently runs on Linux, MacOS X and Microsoft Windows and makes use of SSE2optimizations on x86 and x86_64 platforms. So far, its main use has been as atestbed for algorithm development in computer graphics, but there are manyother interesting applications.

Mitsuba comeswith a command-line interface as well as a graphical frontend to interactivelyexplore scenes. While navigating, a rough preview is shown that becomesincreasingly accurate as soon as all movements are stopped. Once a viewpointhas been chosen, a wide range of rendering techniques can be used to generateimages, and their parameters can be tuned from within the program.

[pic]

Mitsubainstances can be merged into large clusters, which transparently distribute andjointly execute tasks assigned to them using only node-to-node communcation. Ithas successfully scaled to large-scale renderings that involved more than 1000cores working on a single image. Most algorithms in Mitsuba are written using ageneric parallelization layer, which can tap into this cluster-wideparallelism. The principle is that if any component of the renderer produceswork that takes longer than a second or so, it at least ought to use all of theprocessing power it can get.

The rendereralso tries to be very conservative in its use of memory, which allows it tohandle large scenes (>30 million triangles) and multi-gigabyte heterogeneousvolumes on consumer hardware.

 [pic]

Mitsuba supportsthe most commonly used scattering models: Lambertian surfaces, idealdielectrics & mirrors as well as the the Phong & anisotropic WardBRDFs. A range of microfacet models are also available, including rough glass,plastic & metal. Subsurface scattering can either be simulated using aBSSRDF approach, or more rigorously using volumetric light transport.

For volumes, thesupported scattering models are isotropic, Henyey-Greenstein, Kajiya-Kay fiberscattering and micro-flakes.

  
[pic]  

 

 

A wide range ofrendering techniques are available, including:

·        Ambient occlusion

·        Direct illumination

·        Monte-Carlo path tracer which solves thefull Radiative Transfer Equation

·        Photon mapper with irradiance gradients

·        Adjoint particle tracer

·        Bidirectional path tracer

·        Instant Radiosity (hardware-accelerated)

·        Progressive Photon Mapper

·        Stochastic Progressive Photon Mapper

·        Path Space Metropolis Light Transport

·        Primary Sample Space Metropolis LightTransport

·        Energy redistribution path tracer

[pic]   

 

Mitsuba cancompute global illumination solutions in scenes containing large isotropic oranisotropic participating media. The underlying volumes can be represented assparse voxel octrees or as hierarchical grids, where grid cells are directlymapped from files into memory.

  [pic]

 

Mitsubainternally uses a O(n log n) SAH kd-tree compiler with support for primitiveclipping (aka. perfect splits). The ray tracing core is built on Havran's fasttraversal algorithm. On Intel platforms, it is possible to trace coherent rayspackets using SSE2. Mitsuba supports analytic shapes such as cylinders andspheres and makes use of additional SSE2 accelerations when working withtriangle meshes, which allows it to intersect up to 4 triangles at a time.

 

Data exchangewith the major modeling packages is supported using the COLLADA file format.Mitsuba can read DAE files andconvert them into its native XML-based file format. In the future, additionalintegration is planned, e.g. plugins for Maya and 3ds max.

 

Other featuresare:

·        Spectral rendering, black body radiationand dispersion

·        Customizable image reconstructionfilters

·        High dynamic-range input/output usingthe OpenEXR format

·        Deterministic Quasi-Monte Carlo sampling

·        Adaptive integration

·        Depth of field

 

图片渲染软件LuxRender

LuxRender 是一款开源免费的图片渲染软件，功能一点儿也不输收费软件，提供物理真实的、无偏差的图片渲染功能，它是根据渲染方程来模拟光的传输，生成物理真实的图像

LuxRender 需要其它3D程序来创建包含材质设置和相机设置的场景，然后通过导出脚本，将场景的所有信息都写入一个LuxRender文件格式的文件中

用LuxRender打开这个文件后，要做的事情只是渲染这个场景。LuxRender基于PBRT项目，但不同之处在于它关注的是产品渲染和艺术效果而非学术和科学目的

LuxRender包含一个交互式的用户界面和多个不同的3D程序的导出脚本，包括Blender，Maya和XSI。提供了命令行工具并且还提供 C/C++ API 以使LuxRender能被集成到第三方应用程序

LuxRender自0.8版本开始提供OpenCL加速渲染功能。

能否详细比较一下LuxRender、Cycles、Mitsuba这三款渲染器的优劣？

有哪些可以用a卡加速的渲染器？

找支持OpenCL的渲染器呗。开源软件Blender的GPU渲染貌似支持A卡。

Furryball渲染器

编译Luxrender

http://www.cppblog.com/chaogu/archive/2011/05/03/145603.html?opt=admin

渲染引擎对比：Cycles和其他

http://www.blenderget.com/render-engine-comparison-cycles-vs-giants/

http://www.boyunjian.com/do/article/snapshot.do?uid=5021297810389985048

vary2.3_for_3dmax2013english64bit.zip   版本要匹配！

参考列表：

https://blog.csdn.net/liqiang981/article/details/52403259

https://blog.csdn.net/mangobar/article/details/50910608