Post Magazine

January 2014

Issue link:

Contents of this Issue


Page 49 of 51

review Chaos Group's V-Ray 3.0 O By Toni Bratincevic 3D Artist Newport Beach, CA V-Ray's evolution includes a simplified UI. X I created this image while testing the new V-Ray 3.0. 48 ne of the most innovative, fast and stable raytracers today, V-Ray has become a synonym for quality. It's responsible for popularizing and pushing what was once a slow and not production-ready algorithm (raytracing) into one of the most popular ways of delivering high-quality renders today. Although Version 2.4 was everything one could have wished for, with Version 3.0 just around the corner, Chaos Group continues on a trajectory of delivering upgrades to one of the best raytracing engines. While introducing some new features, V-Ray 3.0 also delivers a simplified workflow and focuses on optimizing and improving its own raytracing core. One of the most important new features is the introduction of the new progressive render engine. Users now have the ability to choose between old image samplers (DMC, Adaptive Subdivision, Fixed), which were all based on bucket rendering, and the new progressive renderer, which refines the image until it reaches a certain time or quality limit. In V-Ray RT users were limited to brute force global illumination, which can sometimes lead to slow render times. With the new progressive renderer, users can have a choice of (for example) using Irradiance cache and Lightcache as global illumination engines, delivering much faster results while still having the ability of seeing the whole image refining as it is rendering. In addition to this, raytracing speed is improved, especially in cases where brute force global illumination is used. Optimizations were done on the V-Ray raytracing core, with additional choice to use the embree raytracer, which can speed up rendering with specific scenes. As in any production pipeline, rerendering and tweaking shaders, textures and lighting requires constant rerendering, sometimes even 100 times before getting a satisfying result. In the old V-Ray, pros usually used rectangular selection over a rendered image to set borders for rerendering, but in V-Ray 3.0, with the introduction of image masks, this workflow becomes much more efficient. As a mask, users can decide to use selected objects, layers, textures or define a custom objects include/exclude list. Assuming that you have already done a test render of the whole scene, by selecting one object in your viewport and setting Render Mask to Selected, clicking render again will tell V-Ray to only rerender the area of the image covered by the selected object, while leaving old render intact. Post • January 2014 Support for Alembic file format, OSL shaders, deep images and Open EXR 2.0 are very important steps for popularizing V-Ray in high end movie work. Deep images store information for all individual image samples taken during rendering, including their depth values, and allow more correct post processing effects like DOF and motion blur without getting artifacts specific to old image formats. With support for OSL shaders, users are given flexibility to extend V-Ray shading horizons, making it even more powerful and a good choice for high-end production where it is common practice to write specialized shaders. In combination with the Alembic format, new VRMat functionality will help close connectivity gaps between 3D apps. V-Ray VRMat materials defined in 3DS Max can be exported and merged into Maya without any additional scripts or apps. In time, this will open up the possibility of creating a library of materials that are application independent and can be imported into any available V-Ray platform. The user interface has been simplified for new users. There is a choice now of Basic, Advanced and Expert UI, which allows users with different levels of experience to modify and simplify the UI based on their needs. In Basic UI, only important options are exposed, which for artists new to V-Ray, can be a big plus in dealing with the complexity of a new render engine. Hair rendering is one of the most complex situations to solve when using raytracing render engines since it includes a lot of thin strands of hair requiring many samples to get satisfying and noise free result. With old image sampling methods, like the V-Ray Adaptive subdivision sampler, resolving noise in hair could have led to very long render times. To solve these problems the Adaptive DMC sampler was a much better and faster choice, and with hair shading optimizations and faster raytracing in V-Ray 3.0, hair rendering speed is noticeably improved, sometimes delivering renders three times faster than Version 2.4. In V-Ray 2.4, the subsurface shader was limited to screen space density estimation of subsurface samples, which sometimes led to an unstable subsurface solution manifested as flickering. V-Ray 3.0, however, addresses these issues and allows users to decide between three different ways of calculating a subsurface effect. The old method is still there, but two new methods are introduced. The first one is a fully raytraced SSS effect, which produces the best result you can get VITAL STATS MANUFACTURER: Chaos Group PRODUCT: V-Ray 3.0 WEBSITE: PRICE: V-Ray 3.0 for 3DS Max user license plus one Render Node: $1,050; Single V-Ray Render Node: $350 · stable with improved speed · simplified UI for new users but takes more time to render. Since it is brute force, there is no prepass and the solution can't be saved for reuse. This brute force method is becoming extremely popular even in high-end production because of very predictable results. Object method is based on a fixed number of samples and is not camera view dependent as in the prepass method, which means as an object moves away from the camera it will keep the same number of samples, reducing flickering in many situations. Managing data when using V-Ray distributed rendering can sometimes lead to unexpected results and problems like missing textures and unresolved paths, but with V-Ray 3.0 these issues are completely resolved by allowing nodes in a distributed rendering network to collect all data needed for rendering locally. In local networks, with relatively slow speeds, this can lead to some speed improvements because some files are copied and stored locally instead of on a network. Although there are a lot of new features in V-Ray 3.0, most of them are evolutionary steps over Version 2.4. As expected, Chaos Group did an amazing job again, but it is up to users in the end to decide if it is an update worth upgrading to. I know for sure that the progressive image sampler alone made my workflow much faster, efficient and more fun, and considering how many free upgrades V-Ray 2.x got from 2.0 to 2.4 version, I am sure 3.0 will be a great investment for many of us.

Articles in this issue

Links on this page

Archives of this issue

view archives of Post Magazine - January 2014