V-Ray 3.0 In Review
June 30, 2014 12:20 am
This time I'm going back into the pre-rendered images realm to write about the latest photorealistic renderer by Chaos Group, V-Ray 3.0. V-Ray is used a lot in architectural rendering, product design and animation because of its unsurpassed photorealism. This is the first time I have used V-Ray, so this article will be very different to others I’ve written, because I am not familiar enough with the software to do a fair comparison between version 3.0 and any previous version.
One of the first things you'll notice when you start using V-Ray is that the renderer is complex. V-Ray lets you fine-tune many different parameters to get the results you want. However, all that fine-tuning comes with a price: it can be very difficult to get the right settings for your specific scene setup.
Luckily, the developers have taken some steps to make it easier. V-Ray now includes a Quick Settings window that allows you select the type of scene you want to render, and V-Ray will automatically configure the best settings for that scene type. My tests showed that using these presets yielded very good results, so it’s very good to see you can start producing good looking renders without having to learn how all the different settings work.
Chaos Group developers have also tweaked the render settings windows to make them easier to manage. You can toggle between three different views, from Basic to Expert, giving you access to a different number of settings. Since I am not familiar enough with V-Ray, I use the Quick Settings window to set up my scene and then use the render settings to tweak what I need.
V-Ray features different global illumination methods (engines) that you can use to render your scenes, and you can also use two different engines at the same time, one as primary engine and another one as secondary engine. The two engines are used to create different GI solutions that are combined in the final image. The primary engine takes care of objects that are visible to the camera (or reflections) and the second one is used for other points used for global illumination calculations, but are located off-camera. At first, I wondered why the software had so many rendering engines, but then they explained that some are better suited for certain types of scenes. For example, if you were to take render engine A to render an interior, it would work but render engine B would render the image much faster.
Rendering global illumination can take a long time, even on fast renderers. This can be your worst enemy, because it means you’d need to render a test image, then tweak your setup, render again, tweak again… This would also mean you’d usually render at small resolutions because you don’t want to spend too much time rendering test images at higher resolutions. Luckily, V-Ray includes a module called V-Ray RT. V-Ray RT is a (near) real-time rendering engine that uses your GPU to render a preview of your render as you work on your scene. Because it is real-time, you can make any changes to your scene and see them immediately in the render view. This includes modifying materials, lights, moving the camera, or anything you can think of.
V-Ray RT does not support all the different things the software render does, due to videocard limitations. Basically, even if videocards have a lot more cores than CPUs, they perform calculations differently. However, depending on your scene, you can even opt to use V-Ray RT in favor of the software renderer. Chaos Group showed me this animation (http://vimeo.com/90509568), which was completely rendered using V-Ray RT. You can also see a behind-the-scenes video, where you see V-Ray RT rendering the animation in real-time, during motion capture (for this setup they were only using one videocard, so to get the real-time playback they had to limit the image quality and that’s why it displays grainy) https://www.youtube.com/watch?v=nnaz8q6FLCk
You have three different calculation methods for V-Ray RT: CPU, CUDA and OpenCL. I’m inclined to think everyone using Max will use either an Nvidia or AMD card (I will leave the CPU method out for the moment), so you select CUDA or OpenCL, depending on the brand of the videocard you’re using. Ironically, even if OpenCL is supposed to be “open,” you can’t use OpenCL on Nvidia cards (it will cause V-Ray, and Max, to crash). I think this is worth mentioning, because, if you’re as curious as me, you’d at least try to see which one is faster.
On the software side, V-Ray also lets you use progressive render, so you can get a fairly good idea of what your scene looks like very quickly. Basically, the progressive render renders the whole image in a very crude/grainy way, and then starts iterating on it to refine the image, add more details, until it reaches the final image quality. What you see is very similar to V-Ray RT, because the image gradually takes shape (just like when you use V-Ray RT) and you can start getting a good idea of what the final lighting will look like fairly quickly. If you are not pleased with the look, you can just cancel the render, tweak, and render again. The result here is similar to using V-Ray RT in CPU mode.
V-Ray can use a few different file caches to store data and reuse it on demand. This can save time when rendering animations, but also serves as a way to reduce flickering in your animations (if you’ve followed my articles, then you know I actually reviewed a plugin that removes flickering from video).
Expert V-Ray users mention V-Ray 3.0 is around 3 times faster than the previous iteration, which is a huge improvement over previous versions. Depending on your scene, this may mean a few seconds or minutes, but when you start considering animations, saving a few minutes per frame quickly translates to saving hours in full scenes.
I should mention www.v-ray.com, a website Chaos Group is using to announce new software versions, and provide an overview of new features. The long term plan is to use this site as a community-centric website for all things V-Ray, and provide more resources as well. If there are things you want to know about V-Ray that were not covered in this article, you should definitely check this website out.
If you are a V-Ray user there’s a big chance you have already upgraded to version 3.0, but if you haven’t, the increased render speed should be reason enough to try the new version out in case you’re considering upgrading. As I said before, this is my first time using V-Ray, but I have used other renderers in the past, including MentalRay, and when it comes to photorealistic rendering, V-Ray wins. The fact that you have access to many different settings means you can fine-tune your renders the way you want, but they make the renderer harder to learn. On the other hand, it’s good to see Chaos Group take the right steps to make V-Ray more accessible to new and inexperienced users.
As you may know, I work in game development, but if I ever needed to render photorealistic scenes for a project (or a game cinematic), V-Ray would be one of my first options (if not my first). For this article, I used the 3ds Max version, but V-Ray is also available for Maya, so I wouldn’t even need to switch 3D applications. Definitely give V-Ray a try!
Sergio Aris Rosa [nemirc], is Sr. Staff Writer for the Renderosity Front Page News. Sergio discusses on computer graphics software, animation techniques, and technology. He also hosts interviews with professionals in the animation and cinematography fields. You can follow him on Twitter, and if you want to see what he's up to you can visit his blog.
June 30, 2014
Please note: If you find the color of the text hard to read, please click on "Printer-friendly" and black text will appear on a white background.
Please take a moment to join Renderosity's Newsletter List to receive news and information from Renderosity!