Best Vray Render Settings For 3ds Max

[Maitane Roderiques]

Increasing both the quality and size of the render will also increase the render time so make sure to pick settings that are balanced with how fast your machine CPU/GPU is and your level of patience.
Screen Shot 2021-02-01 at 7.04.37 AM17441558 209 KB

I have a scene where i have roughly 200 trees (3 different types) 100 bushes (1 type) all Archmodels from evermoton.org and the polycount is killing my render times. Its not the material as i am rendering the scene in "material override", and so i was wondering if there's a way to reduce this and speed up render times?

I don't know which method is the best, never had more than 10-15 trees in a scene but I would try the pro optimizer modifier. It will reduce the polycount but I think the trees will still look good from a distance.

If you're using 2.0 or later, set it to (0) to give V-Ray full access to as much memory as it needs. This will obviously take some testing+tweaking to ensure your computer won't crash-I've read some people set this number to somewhere around 75% of their system memory.

I have a scene with 103 trees, 63 bushes, 664 plants, and 500,000 grass patch proxies, and 3D Ivy, all rendering at once. it uses about 14GB of RAM when rendering, 1.8 GB used by Max when the scene is open, and renders on my Laptop.

Looking at your setting for Dynamic memory as they are the defaults, having 400MB limit would mean your machine would use up to that and then start paging the hard drive for the rest of the GBits it needed. Paging can be very slow.

Essentially, the more dynamic geometry you have, the higher the limit could be, up to a point which will differ for every scene. It's ovly relevant for dynamic geometry such as proxys, displacement etc. Shame on your 4GB limit though! But at least your were able to quadruple the RAM usage for that particular render. It opened your RAM up for usage and you saw the results so that is great!

First, we go through the default V-Ray render settings that are optimized for a large variety of scenes. With their help as a starting point, we learn how to make quick render previews that can be used for light adjustments. Then, we prepare the scene for a production render. At the end, we consider interior animation settings.

V-Ray 6 solves those issues with only two tweaks to the default settings: Light Cache Subdivs, and Retrace amount. While the suggested settings work for most cases (interior, exterior, environments, pack shots, simulations), we suggest you to test them across a set of representative frames before committing to the final sequence rendering.

Notice that depending on a number of factors (scene type, lighting, shaders, geometry), the same noise threshold may produce marginally different results on a scene-by-scene basis. A good tip is to always test before committing to final.

I am currently doing a render in Vray 3DS Max for one of our projects for the on-site sign. The architect here wants a high res. image because people will be able to go up to the image to have a look.

Basically this needs to be done asap, as in yesterday. I have everything set-up but i am struggling to render it at a high res. in a adequate time frame. Currently I am rendering it on my home PC and its indicating 35-40 hours (64bit, windows 7, 6 core, 8GB RAM)
Do you have advice on the rendering settings?

To actually compare render engines you have to have the same scene with the artistic approach from the same person with the same technical level in both engines.
from my experience using various render engines i believe that all of them have their pros and cons since you mentioned vray this render engines needs alot of tweaking to get photorealistic result compared to lets say corona render, some render engine gives you full control others help you to get the fast result of photorealism i could say achieving a stylized look in corona would be hard compared to achieving it in vray.
I believe the main difference in Vray vs Corona or Cycles vs Octane is being a biased render engine or unbiased render engine corona , octane are both unbiased render engines while cycles, vray they are both biased render engines you could read more about them in this article : -and-unbiased-rendering-algorithms-which-is-better/

From my perspective yes render engines like corona and octane are able to achieve photorealistic results easier than cycles because there is not alot of tweaking to do so they handle lighting and atmospheric effects (what i believe octane excel at) better than cycles immediately .

The high illuminated scene i think is not conclusive, it depends on materials reflectivity and or sun energy - it seems to not be the same. In the 2 render engines i would have expected some GI from the colour balls.

First of all, I have attached a screenshot of the material and exactly the walls. The same roughness, the same color, and I found for myself that this is the ratio of reflectivity will be the same. Like the metal - the roughness is the same. Again - different bump and works differently, not just its values, but how the renderer sees the map. In cycles two values to customize this, in octane one and the result is different.

The next thing is to note in general that cycles and octane are both unbiased. And they both produce phenomenal results.
The point is that in my opinion octane is the best renderer in the world in terms of quality, without nuance or compromise.

Different with the displacement. In octane it is the possibility to use it without subdiv but with uv and without procedural maps, in cycles on the other hand there is adaptive subdiv, due to which already real displacement works better in cycles.

No matter how you spin it, we have blenderartist where we see a lot of works, absolutely different level and orientation. And many of them are amateur, and even those that are professional are still from a specific niche, resulting in a different sample.

Further, most addons, objects and other things are made in the same way. Plants in a number of addons are configured through mix node for leaves, rather than through add, which distorts their display, compared to the same octane, and any other renderer, and in fact in a lot of works used these libraries of vegetation, rather than manually customized trees maxtree for example. In Geo-scatter is done through add!

Cause and effect are different things. You are talking about the highest level, about large projects, the realization of which falls on the shoulders of specialists/companies, who use the software that they use in their specialized work.

Moreover, you gave an example of UE, which is not even a path tracer. And even its path tracer is worse than cycles. And there is no point in talking about anything at all - path tracer of such a level is in any case superior to UE in its graphical capabilities.

So the conclusion is the opposite of yours - an artist, or a team, as in this case, who are paid for their work, who use the corresponding library. They had the guys from Gaea help them create a great environment.

In general, my message is that Photorealism and Physical Correctness are different things. You can achieve photorealism with gouache on paper. And you can lose it if the basics are broken in any complex software, so the artist is paramount.

With V-Ray Next for Maya, I can now utilize the viewport and GPU much more, as I can isolate specific areas of my scene with ease. As an instructor, when sharing V-Ray techniques with students or trainees, this offers an extreme advantage, allowing me to immediately isolate different pieces of geometry, lights and even specific nodes within a shading network just by clicking a button in the IPR.

I gradually raise my Min Shading Rate until the areas of secondary rays are clean, leaving the Image sampler Threshold with my single quality settings. A value of 0.1 will give me a noisy-yet-fast render and a value of .01 will be close to the final.

For scenes with depth of field and motion blur, I change to a heavier AA sampling setup. I usually start with my Max subdivs around 24, but if there is a soft focus or a lot of blur, it can go as high as 100. This also means I can ease off the Min Shading Rate and have the renderer focus more on the anti-aliasing.

In previous versions of V-Ray, I would need to hide lights or objects with display layers to isolate them but now I can do all of this by simply selecting them in the light lister. During look development, I can isolate certain parts of my shading network to help debug or see the result of my shading nodes.

V-Ray Next for Maya has progressed the evolution and simplification of my rendering workflow. I rely heavily on the Min Shading Rate and Max subdivs and stay out of a lot other render settings. I also keep local subdivs off almost all the time and can be confident that I will get a clean image with far less effort by dialing individual subdiv values.

In earlier versions of V-Ray, I used local subdivs and divided them, which made the workflow more intensive. You may accidentally leave a light or material setting too high and end up bogging down your whole scene. I often found myself experimenting with different adaptive amounts to try and get a faster render.

Stephen DeLalla currently works in the industry as a Lead Look Development Artist. As a former instructor at Gnomon School of Visual Effects, where he created the first V-Ray curriculum to modernize the lighting instruction at the school, Stephen is an expert in all things V-Ray. He is also the author of two instructional video series focusing on V-Ray for Maya, available through the Gnomon Workshop.

Not all Render Elements get picked up by Deadline. In my testscene, only the Sample Rate element gets picked up by the vrayscene parser. Z-depth and Multi Matte get ignored. I did not test any other elements.

Even when animation is disabled in Maya, the filenames have a framenumber at the end. This framenumber is missing in the Draft assembly config file so the assembly job fails. When animation is enabled, even with just one frame, Deadline picks it up correctly.

c80f0f1006