I wanted to know if there is a solution to build High-End Car Paint. I heard something about a very expansive solution for Mental Ray and am wondering if VRay has something similar.
We are possibly going to make a job for stills and are getting car paint examples. What would be the best workflow to get these examples transferred to a correct shader?
Does anybody have experience with that? I would like to know how you are going to build high-end car images…regarding the paint…
Thanks in advance and sorry for the generally question.
Yes, it’s entirely possible. You don’t really need an “expansive solution” to do good car paint. You simply need a base diffuse layer, a reflective “metallic/pearl” layer with a lower gloss value, and then a clear layer with crisp reflection.
That´s clear. But what I´m talking about is, that if you get a car paint example from a car manufacturer, is there a solution to rebuild it in a “physically” correct way? That it reflects in this angle like this and in a different angle like that…
You would then use a VRay-Blend Material, with three layers, I guess. Diffuse, Flakes, Gloss…am I right?
If they gave you information about a finish, it would probably be full of terms like reflectivity, energy, IOR, etc. You should be able to replicate most of it with the VRay shader.
I’d have to check with Vlado, but i’m pretty sure Energy is a normalized value of the material. So an energy value of 1.0 would mean diffuse 128,128,128 + reflection 128,128,128 = 1.0. Or something. I’d have to see what the technical documents of the material are, rather than just assume what’s on them.
Sounds reasonable to me. We are not so much into it, that we would have known that energy is put together bei diffuse and reflection.
Right now, we didn´t receive any material example yet. Just generally speaking. But if we are getting one, we are going to supply the measures written down on it. Maybe we can figure something out.
We heard something about a method where you can scan a material example, and let build a Maya-mentalRay-Shader out of it. With the HDRI it would be the perfect reproduction of the real world…I guess there is no solution for VRay right now to do something similar to this?!?
Greetings from Germany and thanks for your answer!!!
I work on car commercials pretty much every day at work, and we get paint samples all the time. It’s usually pretty easy to get 90%+ “accurate” looking shader just doing it by eye. The rest usually kind of depends on your lighting.
I forget who it was I saw last year at Siggraph, but they said they had a BRDF scanner up and working for car paint, and the results they got for paint shaders were very very similar to the results they got just eyeballing it, nearly indistinguishable if I remember correctly.
Yes, a BRDF scanner is what I meant, but didn´t know the word. But it´s too expansive and it´s more built for Maya with mentalray pipelines. Though we are working with max and VRay. Wanted to know if there are any inventions going on to implement such a pipeline into Max and VRay.
So you get 90% out of a metarial example by eye. Sounds good to me! How long are you tweaking on such a car paint regularly?
I’d say half of it is getting a paint sample in an environment you can recreate - shiny objects are going to vary massively depending on their environment so if you want to recreate something you’ve got to consider that half of the look is being produced by what’s around the material and the other half by the material settings.
I’ve done alot of car paints, so it doesn’t take me hardly any time at all to get the shader pretty close. I imagine once you have done a few of them, you could bang one out pretty close in 5 minutes? Assuming you already have a test lighting environment setup.
It all depends on the environment, that´s for sure. If it´s a studio setup and you have to see all the lines of the cars, it´s pretty hard to make a good car paint. respectively it´s more the light setup than the paint…it´s all about reflection!
Vray is capable of a lot. I’ve even had pretty good results using the Vrayblend material. If you do a search on this forum, there’s a lot of talk about carpaint methods.
One of the artists here (Suurland) has had very good car work using Vray:
Indeed so. Particularly in a studio setup, the lighting can makes the whole car look bad, if it’s not well lit. Even shooting a real car a car in a real studio. Even if the shader on the car is perfect, “bad” lighting (in any software) will will ruin the look of the car. However, good lighting with “bad” shaders will look equally bad. The model has to be good, the shaders have to be good, and the lighting has to be good And on top of that, well done “post” work too. Even real cars shot in a studio go through post work to make them look better. Not just painting out things, but enhancing the contrast, making the wheels and headlights “pop” more, and so on.
The mark 1 eyeball as fitted in human beings is all the scanning you need. combine this with a detailed technical understanding of vray, and in particular the vrayblend material. the most important thing is tweaking the material to your scene’s lighting, something you would spend just as long on even if you started with a BRDF scanner. Im guessing you could say the same about any other software, maya/mental ray included
Exactly - if you had a photo of a car, a brdf scan of it’s material and put that into a 3d scene, the cg car isn’t going to look the exact same as the car in the photos since the lighting and environment is probably different - the major thing is taking something that’s accurate and then changing it to what you want / adjusting to taste. If you’re working with a client and say something is technically accurate that could be quite different from what they actually want or expected. They could want the look of the material of the car in the photo but in your new lighting / environment setup so it’s a case of making adjustments until you get in the area of what you want.
There has been done some scanned BRDF stuff for VRay. Mind you that this wont necessarily give the desired look. For one it is quite a hassle to get the scanning
done and then it is hardly tweakable at all. So if the client wants a different result
that can get very difficult. In addition you’d need to scan all paints you wanna do. As of
course every paint is pretty different.
Having a decent Base Setup and tweaking it as needed for a specific result / environment currently seems the most feasible way to me.
And on top of that, well done “post” work too. Even real cars shot in a studio go through post work to make them look better. Not just painting out things, but enhancing the contrast, making the wheels and headlights “pop” more, and so on.
)