Vray GI and Color transmittance tests [maxwell] involved

[Da_elf]

i had done this render for the fun of it (sorry for the nudity) the only lighting comes from the direct light and the volumetric light (no color mapping used) sorry aobut the image size forgot it was on HD size.
http://www.elfpro3d.com/vray/volume.jpg

[RErender]

Rerender yes I know what you mean, but, we'r doing testing here nothing more.

Yep I fully understand that at first I wasn't sure of the point, but it is clear to me now. I too am very interested in ascertaining the way to get predictable and correct lighting results, I would like to be able to create a render as a light study and be confident that it is correct within reason.

[Morbid Angel]

rerender - indeed. Imo a lot of the render especially of the interriors out there are probably done with wrong scale, thus producing the unwanted burn from vray lights and incorrect darkening and etc. And by incorrect scale I mean, not necesserely wrong all around, but wrong to vray light mesurments.

Vlado, does gi also depend on the scene scale? If you think in real world, it would take more light to illuminate a larger area then a smaller one (in interrior space)....or is it based on a different formula?

[Morbid Angel]

Vlado, I have replicated your setup in example 1. While it gave me an identical result to your first test, where the point right below the light is 128,128,128. However moving this light from 50 units to 10 units close to a plane has produced an extreme burn as in my previous cases. This burn effect is present in many vray renderings todate. Im wondering if there is something to be done to change this. One might say that the lights intensity is too strong for being 10 meters up above the ground...then the same is applied to the light being 50 meters up and being too weak.





also notice the change in lighting on the side walls. On img 2 the walls are brighter, even though the light stays exactly the same distance from the side walls as in image 1. For some reason, which I cant explain the rgb value on the sides went from 30 30 30 to 34 34 34.

[dlparisi]

also notice the change in lighting on the side walls. On img 2 the walls are brighter, even though the light stays exactly the same distance from the side walls as in image 1. For some reason, which I cant explain the rgb value on the sides went from 30 30 30 to 34 34 34.

This makes sense to me. Since it would seem that more light is being reflected off of the floor now than in img 1. My guess is that the increase on the walls would add up to the decrease on the ceiling and other areas.

[Morbid Angel]

dlparisi - physicly correct speaking you are correct. But in this condition, the surfaces are non refletive. There is no gi bounce so technicly there shouldnt be a change.

[vance3d]

i didn't think there was any gi involved in this test so there would not be any reflected light.

I don't know why you would get a change in the sampled value unless the area you sampled on one of the images was not the brightest area but an area right next to it.

I don't know how much you would perceive a change in value with the range you described in your test. visually yes the walls look brighter on image two but thats because the floor is much darker since less of it is in range of the light.

V Miller

[dlparisi]

dlparisi - physicly correct speaking you are correct. But in this condition, the surfaces are non refletive. There is no gi bounce so technicly there shouldnt be a change.

Oops, missed that point (i.e no GI). It's hard to tell in your images but were the walls always this intensity and it's just more apparent now because it's right up against the black floor?

David

[Morbid Angel]

yes, the walls are identical distance and color of the floor plane to vray light.

[vlado]

Dmitry, in both your images, the brightest part of the walls are nearly the same - RGB 33, 33, 33 in the first image, and RGB 32, 32, 32 in the second, which is close enough. However, it seems that you used a sharpening AA filter, which enhances the edge between the walls, which is why it jumps to 34, 34, 34 just near the edge.

As for the burn-out, if you calculate the intensity for the part below the light, you will see that it is way above 1.0, which is quite correct.

For a perfect point light, the closer you get to the light, the brighter surface you will have, without any upper bound - e.g. the surface becomes infinitely bright as you get closer to the light.

This is not the case with actual area lights, which do have an upper limit on how bright the surface gets, but usually this limit is also very high.

In general, burn-outs, however undesirable, are physically correct and can be observed in actual photographs. The only way to deal with them is to use suitable color (tone) mapping.

Best regards,
Vlado

[Morbid Angel]

I see, Vlado, I used quadratic filter on my image. But never the less, my question to you would be if vray light is made to work in unit scale of 1 u = 1 m, what is the correct settings if my scenes are in 1u = 1 cm? would it be from 1.0 to 0.01?

[RErender]

It would be a factor of 100^2 or 1.0 to .0001