|
|
VRay 1.0 bug or not?
About Truespace Archives
These pages are a copy of the official truespace forums prior to their removal somewhere around 2011.
They are retained here for archive purposes only.
VRay 1.0 bug or not? // Archive: Tech Forum
Post by Burisman // Apr 6, 2007, 3:47am
|
Burisman
Total Posts: 128
|
Hi guys, I noticed a strange phenomenon which I can't explain, but since I am not such a well-versed, experienced TS user as the most of you, I cannot work out whether to call it a bug, or just something I do wrong.
I am creating this scene, which is still in a very rudimentary stage (yep, I started with the easy bits), in which I want to place some light boxes on a table. These light boxes consist of an empty cube frame inset with 5 planes that contain 5 images. I used sub surface scattering to give these planes a 'parchment'-like look. The right-most box contains a local light. I copied it and placed it to the left, without the light; to remove the planes and replace them later with some other engravings. (All in all this scene will represent the study of a Faust-like person ... I am now frantically trying to model a human :o , and following the tutorials).
But here the strange thing happens. I noticed that the picturesof the left cube still shine very bright, notwithstanding the fact that there is no light source in the box. (You can verify this by looking in the mirror underneath the boxes, you sea that they are empty and that the right one has a light within, you also see the reflection of the inside of the left box, again very light).
This shouldn't happen, I think, because, as I understand is, SSS is using an external light, and not some light emanating from the material (the plane) itself.
You can see in the image that there are only 2 significant lights in the scene: 1. the bulb in the box on the right; 2. a small spotlight to illuminate the framework of the model house (but that doesn't shine in the left-most box.)
Anyone any ideas?
(In case it might interest you, the black and white images are details from 19th century engravings, made after paintings in the British Royal Collection, this collection of engravings was given to me once by a mad old aunt who crossed the world on freighters in the fifties, she died years ago, or she evaporated together with the colossal amount of cigarettes she smoked; the colour image is from a acrylic painting on wood panel I made some years ago, depicting an African mosque as seen by the French adventurer, writer and weapons smuggler Henry de Monfreid). |
Post by TomG // Apr 6, 2007, 4:17am
|
TomG
Total Posts: 3397
|
Hi,
hard to say. Try painting the box with a non-SSS material.
Also note that SSS is not about light passing through the object - instead light enters the surface, then bounces around and re-emerges. This means that subsurface scattering also has a more illuminated look when light hits it from in front too - light will hit your left cube from the outside, bounce inside the surface and come back out toward the camera, giving it more of a glow and more of a brighter look than a shader that is not SSS.
So it could be that this is correct for your scene :) It would be expected behavior for SSS materials to do that - not only will light pass through your skin from the back (as seen when you shine a flashlight through your hand), but it will also be captured just inside the surface, bounced around, then re-emitted from a front facing light too, or of course light from any direction that hits the surface - and indeed it will not re-emerge at the same point, so light hitting at (10,10) on the surface could cause the point at (11,11) to gain illumination even if (11,11) is not hit directly by any light.
One test would be to make your right box have a spotlight in it that is casting light on just a small portion of one panel from the back. You can then tweak the SSS and how it is responding to that light, as you can compare the areas where the inside spotlight is shining through and the areas that have no light from the inside shining through - you can adjust the material to maximise that difference, then return to an omni light on the inside.
EDIT - PS I can definitely see the difference in the top and left panels btw, they are brighter in the right cube. The front image of the buildings is harder to tell anyway, since the image has illumination data in it and the eye can't really pick apart the illumination already in the image of the buildings, and the illumination being generated in this scene from the SSS effect.
HTH!
Tom |
Post by Burisman // Apr 6, 2007, 4:22am
|
Burisman
Total Posts: 128
|
Thanks Tom,
That really sounds very reasonable. I assumed indeed that the light had to shine on the material, was diffused a bit, and then came out. But it's more subtle the way you explain it, in fact the object can catch light that is passing through the scene and redirect it towards the camera, kind of light-catcher. :) Well I will set up the experiments to explore this further. Very interesting by the way.
Thanks for the trouble of answering Tom! |
Post by Burisman // Apr 6, 2007, 4:27am
|
Burisman
Total Posts: 128
|
EDIT - PS I can definitely see the difference in the top and left panels btw, they are brighter in the right cube. The front image of the buildings is harder to tell anyway, since the image has illumination data in it and the eye can't really pick apart the illumination already in the image of the buildings, and the illumination being generated in this scene from the SSS effect.
HTH!
Tom
You are right there. I see it too. I just didn't expect the light to be so strong ;) I'm happy it's not a bug but just me :p
Thanks again! |
Post by Jack Edwards // Apr 6, 2007, 1:36pm
|
Jack Edwards
Total Posts: 4062
|
You also might want to check your luminance and diffuse settings. Luminance should be 0 except for light casting surfaces. Diffuse should be low for reflective or trasmissive surfaces like glass, but up to as much as 80% for matt surfaces.
I forget what I usually use with SSS, but I've still haven't gotten that shader to work exactly the way I've wanted it to yet.... so if you get some good results, you should post your settings so the rest of the community can benefit. :D
-Jack. |
Post by Burisman // Apr 7, 2007, 5:51am
|
Burisman
Total Posts: 128
|
Many thanks Jack. In fact upon checking the SSS settings I discovered that I had left some luminance, and even specularity and refraction :o . So turning everything off which pertains to reflective surfaces, and giving diffusion full throttle, that made a difference.
I also discovered some local lights, I had forgotten about :o
But still the surface seems so capture light that is randomly passing through the scene. It is working as a light amplifyer. Well, of course, it is supposed to, but I can't predict the effect very well.
To follow up your suggestion I am now systematically working through a large number of possible settings, starting with 1 infinite light, perpendicular to the grid, then incrementing the angle. I now also notice the interesting effect of the thickness of the material. So there are many possible permutations, given the amount of settings for SSS :rolleyes:
Still the SSS shader produces nasty grainy artefacts on the otherwise very smooth images. But they seem to diffuse now.
I will post the results of this experiment soon.
Thanks again:) |
|
