Page 1 of 2
bump map + normal_smoothing?
Posted: Thu Jun 05, 2008 11:02 pm
by suvakas
Hey,
I can't get the bump maping to work with normal smoothing.
Here are two images, 1st has a teapot with a bump map applied and 2nd without it.
Posted: Thu Jun 05, 2008 11:47 pm
by Zom-B
Seems to be related to
this
Sould the connected to Sun Lightsource...
simply subdivide your mesh until your RAM bursts and you'll get nice shadows
Posted: Thu Jun 05, 2008 11:50 pm
by OnoSendai
hmm, can you do a test with a very small meshlight in the same position as the sun, instead of the sun+sky?
Posted: Fri Jun 06, 2008 12:12 am
by supraserv
I had the same problem months ago when I was creating my Egyptian scene. I noticed the same in some Maxwell renders elsewhere. I made a thousand tests but they didn't help.
Posted: Fri Jun 06, 2008 12:53 am
by suvakas
OnoSendai wrote:hmm, can you do a test with a very small meshlight in the same position as the sun, instead of the sun+sky?
Still same...
Posted: Fri Jun 06, 2008 2:51 am
by cookieofdoom
I'm pretty sure that kettle would never work right anyway.
Posted: Fri Jun 13, 2008 9:44 am
by suvakas
Hey Ono,
What do you think is the cause of this issue?
The problem is still there in 1.1.5 too...
Posted: Wed Jul 30, 2008 5:08 pm
by v_mulligan
Hmm... I've seen something similar on Renderman-compliant REYES renderers. If you write a shader that calculates the bumped normal by perturbing true geometric normal rather than the smoothed normal, it looks a bit like that. I don't know why it only seems to be affecting direct light, though, and not reflected light. Of course, Indigo's a bit of a black box to me. Only OnoSendai can really say what's going on here.
Posted: Wed Jul 30, 2008 5:13 pm
by v_mulligan
Hmm... Looking at it again, it looks like each of those polygons is either fully illuminated or fully in shadow, with regards to the emitter, but has proper intermediate levels of shading with regards to reflected light. Is there any chance that something that ought to be a float or a double is coming out as an int or a bool? Total stab in the dark, here.
Posted: Wed Jul 30, 2008 9:53 pm
by Deus
You are right mulligan it looks like the albedo direct lighting is completely broken
Ono maybe you should try general diffuse
color = cos( normal dot lightdir )
Posted: Thu Jul 31, 2008 6:49 am
by v_mulligan
Don't you just mean normal.lightdir? The dot product of two vectors is equal to the product of their lengths and the cosine of the angle between them.
Posted: Thu Jul 31, 2008 8:10 am
by BbB
That always happens with sun light and bump maps on low-geometry meshes. No way around it but to increase the polys. Pretty much all renderers are affected by the way. Not an Indigo singularity.
Posted: Thu Jul 31, 2008 8:19 am
by v_mulligan
BbB wrote:That always happens with sun light and bump maps on low-geometry meshes. No way around it but to increase the polys. Pretty much all renderers are affected by the way. Not an Indigo singularity.
Um... no. REYES renderers certainly don't give that sort of image artefact unless you've done something wrong.
Raytracers might with very bright light since some edge polygons will be shadowed by other edge polygons. That still doesn't explain why the direct illumination seems to be uniform across the lit half of the surface, though.
Posted: Thu Jul 31, 2008 8:20 am
by suvakas
@BbB
Do you mean other unbiased renderers? Cause I haven't seen this kind of artifact with any other 3ds Max renderers i use.
Posted: Thu Jul 31, 2008 9:04 am
by zsouthboy
Yeah not only that, I don't think this used to happen with Indigo, as well, but we never really had closure on it..