Maxwell Render | Official public forum

Hey all -

I'm having some bump map weirdness and I can't figure out what is going on. Here's the result, and a lo-res of the bump map:



The left side object is a bevelled plane done with geometry, the right/back one is a flat plane with bump mapping applied. For some reason the bump map just won't work properly. It looks inverted no matter what I do, but it's actually not inverted, it just seems to react like crazy to the reflection of the floor.

The material is the same on both models, just the one on the right has the bump, the modelled bevel piece has no bump.

I've tried with and without strong reflections and it doesn't matter. Something weird going on, but is it me or 1.5 or what? Any thoughts?

I also posted the scene file here if anyone wants to have a look (it's exported to Studio from Max)

http://www.simmsimaging.com/CG/marble_wall_scene.zip

thanks in advance for any help!

b

I am having similar problems. And another thing that works in a strange way is that bump amount seems dependent to the size of the map. With big maps such the ones i use for tiled surfaces, i get almost no bump at all.

is your map RGB or greyscale?? this used to matter, though I'm not sure it still does

Thanks Lebbeus - it's RGB.

I've been continuing to test around this while waiting for some replies and I'm finding some further oddness. In the render I posted the marble bitmap was used twice (two BSDF) but it was a different copy of the same image used on each layer. The file was identical, but two locations. By switching it to one file it did seem to fix the reverse look, but it's still all over the place for weird reflectance/brightness.

Strange.

b

I, too, have came across strangeness in the M~R bumpmap area. (Please note that I love Maxwell and am not here to rag on it. I only wish to understand and/or help make it better through exposing strangenesses to Juan and team. ) Here's my silly test bump map (the original is a .png):



Here's my render of this map on a square plane (with a color map added) done fully in LightWave. LW renders it out as I would expect:



And here's my M~R of the same mapped object:



The scene has only one light as the shadow conveys. Consistent with simmsimaging's findings above, the M~R bump mapping is not acting as one would expect. All the beveled stuff looks fine, but most everything else...
Note that the two crisp rectangles at the top are displaying the illusion of being sunken in rather than raised, although the light source would need to be on the other side (right) for even that effect to be completely true.

I would almost conclude from the perfection of the bevels that M~R doesn't like values jumping too great a distance from one pixel to another except that the airbrush spots are also appearing strange in the M~R despite then having plenty of grayscale to avoid great jumping (at least that's my assumption).

So, am I (and Brett) merely doing something incorrect in setting up our maps, or is this a borderline bug? Thanks for listening!

what happens if you blur this bump-map a bit?

quick test:
bumpmap:


rendering:



more blur=more bump!

So you mean that for crisp bump edges like the ones of the white wall in alpha times we'll have to wait for displacement?
I am a bit disappointed because until then all my bumpmaps are almost unusable, expecially for huge real scales like the ones for tiled surfaces...

Mattia Sullini wrote:So you mean that for crisp bump edges like the ones of the white wall in alpha times we'll have to wait for displacement?

Yes, I think that crisp+bump is a difficult combination. But if you can accept a little bit "roundness" it's easier...

So in order to use bump maps in MW without weird artifacts we have to soften the maps more than we would in other renderers? I can live with a bit of roundness in most cases, but I"m curious whether that gets rid of the other odd qualities we have been finding.

Thanks
b

Brett, I think your case (1. post) has something to do with the bright, light-reflective floor. It would be very interesting to see the same testrendering with a dark floor!

Brett, I think your case (1. post) has something to do with the bright, light-reflective floor. It would be very interesting to see the same testrendering with a dark floor!

It is less obvious without anything to reflect there, but the problem isn't the floor. I did try it with several combinations of flooring because I originally thought that was the problem too. That's why the test I posted shows the embossing done with geometry in the scene as well. To me that clearly shows it's bump related.

I'm a little swamped at the moment but I'll try a test with a blurred bump map when I have a chance.

b

I can't go back and test it on 1.1, but isn't this something that has cropped up in 1.5? I have a bunch of pretty sharp wood grain bump maps that don't seem to do anything anymore.

simmsimaging wrote:. That's why the test I posted shows the embossing done with geometry in the scene as well. To me that clearly shows it's bump related.

Well, my theory might be wrong, but to me that shows that the bump is receiving light from beneath, while the geometry is not - which is natural, because it's creating "real" shadows.

Well, my theory might be wrong, but to me that shows that the bump is receiving light from beneath, while the geometry is not - which is natural, because it's creating "real" shadows.

But doesn't that suggest that the bump effect is capable of accounting for it's own shape for direct light, but not indirect light? In other words: it self-shadows and reflects properly for emitters, but not the reflected light off the floor, or the actual reflection of the floor?

That can't be right, can it? Am I missing something obvious? (wouldn't be the first time.... )

b