Weird behaviour of bump

Use this forum for general user support and related questions.
Forum rules
Please upload a testscene that allows developers to reproduce the problem, and attach some images.
User avatar
FarbigeWelt
Donor
Donor
Posts: 1046
Joined: Sun Jul 01, 2018 12:07 pm
Location: Switzerland
Contact:

Re: Weird behaviour of bump

Post by FarbigeWelt »

Edit: See next post
Last edited by FarbigeWelt on Fri Aug 17, 2018 6:13 am, edited 1 time in total.
Light and Word designing Creator - www.farbigewelt.ch - aka quantenkristall || #luxcorerender
MacBook Air with M1
User avatar
FarbigeWelt
Donor
Donor
Posts: 1046
Joined: Sun Jul 01, 2018 12:07 pm
Location: Switzerland
Contact:

Re: Weird behaviour of bump

Post by FarbigeWelt »

FarbigeWelt wrote: Fri Aug 17, 2018 6:06 am Well, I struggeld too with the use of bump maps. I baked different normal and different displacement maps with blender. And tested them with luxcore. Depending on what you bake the result gets at least funny. Rational was a torus baked on a plane. The bump throws shadows depending on bump height but image always looked flat if viewed from an angle. If added checkered texture what in one case lead to hovering looking, alternating plates if applied to a flat ‚cube‘. I‘ve tried texture wrinkled with different bump heights on a sphere what let to impressive valleys. And if I applied just some other texture to the sphere I got pressed looking spheres like some heavy pressure reshaped a football. I wonder if 3D procedural textures forms an object whereas images or normal maps keep object flat and renders in best case a an expected shadow from a certain angle of a light source. If 3d textures lead to another result one should be close to add 3d objects instead of images to the bump input with pointer.
Bump seems everything else than an easy, intuitive tool.
Light and Word designing Creator - www.farbigewelt.ch - aka quantenkristall || #luxcorerender
MacBook Air with M1
User avatar
FarbigeWelt
Donor
Donor
Posts: 1046
Joined: Sun Jul 01, 2018 12:07 pm
Location: Switzerland
Contact:

Bumped and a freaking Sphere

Post by FarbigeWelt »

Maybe I''ve set some bad parameters somewhere..
But I get very strange images. There is also some weired memory effect if you use bump maps. If I change the picture or I remove the node input the first applied bump map keeps applied. I have to add a new object and chose the changed material then to ged rid of the first picture. :?:
Did anybody experience similar effects with bump map?
What is wrong here
What is wrong here
Displacement.zip
Displacement
(992.51 KiB) Downloaded 162 times
Does anybody have an idea what happens to the smoothed spheres? :?:
:arrow: I guess it is somehow related to Wasd's observations.
Light and Word designing Creator - www.farbigewelt.ch - aka quantenkristall || #luxcorerender
MacBook Air with M1
User avatar
Egert_Kanep
Posts: 237
Joined: Tue Mar 13, 2018 10:34 am

Re: Weird behaviour of bump

Post by Egert_Kanep »

Check over your values, I opened your file and it is pretty extreme what is going on there. Check your scene scale, objects scale and light intensity also. I also changed camera tonemapping to get better control over scene exposure. And do not confuse bump with displacement, as far as I can tell they are very different. Bump is for adding micro detail, not changing geometry shape. I also encourage you to join discord server, where you can get live support. link - https://discord.gg/chPGsKV
Attachments
Capture.PNG
Displacement_changed.blend
(3.5 MiB) Downloaded 138 times
Last edited by Egert_Kanep on Fri Aug 17, 2018 5:39 pm, edited 1 time in total.
User avatar
B.Y.O.B.
Developer
Developer
Posts: 4146
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: Weird behaviour of bump

Post by B.Y.O.B. »

FarbigeWelt wrote: Fri Aug 17, 2018 4:20 pm Maybe I''ve set some bad parameters somewhere..
Your spheres have a diameter of 2.
You have a wrinkled texture, which outputs in the 0..1 range, then you multiply it with 0.5, and then you use it as bump texture.
So you expect a bump effect of 0.5. Try to think of the values in meters.

This is more than extreme.
Bump/normalmapping is just a fake effect, it can never look convincing with such extreme values.
You can only try to fake a bump effect that would not create much visible surface deformation anyway.
So on a 2 meter sphere viewed from this distance, maybe 10-20 cm is the maximum I would use.

However, your scene is also hard to judge because of other reasons:
- The point light together with very lowpoly geometry creates massive terminator artifacts
- the right sphere is overexposed, the automatic tonemapper is annoying

Here is your scene with these issues fixed, using the same insane bump values:
Attachments
Displacement2.blend
(3.49 MiB) Downloaded 155 times
untitled.jpg
User avatar
FarbigeWelt
Donor
Donor
Posts: 1046
Joined: Sun Jul 01, 2018 12:07 pm
Location: Switzerland
Contact:

Re: Weird behaviour of bump

Post by FarbigeWelt »

Thank you, both of you.
The point light together with very lowpoly geometry creates massive terminator artifacts
Great, now I understand the term terminator artifact. Actually same effect can be expected from high polies in this case but less visible because smaller. First time I've noticed these bricks/stairs on a smoothed object. :roll:
the automatic tonemapper is annoying
But seems to work in less extreme situations, doesn't it.
And do not confuse bump with displacement, as far as I can tell they are very different
Oh, displacement would be a great feature. I understood bump as kind of flat shadow thrower. Bump maps work in some cases very well to add some structure to an incredible boring flat area.

Mayebe I should use the bump map just to simulate i.e. dust and surface scratches.
Light and Word designing Creator - www.farbigewelt.ch - aka quantenkristall || #luxcorerender
MacBook Air with M1
User avatar
B.Y.O.B.
Developer
Developer
Posts: 4146
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: Weird behaviour of bump

Post by B.Y.O.B. »

wasd wrote: Thu Aug 16, 2018 11:21 pm
B.Y.O.B. wrote: Thu Aug 16, 2018 9:16 pm If you plug any texture apart from the normal map into the bump slot, the grayscale values are interpreted as elevation, where 1.0 (white) is 1 meter and 0.0 (black) is 0 meter (or something like that, maybe the 0 m is gray, not sure about that).
If you're not sure, how should I be?:)
B.Y.O.B. wrote: Thu Aug 16, 2018 9:16 pm Now, normalmaps are something different, they directly contain the shading normal.
I've already get the idea, that normal map is something usable.

This is cycles image:
bump-test-e-cycles.png

I've just plugged this map:
http://www.bricksntiles.com/textures/ma ... mp-map.jpg

Everything I've got with luxcore bump is either ugly or very ugly, I won't attach the images.
But you may have the scene. HDRI is the same.
bump-test-e.zip

P.S. About white being 1 meter and grey being zero or something – these are all words, it (maybe) supposed to be like this, but what I see is something else. It'd be great if there was some convincing images. Made to scale.
I didn't write the bump code, I don't know how it's supposed to work, I'm just trying to help (and I am almost as lost as you are on this matter). ;)
I took your file and tested a few values. If I use 0.8 mm as bump value, it somewhat resembles the Cycles result, however the shadows are brighter.
Maybe LuxCore simulates light bouncing from the bumped surfaces onto other bumped surface? ("self-lighting", if you will, analogous to "self-shadowing"?)
No idea.

By the way the Cycles results seems to lack global illumination? (the light bouncing from the chimney)
Attachments
bump_0.8mm.jpg
bump-test-e.blend
(800.04 KiB) Downloaded 146 times
User avatar
Egert_Kanep
Posts: 237
Joined: Tue Mar 13, 2018 10:34 am

Re: Weird behaviour of bump

Post by Egert_Kanep »

Also I think bump textures should be linear, meaning 1 for gamma, seems more logical that values are used in linear fashion. It is not a rule, but i think bump height should be dependent of object scale, let's say a 20th of object's scale. In case of this chimney, I would make it 50cm x 50cm and bump depth maximum 2.5cm. Although probably cement would be 1 cm inwards, but my point is that if you stick to this ''rule'' you won't get weird results.

So don't go using 10cm bump on a tiny napkin and you can live happily ever after!
User avatar
B.Y.O.B.
Developer
Developer
Posts: 4146
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: Weird behaviour of bump

Post by B.Y.O.B. »

FarbigeWelt wrote: Fri Aug 17, 2018 6:23 pm Mayebe I should use the bump map just to simulate i.e. dust and surface scratches.
It depends on the size of your scene.
If you want to render the earth in space, you can use bump to simulate whole mountain ranges.
But if you zoom in further, bump won't be enough at some point, you will need true displacement.
User avatar
FarbigeWelt
Donor
Donor
Posts: 1046
Joined: Sun Jul 01, 2018 12:07 pm
Location: Switzerland
Contact:

Re: Weird behaviour of bump

Post by FarbigeWelt »

B.Y.O.B. wrote: Fri Aug 17, 2018 6:51 pm
FarbigeWelt wrote: Fri Aug 17, 2018 6:23 pm Mayebe I should use the bump map just to simulate i.e. dust and surface scratches.
It depends on the size of your scene.
If you want to render the earth in space, you can use bump to simulate whole mountain ranges.
But if you zoom in further, bump won't be enough at some point, you will need true displacement.
:mrgreen:
World.Sea, Moon.Grey and Stars
World.Sea, Moon.Grey and Stars
Light and Word designing Creator - www.farbigewelt.ch - aka quantenkristall || #luxcorerender
MacBook Air with M1
Post Reply