koobas.hobune.stream
How to create 3d texture effect
Specific Texture question here:
I saw this really cool effect in a video tutorial: https://youtu.be/lmLG4nC9tm0?t=723
I looked up how to make normalized maps to give my textures depth but its not nearly the same. How did this object in the video get away with creating depth with nothing but the texture?
Are height Maps and normal maps different things? Does it really make your game perform better because theres less geometry to render?
Thanks!
Take a look at this:
https://www.pluralsight.com/blog/film-games/bump-normal-and-displacement-maps
A height map and a bump map are arguably the same thing in which they are both greyscale images that tell unity depth information. A normal map is an RGB texture which gives unity depth information. They both attempt to send depth information in an attempt to fake lighting/shadows.
Do not be mistaken, they carry an extra workload as opposed to just having a regular texture on the model face in question, but they are less expensive than creating a super intricate mesh. Having less geometry is good for performance, but I think it’s a balancing act between poly count and nice looking textures, but there are a lot of things to consider, notably what are your target platforms. If you are targeting smartphones, there are a lot more limitations than that of desktops.
I personally like faking as much as possible in the texture itself (no bump or normal maps, just detailed textures), but normal maps can give a nice effect if you want that extra depth.
Hopefully that helps,
Cheers
While you are right about normal maps and bump maps, the video looks like an actual displacement map. Though displacement maps are quite expensive. I guess it's a displacement map based on the preview of the image site where the texture came from which was used for the crate. I guess it was this one. If you switch the preview mode (at the bottom left corner) to "sphere" (3rd option) you'll see that the geometry is actually displaced and not just some normal map lighting.
As I understand it, a displacement map will actually create more geometry, correct? So the benefit of using them is that when a face of the model is not being rendered, the extra geometry will not be calculated? Because it would seem to me that having predeter$$anonymous$$ed vertice positions on a mesh would be much more optimized than calculating the displacement every frame. Or am I missing something?
Yes, in most cases actual displacement mapping doesn't make much sense. Ins$$anonymous$$d of displacement mapping one could also use parallax mapping (here's a more detailed article). Though parallax mapping, like normal mapping, doesn't change the actual geometry and therefore the shape / outline wouldn't change. So it might be parallax mapping in the video. Unity has some parallax shaders. This might be also interesting
Cool! Thanks for the insightful answer, I was initially miffed about not being able to make boxes as cool as the guys video.
But if its in fact more work to do displacement mapping then I'm content with sticking to sculpted boxes :)
