Download seamless drip normal and heightvasilnatalie on DeviantArthttp://creativecommons.org/licenses/by/3.0/https://www.deviantart.com/vasilnatalie/art/Download-seamless-drip-normal-and-height-708926016vasilnatalie

Deviation Actions

vasilnatalie's avatar

Download seamless drip normal and height

By
Published:
9.7K Views

Description

www.mediafire.com/file/0ic70i0…

Same as with the chainmail.  4096^2 px textures.  CC0/public domain, no rules, use as you wish.  Seamless in both U and V (duh on the V, it's flat.)  If you want the mesh I made to make these pics, just ask.  Height map's normalized to the image, run it through levels or something if you want it to be in some particular scale.

This was much more difficult for me than the chainmail.  I'm sure there are some tricks to make it easier that I just haven't quite figured out yet.

My original goal was to use this to make a simple effect that does some UV distortion and uses normals and depths to imitate fluid flow.  But.

About 9/10ths of the way through this, I had a cigarette and started thinking about how I could do a drip physics simulation effect.  I know there are liquid sims that exist, but not like I was thinking about.  I think doing a drip liquid sim would be very doable and could look great (although it sure would be nice to be able to use vertex displacement instead of just doing it with normals and depth, but no displacement in MMD because shader model 3.0.)  It could collect in crannies and drip in the direction of gravity.  To do this, you would need a transformation matrix from UV space to world space-- a "tangent frame."  This is the exact same thing you need to do to make anisotropic highlights.  I had a hard time thinking about how I could even test my baby steps toward anisotropic highlights in the past, but I think I have some ideas now, and may revisit it some time.  Note that just because something has anisotropic highlights doesn't mean it looks good.  Pretty hair pictures don't just have anisotropy, they have a lot of vertices-- maybe 100k strands, each modelled as a curve rather than a set of polys-- and they have better tangent calculations than can be done in a 3.0 shader.

And I have a feeling this concludes my current spurt of modelling, and that I'm going back to some effects programming in the next few days.
Image size
2048x1024px 188.42 KB
Comments1
anonymous's avatar
Join the community to add your comment. Already a deviant? Log In
animaniac72's avatar
This looks pretty cool.  Thanks.