Planetside Software Forums

General => Image Sharing => Topic started by: WAS on October 19, 2020, 08:31:13 PM

Title: Mudcracks Shader
Post by: WAS on October 19, 2020, 08:31:13 PM
Makin' a mudcracks shader. I've made tons but never really treated the scalars with care. Also packing in a bunch of features.

MPD 1 AA 8 render with GISD sliders maxed only little over and hour so pretty optimized for voronoi stuff.
Title: Re: Mudcracks Shader
Post by: Dune on October 20, 2020, 02:40:59 AM
Looks very good.
Title: Re: Mudcracks Shader
Post by: Hannes on October 20, 2020, 04:28:17 AM
Holy cow, what a network!!!

The cracks look great!
Title: Re: Mudcracks Shader
Post by: RichTwo on October 20, 2020, 05:50:03 PM
Echo @ Hannes!  Nice result but worth the render time?  Depends on the desired result I suppose.
Title: Re: Mudcracks Shader
Post by: WAS on October 20, 2020, 06:59:53 PM
Quote from: RichTwo on October 20, 2020, 05:50:03 PMEcho @ Hannes!  Nice result but worth the render time?  Depends on the desired result I suppose.
Pretty fast render time considering settings and resolution on a Ryzen 2600. I can preview 2k at MPD 0.6 AA3 in less than 10 minutes. Clearly not fast for rendering animation but what is in TG when it comes to making anything realistic and than animating it. And I highly doubt in an actual scene you're rendering a full view of mudcracks for no reason which lessens it's appearance and load on render times.

Those same settings on my cliff setup you use would be far longer due to the chaotic nature and blown out scalars, which use same base noise, though this uses 5 function noises (3 or 5 depending on scale variation settings actually used).

Probably should note I also am rendering with 6 cores of 12, so I can run other software like Blender/Character Creator/Daz etc.
Title: Re: Mudcracks Shader
Post by: Kadri on October 21, 2020, 01:18:40 AM
Looks good.
Title: Re: Mudcracks Shader
Post by: WAS on October 21, 2020, 01:46:44 AM
Thanks. I want to warp the cracks by the mudcracks heightmap which makes them look so much more real, but man, just one warp and render time seems to double :( may still add it as an option.
Title: Re: Mudcracks Shader
Post by: Jo Kariboo on October 22, 2020, 09:42:55 AM
A great image of Mudcracks and the other two demonstrating  the shader and node connections is impressive!
Title: Re: Mudcracks Shader
Post by: N-drju on October 23, 2020, 05:45:02 AM
This looks fantastic. But there has to be a way to achieve the same with the tenth of that node network. It's insane... I wonder how long it took you to build that little monster. :(
Title: Re: Mudcracks Shader
Post by: WAS on October 23, 2020, 01:49:51 PM
Quote from: N-drju on October 23, 2020, 05:45:02 AMThis looks fantastic. But there has to be a way to achieve the same with the tenth of that node network. It's insane... I wonder how long it took you to build that little monster. :(
I mean you could slap stuff on a voronoi setup and hope for something similar. XD

This is all very deliberate use of scalars to masks from scalars to conform to the mudcracks shapes.

The basic concept is clamping voronoi diff untill you have a nice plateau. Then you use the original voronoi inverted and adjusted and masked by the faults to create lifts where mud dried too fast and pulled away from other cells when they cracked. Next is adding voronoi cell scalar at same scales and seeds and after making it absolute, can use it as a mask for where mudcracks have collapsed. Next is basically my fractures and tension shader adjusted for the setup. Optionally warped by the mudcracks to conform the cracks to the voronoi shapes. Then we add some microdisplacement, and colour shaders. A lot of shader bloat is in conditional to allow for user configuration otherwise they would be lost in that node network without knowing how to disable/enable/adjust settings. All the effects of course varied by Perlin for natural effect too.
Title: Re: Mudcracks Shader
Post by: WAS on October 23, 2020, 01:52:58 PM
A note a lot of these iterated/phased shaders could go by much faster and easier is we had a loop conditional, as well as vector inputs for transform shaders. There are dozens of ideas I've wanted to do but TG lacks the logic and control. A conditional colour/shader node alone would help decrease the frequency of conditional scalars used too.