Non Photorealistic Shading


This page shows geometries shaded in such a way that they are intended to mimic the "look" of images recorded by a scanning electron microscope.



Final Animation



Side by Side Comparison



Other Reference Images



Visualizer (with displacement) and Occlusion Render (without displacement)



Technical Issues and Problems

Predominately the issues I had with this project related mostly with displacement. I had to fight a decent bit to blend my displacement texturemap as well as the procedural noises that were being used.I ended up solving this issue by essentially restarting my modeling on the salt cubes. I found that integrating multiple displacement maps from zBrush Quixel and procedural noise caused a great deal of poly cracking along the seams of UVs and in some instances wouldn't even show up. By redoing my models and removing the seams with a normal-based projection method and running the maps through a PxrSeExpre expression node I was able to mix, control and balance multiple shaders that were in use. The Pixar Expression node read as such and please see below images for connected attributes.

rockDisp = floatInput2 - .5;
alligatorSalt = floatInput1 - .3;
resultF = rockDisp + alligatorSalt/3;
resultF
			

The additional shaders where acquired from https://docs.chaosgroup.com/display/OSLShaders/Additional+procedural+noises#Additionalproceduralnoises-Simplexflownoise%C2%A0

SEM_Edge.html
Alligator_Noise.html
Fractal_Simplex.html (Chaos Group Collective)


Salt Shader Hypershade Network, Snow Shader Hypershade Network, and Back Plane Hypershade Network



Self Evaluation

Overall I would like to believe that I succeeded in this assignment. Despite my technical hurdles with displacement I was able to solve those as stated above in the technical issues section, more interestingly however is how I managed the more interesting elements, the rainbows and the snow-spikes. The Snow Spikes were a familiar thing for me. I used a selection set to select the desired faces by which I wanted to scatter my spikes across and, using MASH, I controlled the angle and scale by which the spikes were oriented, and then by using a color ID node and exporting vertex colors the spikes inherited the colors from the snowflake beneath them. Though what of those rainbows? How'd those get there? Well, I used a ramp and fed it into multiple edge shaders of different levels of strength. By taking advantage of the curvature strength I could create my own layers of color to be fed into my base color, wherein I can add the edge color to be a flat semi-complimentary color. I chose the flat color as the edge so that we could still see the fake "lighting" affect across the varied surface colors. Were I to do this project again I think I would've addressed the snow flakes more than the salt. Compositionally I enjoyed the layout of having the seperate snow and salt crystals, but in many of the snow reference images the formations were much more intertwined and the interplay there may have been more of a modelling challenge than working off of several different basemeshes that I sculpted and then positioned differently.