You are viewing the course site for a past offering of this course. The current offering may be found here.
Lecture 19: Intro to Physical Simulation (6)
kmkranen

To tie this back to what we learned earlier this semester about rendering, rendering hair is actually very complicated for multiple reasons. One is simply because of the shape of human hair! Each hair is actually many overlapping scales, making the surface of each hair very bumpy on a microscopic scale. Additionally, each hair refracts and absorbs light! Lastly, different colors of hair scatter light differently. If you're interested in how models overcome these difficulties, check out this paper! http://www.graphics.stanford.edu/papers/hair/hair-sg03final.pdf

rsha256

Why is it hard to check if a hair strand intersects with another hair strand and stop it if so? It doesn’t seem like computation is an issue with GPUs readily available nowadays (and how we mentioned the usage of Navier-Stokes to model sea water in a previous slide — which sounds much more computationally intensive).

wangdotjason

Nvidia has done some interesting work on hair rendering in the past decade. Most notably, their work on Nvidia HairWorks, which aims to render realistic looking hair in real time. HairWorks primarily usess tesselation to make hair appear smoother and curvey, and interpolation to add more hair volume without having to compute more individual hairs. (https://developer.download.nvidia.com/presentations/2008/SIGGRAPH/RealTimeHairSimulationAndRenderingOnGPU.pdf)

SeanW0823

How high does the resolution of the spring model have to be for the hair to look real? Would we have to take into account how hair thickness varies from its root to the end?

l-winston

modeling hair seems like a very challenging task given the complex and highly varied physical properties of hair, not to ment ion the large amount of primitives involved

ShaamerKumar

I remember earlier how rendering hair was very difficult for movies like tangled, I wonder how they have optimised this

StaffDanCubed

Might be a random question, but I was looking at all the comments and just realized that "human hair" really cannot be represented by a single model because of how much it varies under different conditions. Wet hair behaves very differently from dry hair, dirty hair from clean hair, etc. I wonder if there has been efforts to model hair under these variety of conditions? And if so what are the challenges presented by them?

ess3ncez

To optimize rendering performance, LOD (level of detail) techniques are used to reduce the complexity of hair models depending on the distance from the camera. When a character is far away from the camera, their hair will be rendered with fewer individual strands, while close-up shots will have more detailed hair. (Very similar to the idea of Mipmaps!)

LeslieTrue

Question: I know that hair or fur should be difficult to render but I'm wondering how do artist usually do when creating the "backbone" of these hair before rendering? Do they typically draw it piece by piece?

rheask8246

Animating curly hair was actually quite difficult in the past. Pixar developed new software to simulate curly hair, based on an elastic rod model. Here's a paper on how they did it: https://graphics.pixar.com/library/CurlyHairA/paper.pdf

You must be enrolled in the course to comment