计算机图形学-4.ppt

Motion Blur Test Cases regular grid, 1 sample/pixel jittered grid, 1 sample/pixel regular grid, 16 samples/pixel jittered grid, 16 samples/pixel * Monte Carlo Path Tracing: Pros and Cons Arbitrary geometry – point primitive (just ray—object intersection). Very low memory consumption. Theoretically “best” possible with infinite (time) resources. Paths in both time and space. View dependent (paths start at eye). Does not reuse illumination information. Not efficient for diffuse surfaces. Slow. Noisy. Expensive on caustics. May have high image variance. * Insert the two cases of using other PDFs to find the cosine estimator Projection into Hemisphere Uniformly sample points on the unit disk and just project them up onto the unit hemisphere. Vertical direction: Circular direction: * H2(N) samples ? ? Forming the Sample Ray To form sample ray: Have surface point P, Hemisphere sample point is S=(x,y,z) where (x,y) comes from (r,?) sample: And * H2(N) ? ? S P Back to Monte Carlo Path Tracing Summary Choose sample paths according to an appropriate Probability Density Function. Evaluate function at sample. Average these appropriately weighted values using a suitable estimator. Error ? slow convergence (twice as good ? need 4 times more samples). Variance in estimator ? noise in image. Good for high dimensional functions. Theoretically best we can do (good estimator and more samples ? better image). * Monte Carlo Path Tracing For each pixel, estimate radiance value as (appropriately chosen and weighted average of (lots of)) random paths along rays from eye through the pixel. Compute radiance value for each visible point from Rendering Equation: Self Emission Direct Illumination (shadow rays) Indirect Illumination (Recursive) Keep lengths of recursive paths manageable – Russian Roulette. * Cutler Durand, MIT More Samples = Higher Cost but Less Noise! * 8 samples/pixel 40 samples/pixel 200 samples/pixel 13 sec