Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
CSE168 Computer Graphics II, Rendering Spring 2006 Matthias Zwicker Last time • Realistic camera models • High dynamic range imaging Pinhole cameras Pinhole cameras Problems • Small pinhole gathers little light, requires long exposure • Larger pinhole reduces sharpness Lenses Pinhole Lens 6 sec. exposure 0.01 sec exposure Thin lens model • Approximative model for well-behaved lenses • All parallel rays converge at focal length • Rays through the center are not deflected Thin lens model • All rays passing through a single point on a plane at distance in front of the lens will pass through a single point at distance behind the lens • Thin lens formula Depth of field • Blurriness of out of focus objects depends on aperture size Aperture Depth of field Ray tracing using a thin lens model • Place image plane at distance D from lens plane • Generate primary rays with random origin on lens aperture Pinhole Thin lens Object in focus Primary rays Image plane Primary rays Image plane HDR imaging • Image sensors, display devices have a smaller dynamic range than radiance in real/simulated scenes • Dynamic range: ratio of brightest to darkest pixel in an image HDR photography • Acquire several images with different exposures • Recover a HDR intensity for each pixel [Wikipedia] Tone mapping • Compress dynamic range of image without losing detail for low dynamic range display [Wikipedia] Multi-scale methods • Decompose image into coarse and detail layer • Compress only coarse layer, then recombine Today • Participating media • Subsurface scattering Participating media • So far: light reflection on surfaces, BRDFs • What about interaction of light with volumetric media? [Greenler, “Rainbows, halos, and glories”] Participating media Applications • Clouds, smoke, water • Subsurface scattering (paint, skin) • Scientific/medical visualization (CT, MRI, etc.) Participating media Topics • Absorption, out-scattering, in-scattering, emission • Phase functions • Volume rendering equation • Ray tracing volumes – Absorption/emission – Single scattering Absorption • Absorption cross-section • Rate of absorption Out-scattering • Scattering coefficient • Rate of being scattered Extinction • Extinction: absorption and out-scattering Extinction [Pharr, Humphreys] Transmittance • Optical thickness • Transmittance Transmittance • Multiplicative property • Homogeneous media Beer’s law In-scattering • Phase function Phase functions • Phase angle • Isotropic • Rayleigh: scattering from particles smaller than wavelength of light • Mie scattering: scattering from small spheres, larger than wavelength Henyey-Greenstein phase function • Empirical phase function • Useful for many phenomena (water, clouds, skin, stone) • Average phase angle , asymmetry parameter Backward scattering Forward scattering Phase functions • Unitless • Reciprocity • Energy conservation • Amount of scattering is controlled by Blue sky, red sunset [Greenler] The volume rendering equation • Integro-differential form • Integro-integral form Extinction (absorption, out-scattering) Source (emission, in-scattering) The volume rendering equation • Full solution is very expensive to compute, intractable for complicated scenes • Rendering with simplified models • Emission/absorption only • Single scattering Absorption/emission only • No in-scattering Volume emission • Remember transmittance Ray marching • Monte Carlo approximation • Uniformly distributed samples Ray marching • Incremental computation of transmittance Ray marching T = 1 L = 0 ds = (s_out – s_in) / N for( s = s_in; s <= s_out; s += ds ) { L = L + T * L_ve T = T * ( 1 – sigma_t(s) ) * ds } L = L * ds Absorption/emission only [Pharr, Humphreys] Volume visualization • Visualize scalar data on volumetric grid • CT, MRI scans, … • Scalar data represents some physical property of a material • Medical/industrial applications • Transfer function assigns color/opacity value to each scalar value Volume visualization Volume rendering: ray marching/ray casting as described before Single scattering • Include in-scattering of incident radiance due to direct illumination In-scattering, • Direct illumination from light sources • Phase function • Scattering coefficient Single scattering • Direct illumination needs to be attenuated • Shoot shadow rays Shadow rays T = 1 dt = t_out – t_in / N for( t = t_in; t <= t_out; t += dt ) { T = T * ( 1 – sigma_t(t) ) * dt } L_d = T * L_s // L_s: light source S = sigma_s * p * L_d // p: phase function Single scattering [Pharr, Humphreys] Beams of light [Greenler] [Minneart] Multiple scattering • Path tracing, very slow • Photon mapping “Efficient Simulation of Light Transport in Scenes with Participating Media using Photon Maps”, Jensen, Christensen “Realistic Image Synthesis using Photon Mapping”, Jensen Photon mapping • Volume caustics [Jensen] Questions? Subsurface scattering • So far: BRDFs, light scatters exactly at the point where it hits the surface • Subsurface scattering: light enters the material, bounces around, leaves at a different place • Describe scattering properties of material using an extension of BRDFs Subsurface scattering • BSSRDF: bidirectional surface scattering reflectance distribution function BRDF BSSRDF Subsurface scattering • BSSRDF has 8 degrees of freedom (2 positions, 2 orientations) • Hard to capture in the general case • Brute force Monte Carlo simulation very expensive Subsurface scattering Diffusion approximation • Light distribution in highly scattering media tends to become isotropic • We can a find a diffuse BSSRDF where • Also known as “dipole model” Subsurface scattering Diffusion approximation [Jensen et al.] Diffusion profile Subsurface scattering BSSRDF BRDF [Jensen et al.] Subsurface scattering BRDF BSSRDF Monte Carlo simulation [Jensen et al.] Efficient rendering • “A Rapid Hierarchical Rendering Technique for Translucent Materials”, Jensen, Buhler • Two pass approach • First, compute irradiance at a set of surface points • Second, evaluate diffuse BSSRDF by gathering irradiance samples • Use efficient hierarchical scheme Efficient rendering • Irradiance gathering using the diffuse BSSRDF Irradiance sample Reflected radiance Efficient rendering Irradiance samples Final image [Jensen, Buhler] Efficient rendering • Global illumination and subsurface scattering [Jensen, Buhler] Next time • Guest lecture