|Ravi Ramamoorthi||Computer Graphics Laboratory ,|
|Pat Hanrahan||Stanford University|
We present a new method for real-time
rendering of objects with complex isotropic BRDFs under distant
natural illumination, as specified by an environment map. Our
approach is based on spherical frequency space analysis and includes
three main contributions. Firstly, we are able to theoretically
analyze required sampling rates and resolutions, which have
traditionally been determined in an ad-hoc manner. We also introduce
a new compact representation, which we call a spherical harmonic
reflection map (SHRM) , for efficient representation and rendering.
Finally, we show how to rapidly
prefilter the environment map to compute the SHRM ---our
frequency domain prefiltering algorithm is generally orders of
magnitude faster than previous angular (spatial) domain approaches.
SummaryOur goals are real-time rendering with complex natural illumination and realistic, possibly measured, BRDFs. This paper introduces a new frequency space paradigm for prefiltering and rendering environment mapped images with general isotropic BRDFs. Our approach is based on recent theoretical results by Basri and Jacobs and Ramamoorthi and Hanrahan, wherein they formalize the notion of reflection as a spherical convolution of the illumination and BRDF. We show that frequency space analysis allows for setting sampling rates accurately, and enables compact representations. Further, just as image convolutions are often computed in the Fourier rather than the spatial domain, prefiltering is more efficient in frequency rather than angular space. Our main contributions are:
ResultsThe images on the right illustrate some of our results. More information is found in the paper. Clicking on each of these figures will bring up a high-resolution version.
Figure 1 This figure includes real lighting (measured in the Grace Cathedral, courtesy of Paul Debevec) and many different BRDF models including krylon blue (using McCool et al.'s reconstruction from measurements at Cornell) and the LaFortune model. These images were each rendered at approximately 30 frames per second using our Spherical Harmonic Reflection Map (SHRM) representation.
Figure 2 An illustration of the reflective reparameterization used by us. Reparameterization involves recentering about the reflection vector. BRDFs become more compact, and in special cases (Phong) become 1D functions.
Figure 3 An Overview of our entire pipeline. The inputs are measured illumination and BRDF values. We then find their spherical harmonic coefficients, perform convolution by multiplication in the frequency domain and output the spherical harmonic reflection map or SHRM.
Figure 4 An illustration of SHRMs. Each pixel in a standard reflection cubemap actually stores a low-frequency distribution, making it a view-dependent reflection map. For rendering, we use a dynamic cubemap.
Relevant LinksSiggraph 2002 paper in Gzipped Postscript (4.2M) or PDF (3.3M)