研究目的
To develop a new theory of volumetric light transport for media with non-exponential free-flight distributions, enabling practical rendering algorithms that enrich the level of control over the appearance of participating media.
研究成果
The paper introduces a comprehensive theory for non-exponential volumetric light transport, supporting surfaces and heterogeneity while being reciprocal and practical. It enables a rich range of appearances through parametric and fractal-based models, with implementations showing accurate results and potential for artistic control and efficient rendering in graphics applications.
研究不足
The theory assumes a single, globally fixed free-flight PDF, limiting handling of multiple overlapping media with different transmittance functions. Efficient, unbiased distance sampling methods are not fully generalized for all non-exponential heterogeneous media. Navigating the appearance space is challenging due to non-orthogonal parameters.
1:Experimental Design and Method Selection:
The methodology involves formulating a non-exponential path integral by averaging over stochastic media realizations, using assumptions from neutron transport and atmospheric sciences. It includes deriving a reciprocal path integral and a practical heterogeneity model.
2:Sample Selection and Data Sources:
The study uses synthetic media models, including discrete point distributions and continuous density fields with correlations (e.g., fractal noise), generated computationally.
3:List of Experimental Equipment and Materials:
Computational resources and rendering software (PBRTv3 and Tungsten) are used; no specific physical equipment is mentioned.
4:Experimental Procedures and Operational Workflow:
The process involves Monte Carlo simulations to compute ensemble-averaged transmittance, implementing the path integral in rendering systems, and testing with various parametric models and fractal noise.
5:Data Analysis Methods:
Analysis includes comparing rendered results to ground truth via Monte Carlo experiments, verifying reciprocity, and evaluating performance metrics like rendering speed.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容