A new microflake model with microscopic self‐shadowing for accurate volume downsampling

Abstract

Naïve linear methods for downsampling high‐resolution microflake volumes often produce inaccurate appearance, especially when input voxels are very opaque. Preserving correct appearance at all resolutions requires taking into account maskingshadowing effects that occur between and inside dense input voxels. We introduce a new microflake model whose additional parameters characterize self‐shadowing effects at a microscopic scale. We provide an anisotropic self‐shadowing function and microflake distributions for which the scattering coefficients and the phase functions of our model have closed‐form expressions. We use this model in a new downsampling approach in which scattering parameters are computed from local estimations of self‐shadowing probabilities in the input volume. Unlike previous work, our method handles datasets with spatially varying scattering parameters, semi‐transparent volumes and datasets with intricate silhouettes. We show that our method generates LoDs with correct transparency and consistent appearance through scales for a wide range of challenging datasets, allowing for huge memory savings and efficient distant rendering without loss of quality.

Real-time Rendering of Small-scale Volumetric Structure on Animated Surfaces

“…For example, when the texture represents some plane, it should be opaque when the view direction is perpendicular to that plane, and almost transparent, when it's parallel. There is some research about correct handling of downscaled volumetric textures [21,22], but it focuses on offline rendering. So adopting it to real-time is a topic for future work.…”

SpongeCake: A Layered Microflake Surface Appearance Model

“…The microflake model is successful in representing fibers and fabrics [Zhao et al 2011]. Other work has also benefited from using microflakes for representing other types of materials such as foliage [Loubet and Neyret 2018], or special pigments [Guillén et al 2020]. Heitz et al [2015] introduced a versatile representation named symmetric GGX (SGGX) to efficiently describe microflake distributions using 3× 3 positive-definite matrices, allowing for surface-like and fiber-like microflakes, including convenient control over their orientations.…”

Rendering Discrete Participating Media with Geometrical Optics Approximation

“…Unfortunately, these methods are highly data-intensive and plagued by heavy computation. To improve the performance while maintaining good accuracy, some downsampling strategies [42], [43] are developed. Current rendering solutions for granular materials are also based on explicit geometries and pre-captured optical properties of each individual grain [12], [13], [14].…”