Real-Time Rendering of Glints in the Presence of Area Lights

TL;DR

This work advances real-time rendering of glittery materials under area-light illumination by introducing a physically motivated probability $p_{\Omega_h}$ that a microfacet is correctly oriented to reflect light from an area source to the observer. The method combines a discrete NDF $\hat{D}_{\mathcal{P}}$ with a binomial counting model, and integrates LTC-based approximations for area-light shading to achieve efficient, scalable performance. Key contributions include a formal discrete microfacet theory, an area-light extension that preserves light-source size effects, and practical LTC-based implementations with minimal overhead. The approach enables realistic glints under spatially diffuse area lights and can adapt to point or directional lights via a tunable light-size parameter, making it suitable for real-time rendering pipelines. Overall, the framework unifies stochastic glint modeling with LTC-area-light shading to produce visually accurate glitter with low runtime cost.

Abstract

Many real-world materials are characterized by a glittery appearance. Reproducing this effect in physically based renderings is a challenging problem due to its discrete nature, especially in real-time applications which require a consistently low runtime. Recent work focuses on glittery appearance illuminated by infinitesimally small light sources only. For light sources like the sun this approximation is a reasonable choice. In the real world however, all light sources are fundamentally area light sources. In this paper, we derive an efficient method for rendering glints illuminated by spatially constant diffuse area lights in real time. To this end, we require an adequate estimate for the probability of a single microfacet to be correctly oriented for reflection from the source to the observer. A good estimate is achieved either using linearly transformed cosines (LTC) for large light sources, or a locally constant approximation of the normal distribution for small spherical caps of light directions. To compute the resulting number of reflecting microfacets, we employ a counting model based on the binomial distribution. In the evaluation, we demonstrate the visual accuracy of our approach, which is easily integrated into existing real-time rendering frameworks, especially if they already implement shading for area lights using LTCs and a counting model for glint shading under point and directional illumination. Besides the overhead of the preexisting constituents, our method adds little to no additional overhead.

Figures (9)

• Figure 1: The effect of real-world area lights is shown on a sample of car paint with metallic flakes. (\ref{['fig:photo_flakes:setup']}) The camera and a smartphone display are placed $50cm$ and $30cm$ above the sample. We took two exposure series illuminated by a small (\ref{['fig:photo_flakes:10']}) and a large (\ref{['fig:photo_flakes:35']}) light source, respectively. The shapes of the light sources have been realized using the display of a smartphone, and the reconstructed HDR images have been adjusted in post-processing such that the light sources emit an equal amount of radiant power. The experiment shows that the smaller light source results in sparser but brighter glints, and the larger light source produces a larger number of weaker glints.
• Figure 2: The light source (red) defines a spherical polygon ${{\Omega_i}}$ (green) that is transformed to the halfway vectors ${{\Omega_h}}$ (purple) over which the continuous NDF is integrated.
• Figure 3: Comparison of lobe shapes defined by the continuous BRDF and NDF, as well as their LTC approximations for $\sqrt{\alpha} \in \{0.5,0.8\}$ and $\theta_o \in \{\ang{30},\ang{75}\}$ projected into the tangent plane. The top difference images of the BRDF, BRDF LTC and NDF LTC columns show the signed difference with the NDF. The bottom difference image of BRDF LTC shows the signed difference with the BRDF.
• Figure 4: Given viewing direction ${\omega_o}$ and roughness $\alpha$, (\ref{['fig:fgd:brdf']}) we plot the directional albedo ${\mathrm{FGD}}$ of the BRDF for ${{{F}\mathopen{}\mathclose{\left({\omega_i}, {\mathbf{h}}\right)}}}=1$ and (\ref{['fig:fgd:ndf']}) the potentially reflecting microfacet area ${\mathcal{D}_{\mathrm{PR}}}$. (\ref{['fig:fgd:diff']}) We plot the difference after scaling ${\mathrm{FGD}}$ to minimize the squared error with ${\mathcal{D}_{\mathrm{PR}}}$.
• Figure 5: We compare the appearance of our model (\ref{['sec:infinitesimal_lights']}) with the baseline deliot2023real for perceptually linear burley2012physically roughness $\sqrt{\alpha} \in \{0.1, 0.5, 0.9\}$ under directional illumination. We show two light opening angles $\gamma \in \{\ang{0.26}, \ang{5}\}$ with ${|{{\Omega_i}}|} = 2\pi(1-\cos\gamma) \,[\unit{\steradian}]$, and determine the respective value for $R$ in the baseline method by asserting equality for the success probability of the binomial distribution at normal incidence given $\gamma$ and $\alpha$. The left half of the figure assumes a narrow sun-like light source. The right half of the figure assumes a larger light source where our assumptions produce invalid ${{p_{{{\Omega_h}}}}} > 1$ for small $\sqrt{\alpha}=0.1$ and large $\gamma=\ang{5}$.