We present Optimised Path Space Regularisation (OPSR), a novel regularisation technique for forward path tracing algorithms. Our regularisation controls the amount of roughness added to materials depending on the type of sampled paths and trades a small error in the estimator for a drastic reduction of variance in difficult paths, including indirectly visible caustics. We formulate the problem as a joint bias‐variance minimisation problem and use differentiable rendering to optimise our model. The learnt parameters generalise to a large variety of scenes irrespective of their geometric complexity. The regularisation added to the underlying light transport algorithm naturally allows us to handle the problem of near‐specular and glossy path chains robustly. Our method consistently improves the convergence of path tracing estimators, including state‐of‐the‐art path guiding techniques where it enables finding otherwise hard‐to‐sample paths and thus, in turn, can significantly speed up the learning of guiding distributions.

Russian roulette and splitting are widely used techniques to increase the efficiency of Monte Carlo estimators. But, despite their popularity, there is little work on how to best apply them. Most existing approaches rely on simple heuristics based on, e.g., surface albedo and roughness. Their efficiency often hinges on user-controlled parameters. We instead iteratively learn optimal Russian roulette and splitting factors during rendering, using a simple and lightweight data structure. Given perfect estimates of variance and cost, our fixed-point iteration provably converges to the optimal Russian roulette and splitting factors that maximize the rendering efficiency. In our application to unidirectional path tracing, we achieve consistent and significant speed-ups over the state of the art.

We introduce a practical general-purpose neural appearance filtering pipeline for physically-based rendering. We tackle the previously difficult challenge of aggregating visibility across many levels of detail from local information only, without relying on learning visibility for the entire scene. The high adaptivity of neural representations allows us to retain geometric correlations along rays and thus avoid light leaks. Common approaches to prefiltering decompose the appearance of a scene into volumetric representations with physically-motivated parameters, where the inflexibility of the fitted models limits rendering accuracy. We avoid assumptions on particular types of geometry or materials, bypassing any special-case decompositions. Instead, we directly learn a compressed representation of the intra-voxel light transport. For such high-dimensional functions, neural networks have proven to be useful representations. To satisfy the opposing constraints of prefiltered appearance and correlation-preserving point-to-point visibility, we use two small independent networks on a sparse multi-level voxel grid. Each network requires 10--20 minutes of training to learn the appearance of an asset across levels of detail. Our method achieves 70--95% compression ratios and around 25% of quality improvements over previous work. We reach interactive to real-time framerates, depending on the level of detail.