Rendering - Experimental Ideas & Implementations

Permanent URI for this community

https://diglib.eg.org/handle/10.2312/12635

Browse

Now showing 1 - 20 of 137

4D-rasterization for Fast Soft Shadow Rendering
(The Eurographics Association, 2016) Wang, Lili; Zhao, Qi; Meng, Chunlei; Popescu, Voicu; Elmar Eisemann and Eugene Fiume
This paper describes an algorithm for rendering soft shadows efficiently by generalizing conventional triangle projection and rasterization from 2D to 4D. The rectangular area light source is modeled with a point light source that translates with two degrees of freedom. This generalizes the projection of triangles and of output image samples, as seen from the light, to the locus of projections as the light translates. The generalized projections are rasterized to determine a conservative set of sample/ triangle pairs, which are then examined to derive light occlusion masks for each sample. The algorithm is exact in the sense that each element of the occlusion mask of a sample is computed accurately by considering all potentially blocking triangles. The algorithm does not require any type of precomputation so it supports fully dynamic scenes. We have tested our algorithm on several scenes to render complex soft shadows accurately at interactive rates.
Adaptive Multi-view Path Tracing
(The Eurographics Association, 2019) Fraboni, Basile; Iehl, Jean-Claude; Nivoliers, Vincent; Bouchard, Guillaume; Boubekeur, Tamy and Sen, Pradeep
Rendering photo-realistic image sequences using path tracing and Monte Carlo integration often requires sampling a large number of paths to get converged results. In the context of rendering multiple views or animated sequences, such sampling can be highly redundant. Several methods have been developed to share sampled paths between spatially or temporarily similar views. However, such sharing is challenging since it can lead to bias in the final images. Our contribution is a Monte Carlo sampling technique which generates paths, taking into account several cameras. First, we sample the scene from all the cameras to generate hit points. Then, an importance sampling technique generates bouncing directions which are shared by a subset of cameras. This set of hit points and bouncing directions is then used within a regular path tracing solution. For animated scenes, paths remain valid for a fixed time only, but sharing can still occur between cameras as long as their exposure time intervals overlap. We show that our technique generates less noise than regular path tracing and does not introduce noticeable bias.
Additional Progress Towards the Unification of Microfacet and Microflake Theories
(The Eurographics Association, 2016) Dupuy, Jonathan; Heitz, Eric; d'Eon, Eugene; Elmar Eisemann and Eugene Fiume
We study the links between microfacet and microflake theories from the perspective of linear transport theory. In doing so, we gain additional insights, find several simplifications and touch upon important open questions as well as possible paths forward in extending the unification of surface and volume scattering models. First, we introduce a semi-infinite homogeneous exponential-free-path medium that (a) produces exactly the same light transport as the Smith microsurface scattering model and the inhomogeneous Smith medium that was recently introduced by Heitz et al, and (b) allows us to rederive all the Smith masking and shadowing functions in a simple way. Second, we investigate in detail what new aspects of linear transport theory enable a volume to act like a rough surface. We show that this is mostly due to the use of non-symmetric distributions of normals and explore how the violation of this symmetry impacts light transport within the microflake volume without breaking global reciprocity. Finally, we argue that the surface profiles that would be consistent with very rough Smith microsurfaces have geometrically implausible shapes. To overcome this, we discuss an extension of Smith theory in the volume setting that includes NDFs on the entire sphere in order to produce a single unified reflectance model capable of describing everything from a smooth flat mirror all the way to a semi-infinite isotropically scattering medium with both low and high roughness regimes in between.
Ambient Dice
(The Eurographics Association, 2017) Iwanicki, Michal; Sloan, Peter-Pike; Matthias Zwicker and Pedro Sander
We present a family of basis functions designed to accurately and efficiently represent illumination signals on the unit sphere. The bases are built of locally supported functions, needing three to six basis functions in a given direction. This minimizes the number of memory transactions and bandwidth requirements needed for reconstruction. There are three variations of our basis. All are based on storing coefficients at the 12 vertices of an icosahedron. The first one stores the values directly, together with their directional derivatives and hybrid Bézier patches are used for interpolation. This allows one to achieve quality comparable to 3rd-5th order spherical harmonics while still requiring 27 coefficients for the reconstruction. The second variation encodes the signal in YCoCg space and uses a reduced quality, linear reconstruction for the chromaticity components - requiring only 15 coefficients while marginally reducing the quality. The third option exploits the partition of unity formed by cos2 and cos4 restricted to a hemisphere oriented along the directions of the icosahedron vertices. It uses 18 coefficients for the reconstruction, but trades the additional bandwidth requirements for simpler calculations. The quality of that version is still comparable to 3rd order spherical harmonics (SH). We name the basis Ambient Dice as a reference to both: the Ambient Cube basis - as ours is an extension of some of its properties - and the 20-sided dice commonly used in pen-and-paper role-playing games, which is an icosahedron.
Apex Point Map for Constant-Time Bounding Plane Approximation
(The Eurographics Association, 2015) Laine, Samuli; Karras, Tero; Jaakko Lehtinen and Derek Nowrouzezahrai
We introduce apex point map, a simple data structure for constructing conservative bounds for rigid objects. The data structure is distilled from a dense k-DOP, and can be queried in constant time to determine a tight bounding plane with any given normal vector. Both precalculation and lookup can be implemented very efficiently on current GPUs. Applications include, e.g., finding tight world-space bounds for transformed meshes, determining perobject shadow map extents, more accurate view frustum culling, and collision detection.
Appearance-Driven Automatic 3D Model Simplification
(The Eurographics Association, 2021) Hasselgren, Jon; Munkberg, Jacob; Lehtinen, Jaakko; Aittala, Miika; Laine, Samuli; Bousseau, Adrien and McGuire, Morgan
We present a suite of techniques for jointly optimizing triangle meshes and shading models to match the appearance of reference scenes. This capability has a number of uses, including appearance-preserving simplification of extremely complex assets, conversion between rendering systems, and even conversion between geometric scene representations. We follow and extend the classic analysis-by-synthesis family of techniques: enabled by a highly efficient differentiable renderer and modern nonlinear optimization algorithms, our results are driven to minimize the image-space difference to the target scene when rendered in similar viewing and lighting conditions. As the only signals driving the optimization are differences in rendered images, the approach is highly general and versatile: it easily supports many different forward rendering models such as normal mapping, spatially-varying BRDFs, displacement mapping, etc. Supervision through images only is also key to the ability to easily convert between rendering systems and scene representations. We output triangle meshes with textured materials to ensure that the models render efficiently on modern graphics hardware and benefit from, e.g., hardware-accelerated rasterization, ray tracing, and filtered texture lookups. Our system is integrated in a small Python code base, and can be applied at high resolutions and on large models. We describe several use cases, including mesh decimation, level of detail generation, seamless mesh filtering and approximations of aggregate geometry.
Approximate svBRDF Estimation From Mobile Phone Video
(The Eurographics Association, 2018) Albert, Rachel A.; Chan, Dorian Yao; Goldman, Dan B.; O'Brien, James F.; Jakob, Wenzel and Hachisuka, Toshiya
We describe a new technique for obtaining a spatially varying BRDF (svBRDF) of a flat object using printed fiducial markers and a cell phone capable of continuous flash video. Our homography-based video frame alignment method does not require the fiducial markers to be visible in every frame, thereby enabling us to capture larger areas at a closer distance and higher resolution than in previous work. Pixels in the resulting panorama are fit with a BRDF based on a recursive subdivision algorithm, utilizing all the light and view positions obtained from the video. We show the versatility of our method by capturing a variety of materials with both one and two camera input streams and rendering our results on 3D objects under complex illumination.
Bi-Directional Polarised Light Transport
(The Eurographics Association, 2016) Mojzík, Michal; Skřivan, Tomáš; Wilkie, Alexander; Křivánek, Jaroslav; Elmar Eisemann and Eugene Fiume
While there has been considerable applied research in computer graphics on polarisation rendering, no principled investigation of how the inclusion of polarisation information affects the mathematical formalisms that are used to describe light transport algorithms has been conducted so far. Simple uni-directional rendering techniques do not necessarily require such considerations: but for modern bi-directional light transport simulation algorithms, an in-depth solution is needed. In this paper, we first define the transport equation for polarised light based on the Stokes Vector formalism. We then define a notion of polarised visual importance, and we show that it can be conveniently represented by a 4 4 matrix, similar to the Mueller matrices used to represent polarised surface reflectance. Based on this representation, we then define the adjoint transport equation for polarised importance. Additionally, we write down the path integral formulation for polarised light, and point out its salient differences from the usual formulation for light intensities. Based on the above formulations, we extend some recently proposed advanced light transport simulation algorithms to support polarised light, both in surface and volumetric transport. In doing that, we point out optimisation strategies that can be used to minimise the overhead incurred by including polarisation support into such algorithms.
The Challenges of Releasing the Moana Island Scene
(The Eurographics Association, 2019) Tamstorf, Rasmus; Pritchett, Heather; Boubekeur, Tamy and Sen, Pradeep
A tremendous amount of research has been done over the years using the Stanford bunny, the Cornell box and recently somewhat more complicated data sets. Yet, none of these data sets come close to representing the complexity that production houses and film studios handle on a daily basis. In recent years industry members have lamented this lack of realistic examples, and in return academics have requested that more representative examples be made available. Both of these points are valid, which in turn has led to the release of the Moana Island Scene dataset. However, while it sounds simple, the actual release of such data leads to numerous philosophical and practical questions. The goal of this paper is to present some of the challenges associated with releasing production data for academic use.
Color Clipping and Over-exposure Correction
(The Eurographics Association, 2015) Abebe, Mekides Assefa; Pouli, Tania; Kervec, Jonathan; Larabi, Chaker; Jaakko Lehtinen and Derek Nowrouzezahrai
Limitations of the camera or extreme contrast in scenes can lead to clipped areas in captured images. Irrespective of the cause, color clipping and over-exposure lead to loss of texture and detail, impacting the color appearance and visual quality of the image. We propose a new over-exposure and clipping correction method, which relies on the existing correlation between RGB channels of color images to recover clipped information. Using a novel region grouping approach, clipped regions are coherently treated both spatially and temporally. To reconstruct over-exposed areas where all channels are clipped we employ a brightness profile reshaping scheme, which aims to preserve the appearance of highlights, while boosting local brightness. Our method is evaluated using objective metrics as well as a subjective study based on an ITU standardized protocol, showing that our correction leads to improved results compared to previous related techniques. We explore several potential applications of our method, including extending to video as well as using it as a preprocessing step prior to inverse tone mapping.
A Compact Representation for Fluorescent Spectral Data
(The Eurographics Association, 2021) Hua, Qingqin; Fichet, Alban; Wilkie, Alexander; Bousseau, Adrien and McGuire, Morgan
We propose a technique to efficiently importance sample and store fluorescent spectral data. Fluorescence behaviour is properly represented as a re-radiation matrix: for a given input wavelength, this matrix indicates how much energy is re-emitted at all other wavelengths. However, such a 2D representation has a significant memory footprint, especially when a scene contains a high number of fluorescent objects, or fluorescent textures. We propose to use Gaussian Mixture Domain to model re-radiation, which allows us to significantly reduce the memory footprint. Instead of storing the full matrix, we work with a set of Gaussian parameters that also allow direct importance sampling. When accuracy is a concern, one can still use the re-radiation matrix data, and just benefit from importance sampling provided by the Gaussian Mixture. Our method is useful when numerous fluorescent materials are present in a scene, an in particular for textures with fluorescent components.
Computing Manifold Next-Event Estimation without Derivatives using the Nelder-Mead Method
(The Eurographics Association, 2024) Granizo-Hidalgo, Ana; Holzschuch, Nicolas; Haines, Eric; Garces, Elena
Specular surfaces, by focusing the light that is being reflected or refracted, cause bright spots in the scene, called caustics. These caustics are challenging to compute for global illumination algorithms. Manifold-based methods (Manifold Exploration, Manifold Next-Event Estimation, Specular Next Event Estimation) compute these caustics as the zeros of an objective function, using the Newton-Raphson method. They are efficient, but require computing the derivatives of the objective function, which in turn requires local surface derivatives around the reflection point, which can be challenging to implement. In this paper, we leverage the Nelder-Mead method to compute caustics using Manifold Next-Event Estimation without having to compute local derivatives. Our method only requires local evaluations of the objective function, making it an easy addition to any path-tracing algorithm.
Constrained Spectral Uplifting for HDR Environment Maps
(The Eurographics Association, 2024) Tódová, Lucia; Wilkie, Alexander; Haines, Eric; Garces, Elena
Spectral representation of assets is an important precondition for achieving physical realism in rendering. However, defining assets by their spectral distribution is complicated and tedious. Therefore, it has become general practice to create RGB assets and convert them into their spectral counterparts prior to rendering. This process is called spectral uplifting. While a multitude of techniques focusing on reflectance uplifting exist, the current state of the art of uplifting emission for image-based lighting consists of simply scaling reflectance uplifts. Although this is usable insofar as the obtained overall scene appearance is not unrealistic, the generated emission spectra are only metamers of the original illumination. This, in turn, can cause deviations from the expected appearance even if the rest of the scene corresponds to real world data. We propose a method capable of uplifting HDR environment maps based on spectral measurements of light sources similar to those present in the maps. To identify the illuminants, we employ an extensive set of emission measurements, and we combine the results with an existing reflectance uplifting method. In addition, we address the problem of environment map capture for the purposes of a spectral rendering pipeline, for which we propose a novel solution.
Controlling and Sampling Visibility Information on the Image Plane
(The Eurographics Association, 2017) Lessig, Christian; Matthias Zwicker and Pedro Sander
Anti-aliasing on the image plane is a classic problems in computer graphics. While mip-mapping provides an efficient means to pre-filter texture information, no comparable technique exists for visibility. We address visibility-induced aliasing by exploiting that the Fourier transform of a discontinuity decays slowly only in the normal direction. Pre-filtering is thus only necessary in this direction and, after a coordinate transformation, the corresponding one dimensional problem can be solved analytically or tabulated. The resulting pre-filtered signal can be reconstructed exactly from pointwise samples and we derive corresponding sampling theorems that are tailored to the pre-filtering as well as a set of irregular sampling locations. We demonstrate our methodology for the classical Shannon-Nyquist setting but also for shift-invariant spaces where exact reconstruction kernels with significantly faster decay than the sinc-function are available. Our experimental results demonstrate that our pre-filtering is highly effective and that going beyond the Shannon-Nyquist setting reduces aliasing error further.
Data-driven Pixel Filter Aware MIP Maps for SVBRDFs
(The Eurographics Association, 2023) Kemppinen, Pauli; Aittala, Miika; Lehtinen, Jaakko; Ritschel, Tobias; Weidlich, Andrea
We propose a data-driven approach for generating MIP map pyramids from SVBRDF parameter maps. We learn a latent material representation where linear image downsampling corresponds to linear prefiltering of surface reflectance. In contrast to prior work, we explicitly model the effect of the antialiasing pixel filter also at the finest resolution. This yields high-quality results even in images that are shaded only once per pixel with no further processing. The SVBRDF maps produced by our method can be used as drop-in replacements within existing rendering systems, and the data-driven nature of our framework makes it possible to change the shading model with little effort. As a proof of concept, we also demonstrate using a shared latent representation for two different shading models, allowing for automatic conversion
De-lighting a High-resolution Picture for Material Acquisition
(The Eurographics Association, 2019) Martin, Rosalie; Meyer, Arthur; Pesare, Davide; Boubekeur, Tamy and Sen, Pradeep
We propose a deep-learning based method for the removal of shades, projected shadows and highlights from a single picture of a quasi-planar surface captured in natural lighting conditions with any kind of camera device. To achieve this, we train an encoder-decoder to process physically based materials, rendered under various lighting conditions, to infer its spatially-varying albedo. Our network processes relatively small image tiles (512x512 pixels) and we propose a solution to handle larger image resolutions by solving a Poisson system across these tiles.
Deep Compositional Denoising on Frame Sequences
(The Eurographics Association, 2023) Zhang, Xianyao; Röthlin, Gerhard; Manzi, Marco; Gross, Markus; Papas, Marios; Ritschel, Tobias; Weidlich, Andrea
Path tracing is the prevalent rendering algorithm in the animated movies and visual effects industry, thanks to its simplicity and ability to render physically plausible lighting effects. However, we must simulate millions of light paths before producing one final image, and error manifests as noise during rendering. In fact, it can take tens or even hundreds of CPU hours on a modern computer to render a plausible frame in a recent animated movie. Movie production and the VFX industry rely on image-based denoising algorithms to ameliorate the rendering cost, which suppresses the noise due to rendering by reusing information in the neighborhood of the pixels both spatially and temporally.
Deep Hybrid Real and Synthetic Training for Intrinsic Decomposition
(The Eurographics Association, 2018) Bi, Sai; Kalantari, Nima Khademi; Ramamoorthi, Ravi; Jakob, Wenzel and Hachisuka, Toshiya
Intrinsic image decomposition is the process of separating the reflectance and shading layers of an image, which is a challenging and underdetermined problem. In this paper, we propose to systematically address this problem using a deep convolutional neural network (CNN). Although deep learning (DL) has been recently used to handle this application, the current DL methods train the network only on synthetic images as obtaining ground truth reflectance and shading for real images is difficult. Therefore, these methods fail to produce reasonable results on real images and often perform worse than the non-DL techniques. We overcome this limitation by proposing a novel hybrid approach to train our network on both synthetic and real images. Specifically, in addition to directly supervising the network using synthetic images, we train the network by enforcing it to produce the same reflectance for a pair of images of the same real-world scene with different illuminations. Furthermore, we improve the results by incorporating a bilateral solver layer into our system during both training and test stages. Experimental results show that our approach produces better results than the state-of-the-art DL and non-DL methods on various synthetic and real datasets both visually and numerically.
Deep Partitioned Shadow Volumes Using Stackless and Hybrid Traversals
(The Eurographics Association, 2016) Mora, Frédéric; Gerhards, Julien; Aveneau, Lilian; Ghazanfarpour, Djamchid; Elmar Eisemann and Eugene Fiume
Computing accurate hard shadows is a difficult problem in interactive rendering. Previous methods rely either on Shadow Maps or Shadow Volumes. Recently Partitioned Shadow Volumes (PSV) has been introduced. It revisits the old Shadow Volumes Binary Tree Space Partitioning algorithm, leading to a practicable and efficient technique. In this article, we analyze the PSV query algorithm and identify two main drawbacks: First, it uses a stack which is not GPU friendly; its size must be small enough to reduce the register pressure, but large enough to avoid stack overflow. Second, PSV struggles with configurations involving significant depth complexity, especially for lit points. We solve these problems by adding a depth information to the PSV data structure, and by designing a stackless query. In addition, we show how to combine the former PSV query with our stackless solution, leading to a hybrid technique taking advantage of both. This eliminates any risk of stack overflow, and our experiments demonstrate that these improvements accelerate the rendering time up to a factor of 3.
Diffuse-Specular Separation using Binary Spherical Gradient Illumination
(The Eurographics Association, 2018) Kampouris, Christos; Zafeiriou, Stefanos; Ghosh, Abhijeet; Jakob, Wenzel and Hachisuka, Toshiya
We introduce a novel method for view-independent diffuse-specular separation of albedo and photometric normals without requiring polarization using binary spherical gradient illumination. The key idea is that with binary gradient illumination, a dielectric surface oriented towards the dark hemisphere exhibits pure diffuse reflectance while a surface oriented towards the bright hemisphere exhibits both diffuse and specular reflectance. We exploit this observation to formulate diffuse-specular separation based on color-space analysis of a surface's response to binary spherical gradients and their complements. The method does not impose restrictions on viewpoints and requires fewer photographs for multiview acquisition than polarized spherical gradient illumination. We further demonstrate an efficient two-shot capture using spectral multiplexing of the illumination that enables diffuse-specular separation of albedo and heuristic separation of photometric normals.

Browse

Browsing Rendering - Experimental Ideas & Implementations by Title

Results Per Page

Sort Options