Ray Tracing in Nvidia GPUs

Method for accurately lighting virtual scenes in real-time rendering applications with complex, dynamic scenes. The method involves adaptive importance sampling of light sources in a scene using a cumulative distribution function (CDF). The CDF is generated based on importance estimates of lights in each voxel of the scene. This allows efficient sampling of lights for scene rendering by using the CDF to determine which lights are more important in a region. The CDF is updated as lights move or occlude to adapt to changing lighting conditions. Jittering voxel boundaries and light normal directions helps smooth out artifacts.

Source

Accelerating ray tracing and material shading on parallel processors like GPUs without adding transistors. The method involves compiling material graphs into optimized byte code instructions for execution on the parallel processing cores. The material graphs represent the surface properties and relationships of objects. The byte code is generated by parsing the graphs into expression trees and optimizing them. This allows efficient evaluation of the material graphs using the parallel processing cores without branching or sorting rays.

Source

Spectral rendering of images with volume media like mist, fog, dust, clouds, water, etc., that closely matches rendering in a multi-color scheme like RGB. The method involves identifying a light in a scene, determining an attenuation function characterizing interaction with the medium, fitting parameters based on minimizing color difference, and rendering using the attenuation function and fitted parameters. This allows realistic spectral rendering of lights transmitting through media that matches the appearance in a multi-color rendering.

Source

Render shadows in scenes with foliage like trees and bushes using a hybrid approach that improves performance and realism when ray-tracing shadows of alpha-tested geometry. The approach involves combining noisy visibility samples from raytraced opaque geometry with shadow maps that only contain alpha-tested geometry. This results in hybrid denoised shadows that show penumbras of high quality and hard ray-traced shadows at contact points with a low performance impact. The alpha-only shadow map only contains depth values from geometry that passes the alpha-test.

Source

Optimizing motion vector estimation for translucent objects in images to improve rendering quality and reduce noise. The optimization involves accurately finding motion vectors for pixels in a current frame by searching for matching world positions in previous frames, even when objects are transparent. It uses numerical optimization techniques like Newton's method to minimize the angle between vectors instead of finding closest pixels. This improves performance over naive methods like random walks. The optimized motion vectors are then used for tasks like denoising and temporal anti-aliasing.

Source

Generating realistic water caustics in real-time gaming and animation applications that appear and move like actual physical caustics, while avoiding the need for expensive denoising procedures that can be computationally expensive or prohibitive for real-time use. The method involves using a low resolution caustics map with each point treated as a vector. These vectors are rendered as a mesh with connectors between adjacent points. When a hit point is detected, the position of the hit point replaces the original vertex position. This projected mesh provides sharp caustic patterns using fewer points compared to traditional denoising. The density of triangles is determined based on area to accurately represent brightness.

Source

Rendering signed distance functions (SDFs) in computer graphics more efficiently by improving techniques for tracing rays, computing surface normals, and finding ray intersections with SDF surfaces. The techniques involve: 1. Factorizing SDF coefficients into parameters to compute ray intersections with SDF surfaces using a cubic function. 2. Computing surface normals by interpolating neighboring voxel normals. 3. Finding ray intersections with SDF surfaces using turning points of the cubic function to determine shadowed regions. These techniques improve ray tracing and surface computation efficiency compared to duplicative calculations and separate ray tracing for shadows.

Source

Interleaving textures to improve ray tracing performance for scenes with incoherent light rays. Instead of separately allocating and accessing cache lines for each texture associated with an object intersected by a ray, the textures are combined into a single interleaved texture. This reduces the number of cache lines needed and less texture data is read compared to separately accessing each texture. The interleaved texture contains blocks from multiple textures alternating in order. It can be accessed using multiple texture headers with stride distances between blocks from the same texture. This allows fetching from the interleaved texture instead of separate textures for rays with incoherent intersections.

Source

Hardware acceleration of ray tracing for realistic graphics rendering that provides selective execution of programmable ray operations. The technique allows efficient and flexible ray intersection tests by enabling custom ray operations like selecting the level of detail (LOD) of an object. This improves ray tracing performance by allowing optimized ray operations for acceleration structures like BVHs that have mixed LOD models. It avoids the rigidity of requiring all nodes to support the same programmable ray operations. Instead, the ray operations can be applied selectively to nodes based on their data.

Source

Detecting penumbra regions for shadow denoising using wave intrinsic functions on parallel processors to avoid expensive post-processing for penumbra detection. The approach leverages threads of schedulable units (e.g., warps) used for visibility sampling during ray tracing to identify penumbra regions. Each thread computes a value indicating if its pixel region is in a penumbra using intrinsic functions. This avoids post-processing and allows selective denoising of penumbra regions vs fully lit/shaded regions. It also determines filter parameters based on sampling statistics to adapt denoising for regions.

Source

Hardware acceleration for real-time ray tracing to enable interactive ray tracing for applications like virtual reality and augmented reality. The acceleration is provided by a dedicated co-processor called a traversal coprocessor that sits in the graphics pipeline alongside the streaming processors. The coprocessor efficiently traverses an acceleration data structure like a bounding volume hierarchy to quickly determine ray intersections with scene objects. It handles tasks like ray-bounding volume intersection, ray-primitive intersection, and ray transforms. The coprocessor's dedicated hardware provides faster ray tracing compared to software ray tracing.

Source

Hardware-accelerated ray tracing that provides flexible and efficient ray intersection tests for real-time graphics. The technique allows combining ray operations like instance masking and geometric level of detail testing in parallel processing units. This improves ray tracing efficiency by enabling more complex ray tracing effects while simultaneously improving ray tracing efficiency. It allows dynamically choosing ray traversal based on multiple selection criteria per ray. The ray operations are configured to occur before and after node stack push/pop in the traversal coprocessor. This enables combining ray operations like instance masking and geometric level of detail testing in parallel processing units.

Source

Efficiently generating high quality images using techniques that allow efficient sampling of large emissive textures for lighting scenes. The method involves determining cumulative distribution functions for textures, constructing geometric representations from them, and tracing rays against the geometry to sample the textures efficiently. This hardware-accelerated sampling reduces noise compared to naive random sampling. The approach allows realistic lighting of scenes using large textures without incurring excessive computation cost.

Source

Rendering techniques for media like surfaces with microscopic irregularities and homogeneous slabs using position-free path integrals. It involves sampling directions of light passing through the medium, computing parameters for position distributions based on the directions, computing brightness at the exit direction using the distribution parameters, and rendering images based on the brightness. This approximates path integrals through the medium without the noise of Monte Carlo simulations.

Source

Techniques for sharing irradiance between ray interactions spatially and temporally to improve ray tracing performance for rendering virtual scenes. The method involves using irradiance caches to aggregate and interpolate irradiance samples instead of sampling from every location. Irradiance caches are associated with locations and updated by casting fewer rays based on ranking factors. Irradiance from other caches can be blended to compute irradiance at a location. This reduces the number of samples needed for lighting calculations compared to full ray tracing.

Source

Reducing temporal lag in real-time ray tracing and other dynamic scenes without compromising image quality. It uses a fast history buffer with higher blend weight and clamping to determine a clamping window for the full history buffer. This allows controlling the balance between noise reduction and temporal lag. The fast history buffer is used to clamp the full history value before reprojection to the current frame. The full history is still maintained for accurate smoothing in static scenes without clamping. This fast-full history clamping technique reduces temporal lag in dynamic scenes without adding computational complexity compared to conventional methods.

Source

Efficiently calculating light in 3D scenes with complex environments like indoor scenes using importance sampling. The technique involves generating a mask that identifies regions in the scene where lighting can affect pixels. This mask is continuously updated per frame. It's used to extract an importance sampling function for tracing rays to find light paths from the environment. This improves ray guidance to efficiently find important light transport paths. The mask considers occlusions and identifies areas where rays can hit lights in the environment map.

Source

Ray cone tracing for more efficient and realistic rendering of images using ray cone tracing. The technique avoids a separate G-buffer pass for determining surface curvature during ray cone tracing. Instead, when a ray cone hits a surface, it computes the curvature using differential barycentric coordinates at the hit point. This curvature is then used to determine the width of the ray cone at the next intersection. This width is used to sample mipmaps during texture filtering operations.

Source

Real-time neural network radiance caching for path tracing in 3D rendering applications like computer graphics. The technique uses a neural network to estimate scattered radiance components of global illumination. Short paths are traced from a camera through pixels in an image. The neural network caches radiance predictions at vertices. Longer paths are extended from the rendering paths to train the network. This allows real-time caching without pre-training. Fully-fused neural network execution improves performance by partitioning inputs and loading weights into registers once for parallel processing.

Source

Parallel ray tracing technique for accelerated graphics processing using a physically-based rendering engine. The technique involves assigning each ray to a separate processing core and executing the ray's material shading instructions in parallel. This allows the ray's intersection point and material properties to be evaluated concurrently. The cores can also perform specialized fixed-function pipelines for tasks like importance sampling. The cores return secondary ray directions when intersecting surfaces. This enables efficient ray tracing using a parallel architecture by leveraging the cores' SIMD/SIMT execution models.

Source

Rendering scenes faster while minimizing visual quality degradation compared to traditional methods like ray tracing or rasterization alone. The technique involves combining ray tracing and rasterization algorithms in a hybrid rendering approach. It uses ray tracing to determine occluder objects in the scene, then leverages rasterization to render visible objects. This allows culling of occluded areas using the ray tracing acceleration structure. The hybrid approach improves rendering speed by utilizing ray tracing for occlusion culling and rasterization for visible object rendering.

Source

Ray tracing rendering method that combines rasterization and ray tracing for efficient real-time rendering using hardware acceleration. It leverages the rasterization pipeline to quickly project elements onto a plane and obtain depth values. These depths are used to accelerate ray tracing calculations by converting them to distances to the nearest collision points in 3D. This allows utilizing ray tracing for effects like shadows and reflections while leveraging rasterization for initial projection.

Source

Hardware acceleration of computer graphics ray tracing to enable interactive real-time ray tracing for high-quality photorealistic rendering. A specialized coprocessor, called a traversal coprocessor, is added to the graphics pipeline. This coprocessor accelerates the ray tracing operations like ray-bounding volume intersection tests, ray-primitive intersection tests, and ray transforms. It focuses on the co-processor's role in ray tracing rather than ray tracing algorithms. The coprocessor reduces ray tracing computation time by preprocessing scene geometry into an acceleration data structure like a bounding volume hierarchy. The coprocessor traverses this structure to efficiently find ray intersections with objects.

Source

Ray cone tracing with anisotropic texture filtering for realistic image rendering with reduced aliasing and blur, especially at oblique angles. It approximates ray cone intersections as cylinders to compute ellipse axes and gradients for anisotropic texture filtering. This improves ray cone tracing by extending anisotropic filtering to ray cones instead of just rays.

Source

Hardware acceleration for ray tracing to improve real-time ray tracing performance by reducing false hits and repeated tests when geometry doesn't tightly fit into single, axis-aligned bounding boxes. The technique uses a multi-box bounding volume hierarchy (BVH) with wide compressed treelet representation. This allows more accurate fitting of complex geometry without excess traversal and false positives compared to axis-aligned boxes. The multi-box BVH can be encoded using wide complet formats and provides better single hit behavior for geometry that doesn't fit well into single axis-aligned boxes.

Source

Optimizing the rendering of computer graphics scenes with many lights to improve efficiency by preprocessing the light samples before rendering. The technique involves generating a subset of light samples in advance using reservoir-based spatiotemporal importance sampling (ReSTIR). The preprocessed light samples are stored in a memory structure that avoids cache thrashing and incoherent access. This reduces memory latency and improves cache utilization compared to accessing widely distributed lists of lights during rendering.

Source

Processing ray traced caustics effects in real-time using feedback of photon information between frames. The method involves determining photon patterns in a scene based on interactions with objects that scatter light, and using that information to improve subsequent frame rendering. This improves accuracy compared to ignoring or roughly handling caustics. The photon scattering is adaptive and anisotropic, considering perturbations to light ray parameters.

Source

Hardware acceleration of ray tracing in real-time graphics processing by using a dedicated coprocessor to efficiently traverse acceleration data structures like bounding volume hierarchies. The coprocessor accelerates ray intersection tests against geometry by omitting reporting of provably unused primitives. This reduces ray tracing complexity from O(n) to O(log n) for n objects. The coprocessor integrates into graphics pipelines and complements rasterization for enhanced image quality with photorealistic reflections and shadows.

Source

Scalable parallel construction of bounding volume hierarchies for efficiently generating BVHs on parallel computers. The technique alternates between node and object updates using parallel processing. Nodes are split or converted to leaf nodes. Objects are reassigned to child nodes based on partitions. This allows adaptive BVH construction with variable node/object count, scaling to large datasets. Parallel execution of node and object updates allows balanced resource usage across hierarchy levels.

Source

Real-time neural network radiance caching for path tracing in computer graphics to enable fast and accurate rendering of dynamic scenes. The method involves using a neural network radiance cache to estimate scattered light at intermediate points of light transport paths in a scene. The cache is trained by extending some of the paths to terminal positions and updating the cache weights based on the true radiance values there. This allows the cache to predict radiance at any point efficiently. By reusing rendered paths and sparse sampling, the training is done in real-time alongside rendering. The neural network architecture is fully-connected and fused to minimize memory traffic.

Source

Adaptive ray tracing denoising technique for rendering shadows that improves denoising quality and reduces noise while conserving computing resources. The technique adaptively guides the spatiotemporal filter used for shadow denoising based on variance in accumulated ray-traced samples. If the variance is low, indicating consistent lighting conditions, it skips filtering altogether. This avoids over-filtering when lighting is stable. The variance is computed using accumulated ray-traced visibility samples over multiple frames. The mean and count are also checked to skip filtering if the pixel is fully lit or occluded.

Source

More efficient rendering of signed distance functions (SDFs) in 3D computer graphics. The techniques involve improved methods for tracing rays through an SDF grid, finding intersections, computing surface normals, and shading pixels. Key optimizations include: - Factorized coefficients for cubic ray-SDF intersections instead of duplicated computations. - Interpolating neighboring normals to smoothly blend across grid cells instead of discontinuous normals. - Shadow ray optimization using cubic function turning points. - Factorized cubic function coefficients for faster intersection calculation.

Source

Reducing variance in estimating transmittance through participating media for light transport simulation using a combination of function samples and power series expansion. The approach involves computing the transmittance as a weighted sum of terms in a power series expansion of the media density function. The terms are evaluated using combinations of function samples rather than individual samples. This reduces variance compared to traditional Monte Carlo estimation. The weights for the combined samples are chosen to provide an unbiased estimate. The pivot value for the power series is also computed using function samples to improve convergence. By combining samples and using an optimal pivot, the estimated transmittance has lower variance and faster convergence compared to traditional Monte Carlo methods.

Source

Generating spatio-temporal blue noise masks for real-time image rendering and enhancement that maintain spatial 2D blue noise properties while providing blue noise over time at every pixel. This involves extending the void and cluster algorithm used to generate blue noise masks to handle multiple dimensions, including time. The spatio-temporal blue noise masks improve convergence rates when taking multiple samples per frame or integrating frames over time, without falling into white noise over time. The technique enables blue noise properties in each spatial slice (2D) while providing better than white noise sequences along the time axis (1D).

Source

Hybrid differentiable rendering technique that combines aspects of rasterization and ray-tracing to efficiently render 3D models while supporting advanced lighting effects. The technique involves using machine learning to disentangle geometry, lighting, and material information from input images, and then using rasterization to determine pixel coverage and ray-tracing to compute radiance for covered pixels. This allows efficient rendering of complex scenes with realistic lighting effects.

Source

Accelerating ray tracing and material shading using a physically based rendering engine on parallel processors like GPUs. The rendering engine has a shared set of fixed-function pipelines that execute specific material shading functions. The pipelines are assigned to processing cores instead of entire shader programs. This allows parallel execution of ray tracing and material shading by assigning rays to cores and executing their specific material instructions in the pipelines. It avoids sorting and grouping rays for parallelism as each ray's unique material properties can be processed in parallel.

Source

Accelerating ray tracing and material shading on parallel processors like GPUs by leveraging their fixed-function pipelines instead of relying solely on SIMD/SIMT execution. The technique involves compiling material graphs into fixed-function pipelines with shared execution. Ray processing cores feed inputs to the pipelines, which execute specialized functions on the rays before sending back results. This allows efficient ray tracing without sorting or grouping. The pipelines execute in parallel with cores, improving throughput. The technique also allows compiling material graphs into assembly language for the pipelines.

Source

Accelerated ray tracing and material shading using a physically based rendering engine with fixed-function pipelines and decompressed texture extraction to improve parallel processing efficiency on GPUs and other parallel processors. The rendering engine has shared fixed-function pipelines for complex material graph functions. Rays are sent to a tree traversal unit to find intersecting surfaces, then to the rendering engine for material shading. Pipelines extract texels from compressed blocks without decompressing the whole block. This allows decompression and extraction parallelism. The fixed-function pipelines and decompressed texture extraction help leverage the parallel processing capabilities of GPUs and other parallel processors for ray tracing and material shading.

Source

Distributed streaming of dynamic light probes for realistic real-time rendering on low-end devices without ray tracing capabilities. The technique involves compressing and streaming light field data from a server to clients. The compression reduces bandwidth requirements. The light field data is encoded using techniques like lossy video compression. The server computes the light field for a scene and encodes it into compressed light probes. The clients receive and decode the probes to accurately render the scene using global illumination.

Source

Parallelizing ray tracing on GPUs without adding more transistors to enable faster rendering. The method involves assigning each ray to a different processing core and executing its material shading instructions in parallel. This leverages the parallel processing capability of GPUs without requiring ray sorting or SIMD/SIMT techniques. It allows efficient ray tracing on GPUs by exploiting their parallel architecture.

Source

Adding depth of field effects to omnidirectional cameras for virtual reality applications. It involves modeling the omnidirectional camera as a finite aperture lens with a focal point aligned with the omnidirectional projection. This allows simulating depth of field by projecting rays through the lens and focal point to generate the final image. It extends the pinhole omnidirectional camera model to support finite aperture lenses with depth of field.

Source

Adaptive sampling and denoising of rendered sequences using neural networks for improved image quality and temporal stability in real-time rendering applications with limited samples. The technique involves a neural network that learns to predict optimal sampling patterns for each frame based on temporal feedback from previous frames. It also learns to denoise the low-sample frames to boost effective sample count. This neural network structure, called a warped external recurrent neural network, is trained end-to-end to reconstruct images with reduced artifacts and missing pixels.

Source

Method for rendering images with realistic light simulation using ray tracing. The method involves storing visibility parameters for light samples in a first region, then using those parameters to determine lighting for adjacent regions. This allows more efficient lighting calculations by reusing visibility data. The parameters indicate how much light should be applied from a specific sample. By storing these per-sample values for a region, it allows faster determination of lighting for neighboring regions. This is done by comparing age and distance thresholds to determine when to use the stored parameters versus tracing new rays.

Source

Efficiently handling large numbers of lights in computer graphics rendering by decoupling shading and reuse operations in techniques like ReSTIR. The decoupling improves utilization of ray tracing by avoiding duplicate rays and leveraging visibility information already available. The decoupling allows better distribution of work and reduces bias accumulation compared to directly evaluating all lights for each pixel.

Source

A method of rendering 3D scenes that combines raytracing and rasterization to improve rendering efficiency while maintaining visual quality. The method involves using raytracing to generate acceleration structures that rasterization can leverage. It also involves selectively using raytracing or rasterization based on scene data to balance performance and quality. The idea is to use raytracing for occlusion culling and acceleration, then switch to rasterization for final rendering using the raytracing data as input. This hybrid approach reduces rendering time compared to pure raytracing or rasterization while preserving visual fidelity.

Source

Hardware acceleration of ray tracing for real-time interactive graphics. A specialized ray tracing coprocessor is added to a graphics processing pipeline. The coprocessor efficiently accelerates ray tracing tasks like intersection tests and transforms. It complements the main graphics processor for ray tracing beyond rasterization. The coprocessor reduces ray tracing complexity from O(n) to O(log n) with preprocessing. It enables true real-time ray tracing with physics-based effects like shadows and reflections.

Source

Hardware-based ray-primitive intersection testing for ray tracing that efficiently determines ray-primitive intersections for primitives sharing features like vertices. The technique involves expanding ray tracing hardware to handle intersection testing of primitives sharing at least one feature like vertices, using parallel processing for common features to double the ray intersection testing rate without doubling the hardware. This leverages calculations for shared vertices across multiple primitives using minimal extra hardware.

Source

A parallel rendering technique that improves performance for ray tracing and image filtering by interleaving pixels into partial image fields and processing them in parallel. The full frame is split into multiple partial fields, each with different pixels. These fields are ray traced separately using different rays. The partial fields are then combined to render the full frame. This parallel processing allows more efficient ray tracing and image filtering compared to sequential processing of the full frame.

Source

Parallelizing ray tracing and image filtering for faster rendering by splitting the full frame into smaller partial image fields and processing them independently. This allows parallelization since each partial field can be ray traced on a separate processor or GPU. After ray tracing, the partial fields are combined to render the full frame using image filtering. By interleaving pixels into sets, each partial field has even sampling of the full frame and all the sets together make a complete frame.

Source

Ray cone tracing techniques for more efficient texture filtering in 3D rendering. Instead of tracing rays and separately computing mipmap sampling levels for each texture, the techniques determine curvature at ray intersections using barycentric coordinates. This curvature is used to dynamically adjust ray widths as they hit further surfaces. Texture filtering is then performed with parameter values based on the adjusted ray width. This avoids separate G-buffer passes and reduces computational cost compared to traditional ray tracing and mipmap sampling.

Source