High-Performance Graphics 2012

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/323

Browse

Now showing 1 - 14 of 14

Reducing Aliasing Artifacts through Resampling
(The Eurographics Association, 2012) Reshetov, Alexander; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
Post-processing antialiasing methods are well suited for deferred shading because they decouple antialiasing from the rest of graphics pipeline. In morphological methods, the final image is filtered with a data-dependent filter. The filter coefficients are computed by analyzing the non-local neighborhood of each pixel. Though very simple and efficient, such methods have intrinsic quality limitations due to spatial undersampling and temporal aliasing. We explore an alternative formulation in which filter coefficients are computed locally for each pixel by supersampling geometry, while shading is still done only once per pixel. During pre-processing, each geometric subsample is converted to a single bit indicating whether the subsample is different from the central one. The ensuing binary mask is then used in the post-processing step to retrieve filter coefficients, which were precomputed for all possible masks. For a typical 8 subsamples, it results in a sub-millisecond performance, while improving the image quality by about 10 dB. To compare subsamples, we use a novel symmetric angular measure, which has a simple geometric interpretation. We propose to use this measure in a variety of applications that assess the difference between geometric samples (rendering, mesh simplification, geometry encoding, adaptive tessellation).
Representing Appearance and Pre-filtering Subpixel Data in Sparse Voxel Octrees
(The Eurographics Association, 2012) Heitz, Eric; Neyret, Fabrice; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
Sparse Voxel Octrees (SVOs) represent efficiently complex geometry on current GPUs. Despite the fact that LoDs come naturally with octrees, interpolating and filtering SVOs are still issues in current approaches. In this paper, we propose a representation for the appearance of a detailed surface with associated attributes stored within a voxel octree. We store macro- and micro-descriptors of the surface shape and associated attributes in each voxel. We represent the surface macroscopically with a signed distance field and we encode subvoxel microdetails with Gaussian descriptors of the surface and attributes within the voxel. Our voxels form a continuous field interpolated through space and scales, through which we cast conic rays. Within the ray marching steps, we compute the occlusion distribution produced by the macro-surface inside a pixel footprint, we use the microdescriptors to reconstruct light- and view-dependent shading, and we combine fragments in an A-buffer way. Our representation efficiently accounts for various subpixel effects. It can be continuously interpolated and filtered, it is scalable, and it allows for efficient depth-of-field.We illustrate the quality of these various effects by displaying surfaces at different scales, and we show that the timings per pixel are scale-independent.
kANN on the GPU with Shifted Sorting
(The Eurographics Association, 2012) Li, Shengren; Simons, Lance; Pakaravoor, Jagadeesh Bhaskar; Abbasinejad, Fatemeh; Owens, John D.; Amenta, Nina; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
We describe the implementation of a simple method for finding k approximate nearest neighbors (ANNs) on the GPU. While the performance of most ANN algorithms depends heavily on the distributions of the data and query points, our approach has a very regular data access pattern. It performs as well as state of the art methods on easy distributions with small values of k, and much more quickly on more difficult problem instances. Irrespective of the distribution and also roughly of the size of the set of input data points, we can find 50 ANNs for 1M queries at a rate of about 1200 queries/ms.
High-Quality Parallel Depth-of-Field Using Line Samples
(The Eurographics Association, 2012) Tzeng, Stanley; Patney, Anjul; Davidson, Andrew; Ebeida, Mohamed S.; Mitchell, Scott A.; Owens, John D.; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
We present a parallel method for rendering high-quality depth-of-field effects using continuous-domain line samples, and demonstrate its high performance on commodity GPUs. Our method runs at interactive rates and has very low noise. Our exploration of the problem carefully considers implementation alternatives, and transforms an originally unbounded storage requirement to a small fixed requirement using heuristics to maintain quality. We also propose a novel blur-dependent level-of-detail scheme that helps accelerate rendering without undesirable artifacts. Our method consistently runs 4 to 5x faster than an equivalent point sampler with better image quality. Our method draws parallels to related work in rendering multi-fragment effects.
Algorithm and VLSI Architecture for Real-Time 1080p60 Video Retargeting
(The Eurographics Association, 2012) Greisen, Pierre; Lang, Manuel; Heinzle, Simon; Smolic, Aljosa; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
Aspect ratio retargeting for streaming video has actively been researched in the past years. While the mobile market with its huge diversity of screen formats is one of the most promising application areas, no existing algorithm is efficient enough to be embedded in such devices. In this work, we devise an efficient video retargeting algorithm by following an algorithm-architecture co-design approach and we present the first FPGA implementation that is able to retarget full HD 1080p video at up to 60 frames per second. We furthermore show that our algorithm can be implemented on embedded processors at interactive framerates. Our hardware architecture only requires a modest amount of hardware resources, and is portable to a dedicated ASIC for the use in consumer electronic devices such as displays or mobile phones.
Adaptive Image Space Shading for Motion and Defocus Blur
(The Eurographics Association, 2012) Vaidyanathan, Karthik; Toth, Robert; Salvi, Marco; Boulos, Solomon; Lefohn, Aaron; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
We present a novel anisotropic sampling algorithm for image space shading which builds upon recent advancements in decoupled sampling for stochastic rasterization pipelines. First, we analyze the frequency content of a pixel in the presence of motion and defocus blur.We use this analysis to derive bounds for the spectrum of a surface defined over a two-dimensional and motion-aligned shading space. Second, we present a simple algorithm that uses the new frequency bounds to reduce the number of shaded quads and the size of decoupling cache respectively by 2X and 16X, while largely preserving image detail and minimizing additional aliasing.
Scalable Ambient Obscurance
(The Eurographics Association, 2012) McGuire, Morgan; Mara, Michael; Luebke, David; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
This paper presents a set of architecture-aware performance and integration improvements for a recent screenspace ambient obscurance algorithm. These improvements collectively produce a 7x performance increase at 2560 x1600, generalize the algorithm to both forward and deferred renderers, and eliminate the radius- and scene-dependence of the previous algorithm to provide a hard real-time guarantee of fixed execution time. The optimizations build on three strategies: pre-filter the depth buffer to maximize memory hierarchy efficiency; reduce total bandwidth by carefully reconstructing positions and normals at high precision from a depth buffer; and exploit low-level intra- and inter-thread techniques for parallel, floating-point architectures.
Maximizing Parallelism in the Construction of BVHs, Octrees, and k-d Trees
(The Eurographics Association, 2012) Karras, Tero; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
A number of methods for constructing bounding volume hierarchies and point-based octrees on the GPU are based on the idea of ordering primitives along a space-filling curve. A major shortcoming with these methods is that they construct levels of the tree sequentially, which limits the amount of parallelism that they can achieve. We present a novel approach that improves scalability by constructing the entire tree in parallel. Our main contribution is an in-place algorithm for constructing binary radix trees, which we use as a building block for other types of trees.
Design and Novel Uses of Higher-Dimensional Rasterization
(The Eurographics Association, 2012) Nilsson, Jim; Clarberg, Petrik; Johnsson, Björn; Munkberg, Jacob; Hasselgren, Jon; Toth, Robert; Salvi, Marco; Akenine-Möller, Tomas; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
This paper assumes the availability of a very fast higher-dimensional rasterizer in future graphics processors. Working in up to five dimensions, i.e., adding time and lens parameters, it is well-known that this can be used to render scenes with both motion blur and depth of field. Our hypothesis is that such a rasterizer can also be used as a flexible tool for other, less conventional, usage areas, similar to how the two-dimensional rasterizer in contemporary graphics processors has been used for widely different purposes other than the original intent. We show six such examples, namely, continuous collision detection, caustics rendering, higher-dimensional sampling, glossy reflections and refractions, motion blurred soft shadows, and finally multi-view rendering. The insights gained from these examples are used to put together a coherent model for what a future graphics pipeline that supports these and other use cases should look like. Our work intends to provide inspiration and motivation for hardware and API design, as well as continued research in higher-dimensional rasterization and its uses.
Parallel Patch based Texture Synthesis
(The Eurographics Association, 2012) Lasram, Anass; Lefebvre, Sylvain; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
Fast parallel algorithms exist for pixel-based texture synthesizers. Unfortunately, these synthesizers often fail to preserve structures from the exemplar without the user specifying additional feature information. On the contrary, patch-based synthesizers are better at capturing and preserving structural patterns. However, they require relatively slow algorithms to layout the patches and stitch them together. We present a parallel patch-based texture synthesis technique that achieves high degree of parallelism. Our synthesizer starts from a low-quality result and adds several patches in parallel to improve it. It selects patches that blend in a seamless way with the existing result, and that hide existing visual artifacts. This is made possible through two main algorithmic contributions: An algorithm to quickly find a good cut around a patch, and a deformation algorithm to further align features crossing the patch boundary. We show that even with a uniform parallel random sampling of the patches, our improved patch stitching achieves high quality synthesis results. We discuss several synthesis strategies, such as using patches of decreasing size or using various amounts of deformation during the optimization. We propose a complete implementation tuned to take advantage of massive GPU parallelism.
Clustered Deferred and Forward Shading
(The Eurographics Association, 2012) Olsson, Ola; Billeter, Markus; Assarsson, Ulf; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
This paper presents and investigates Clustered Shading for deferred and forward rendering. In Clustered Shading, view samples with similar properties (e.g. 3D-position and/or normal) are grouped into clusters. This is comparable to tiled shading, where view samples are grouped into tiles based on 2D-position only. We show that Clustered Shading creates a better mapping of light sources to view samples than tiled shading, resulting in a significant reduction of lighting computations during shading. Additionally, Clustered Shading enables using normal information to perform per-cluster back-face culling of lights, again reducing the number of lighting computations. We also show that Clustered Shading not only outperforms tiled shading in many scenes, but also exhibits better worst case behaviour under tricky conditions (e.g. when looking at high-frequency geometry with large discontinuities in depth). Additionally, Clustered Shading enables real-time scenes with two to three orders of magnitudes more lights than previously feasible (up to around one million light sources).
Power Efficiency for Software Algorithms Running on Graphics Processors
(The Eurographics Association, 2012) Johnsson, Björn; Ganestam, Per; Doggett, Michael; Akenine-Möller, Tomas; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
Power efficiency has become the most important consideration for many modern computing devices. In this paper, we examine power efficiency of a range of graphics algorithms on different GPUs. To measure power consumption, we have built a power measuring device that samples currents at a high frequency. Comparing power efficiency of different graphics algorithms is done by measuring power and performance of three different primary rendering algorithms and three different shadow algorithms. We measure these algorithms' power signatures on a mobile phone, on an integrated CPU and graphics processor, and on high-end discrete GPUs, and then compare power efficiency across both algorithms and GPUs. Our results show that power efficiency is not always proportional to rendering performance and that, for some algorithms, power efficiency varies across different platforms. We also show that for some algorithms, energy efficiency is similar on all platforms.
SRDH: Specializing BVH Construction and Traversal Order Using Representative Shadow Ray Sets
(The Eurographics Association, 2012) Feltman, Nicolas; Lee, Minjae; Fatahalian, Kayvon; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
We derive the Shadow Ray Distribution Heuristic (SRDH), an accurate cost estimator for shadow ray traversal through a bounding volume hierarchy (BVH). The SRDH leverages up-front knowledge of the distribution and intersection results of previously traced shadow rays to construct a shadow-ray-specialized BVH and choose an associated traversal order policy which together promote early termination by quickly finding occlusions. In scenes containing large amounts of occlusion, SRDH reduces the number of BVH node traversal steps needed for shadow computations between 22% and 56% compared to average-case traversal through SAH-constructed trees. Evaluating the SRDH using a sparse shadow ray set recorded from a 16 x16 pixel rendering of the scene consistently produces BVHs whose traversal cost is within 6% of those built when all shadow rays are available to the metric at the time of construction. The benefits of the SRDH come at the cost of storing an additional BVH in memory and a 2.4x increase (on average) in BVH construction time.
Adaptive Scalable Texture Compression
(The Eurographics Association, 2012) Nystad, Jorn; Lassen, Anders; Pomianowski, Andy; Ellis, Sean; Olson, Tom; Carsten Dachsbacher and Jacob Munkberg and Jacopo Pantaleoni
We describe a fixed-rate, lossy texture compression system that is designed to offer an unusual degree of flexibility and to support a very wide range of use cases, while providing better image quality than most formats in common use today. The system supports both 2D and 3D textures, at both standard and high dynamic range, at bit rates ranging from eight bits per pixel down to less than one bit per pixel in very fine steps. At any bit rate, texels can have from one to four color components. The system's flexibility results from a number of novel features. Color spaces and weights are represented using an encoding scheme that allows flexible allocation of bits between different types of information. The system uses bilinear interpolation to derive color space coordinates for a texel from sparse samples, and uses a procedural partition function to map texels to color spaces.

Browse

Browsing High-Performance Graphics 2012 by Issue Date

Results Per Page

Sort Options