Accurate sound rendering can add significant realism to complement visual display in interactive applications, as well as facilitate acoustic predictions for many engineering applications, like accurate acoustic analysis for architectural design. Numerical simulation can provide this realism most naturally by modeling the underlying physics of wave propagation. However, wave simulation has traditionally posed a tough computational challenge. In this paper, we present a technique which relies on an adaptive rectangular decomposition of 3D scenes to enable efficient and accurate simulation of sound propagation in complex virtual environments. It exploits the known analytical solution of the Wave Equation in rectangular domains, and utilizes an efficient implementation of the Discrete Cosine Transform on Graphics Processors (GPU) to achieve at least a 100-fold performance gain compared to a standard Finite-Difference Time-Domain (FDTD) implementation with comparable accuracy, while also being 10-fold more memory efficient. Consequently, we are able to perform accurate numerical acoustic simulation on large, complex scenes in the kilohertz range. To the best of our knowledge, it was not previously possible to perform such simulations on a desktop computer. Our work thus enables acoustic analysis on large scenes and auditory display for complex virtual environments on commodity hardware.

We present a novel approach for wave-based sound propagation suitable for large, open spaces spanning hundreds of meters, with a small memory footprint. The scene is decomposed into disjoint rigid objects. The free-field acoustic behavior of each object is captured by a compact per-object transfer function relating the amplitudes of a set of incoming equivalent sources to outgoing equivalent sources. Pairwise acoustic interactions between objects are computed analytically to yield compact inter-object transfer functions. The global sound field accounting for all orders of interaction is computed using these transfer functions. The runtime system uses fast summation over the outgoing equivalent source amplitudes for all objects to auralize the sound field for a moving listener in real time. We demonstrate realistic acoustic effects such as diffraction, low-passed sound behind obstructions, focusing, scattering, high-order reflections, and echoes on a variety of scenes.

An efficient algorithm for time-domain solution of the acoustic wave equation for the purpose of room acoustics is presented. It is based on adaptive rectangular decomposition of the scene and uses analytical solutions within the partitions that rely on spatially invariant speed of sound. This technique is suitable for auralizations and sound field visualizations, even on coarse meshes approaching the Nyquist limit. It is demonstrated that by carefully mapping all components of the algorithm to match the parallel processing capabilities of graphics processors (GPUs), significant improvement in performance is gained compared to the corresponding CPU-based solver, while maintaining the numerical accuracy. Substantial performance gain over a high-order finite-difference time-domain method is observed. Using this technique, a 1 second long simulation can be performed on scenes of air volume 7500 m 3 till 1650 Hz within 18 minutes compared to the corresponding CPU-based solver that takes around 5 hours and a high-order finite-difference time-domain solver that could take up to three weeks on a desktop computer. To the best of the authors' knowledge, this is the fastest time-domain solver for modeling the room acoustics of large, complex-shaped 3D scenes that generates accurate results for both auralization and visualization.

We present an interactive approach for generating realistic physically-based sounds from rigid-body dynamic simulations. We use spring-mass systems to model each object's local deformation and vibration, which we demonstrate to be an adequate approximation for capturing physical effects such as magnitude of impact forces, location of impact, and rolling sounds. No assumption is made about the mesh connectivity or topology. Surface meshes used for rigid-body dynamic simulation are utilized for sound simulation without any modifications. We use results in auditory perception and a novel priority-based quality scaling scheme to enable the system to meet variable, stringent time constraints in a real-time application, while ensuring minimal reduction in the perceived sound quality. With this approach, we have observed up to an order of magnitude speed-up compared to an implementation without the acceleration. As a result, we are able to simulate moderately complex simulations with upto hundreds of sounding objects at over 100 frames per second (FPS), making this technique well suited for interactive applications like games and virtual environments. Furthermore, we utilize OpenAL and EAX TM on Creative Sound Blaster Audigy 2 TM cards for fast hardware-accelerated propagation modeling of the synthesized sound.

We present algorithms for fast quantile and frequency estimation in large data streams using graphics processors (GPUs). We exploit the high computation power and memory bandwidth of graphics processors and present a new sorting algorithm that performs rasterization operations on the GPUs. We use sorting as the main computational component for histogram approximation and construction of -approximate quantile and frequency summaries. Our algorithms for numerical statistics computation on data streams are deterministic, applicable to fixed or variable-sized sliding windows and use a limited memory footprint. We use GPU as a co-processor and minimize the data transmission between the CPU and GPU by taking into account the low bus bandwidth. We implemented our algorithms on a PC with a NVIDIA GeForce FX 6800 Ultra GPU and a 3.4 GHz Pentium IV CPU and applied them to large data streams consisting of more than 100 million values. We also compared the performance of our GPU-based algorithms with optimized implementations of prior CPU-based algorithms. Overall, our results demonstrate that the graphics processors available on a commodity computer system are efficient stream-processor and useful co-processors for mining data streams.