Authors:Savioja, Lauri; Välimäki, Vesa; Smith, Julius O.
Affiliation:NVIDIA Research and Department of Media Technology, Aalto University School of Science, Espoo, Finland; Department of Signal Processing and Acoustics, Aalto University School of Electrical Engineering, Espoo, Finland; Center for Computer Research in Music and Acoustics, Stanford University, Stanford, CA, USA
Although the special purpose graphics processing unit (GPU) found in many personal computers was designed to accelerate complex visual displays, the GPU can also be used for audio signal processing. Several case studies illustrate the use of the GPU in audio applications. For example with additive synthesis, 2 million sine waves were computed in real time, which is about a thousandfold better than a CPU implementation. FFTs and filters also were faster when using a GPU.
Download: PDF (HIGH Res) (4.0MB)
Download: PDF (LOW Res) (613KB)
Authors:Kolundzija, Mihailo; Faller, Christof; Vetterli, Martin
Affiliation:School of Computer and Communications Sciences, Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland
Analyzing the sound pressure field as a multivariate function of spatial location and time illustrates the working principles of gradient and differential microphone arrays. This analysis provides the means for the adjustment of various array parameters, such as array orientation and shape, inter-microphone distances, and microphone signal delays. By using this new perspective, designers have a clear interpretation of gradient and differential microphones and microphone arrays.
Download: PDF (HIGH Res) (641KB)
Download: PDF (LOW Res) (361KB)
Authors:Laitinen, Mikko-Ville; Kuech, Fabrian; Disch, Sascha; Pulkki, Ville
Affiliation:Department of Signal Processing and Acoustics, Aalto University, Espoo, Finland; Fraunhofer Institute for Integrated Circuits IIS, Erlangen, Germany
Reproduction of surrounding applause-type signals is known to produce audible artifacts with parametric spatial audio coders. This arises from inadequate temporal resolution. Better quality results if the temporal resolution is high enough to analyze each handclap separately, which can be obtained by using different temporal windows for different frequencies. Formal listening tests confirm the improved subjective quality.
Download: PDF (HIGH Res) (626KB)
Download: PDF (LOW Res) (375KB)
Authors:Klippel, Wolfgang; Schlechter, Joachim
Affiliation:Klippel GmbH, Dresden, Germany
Accurate measurement of the effective radiation area of loudspeakers is important because this parameter determines the acoustical power and the efficiency. The conventional methods fail for microspeakers and headphones when the surround geometry is complex and the excursions do not vary linearly. Two new methods are proposed to handle these cases: using laser sensor and microphone to measure voice-coil displacement and sound pressure, and using laser triangulation scanner to measure vibration and geometry of the radiator.
Download: PDF (HIGH Res) (3.8MB)
Download: PDF (LOW Res) (900KB)
Novel spatial audio processing and rendering tools are enabling the simulation and control of 3-D audio reproduction in a wide range of contexts. From unusual transducer types to the creation and virtualization of believable audio spaces, papers presented at the 40th International Conference exemplified the impressive range of research in this fast-moving field.
Download: PDF (7.9MB)