AC-3 | Acoustics | Wabe | 2023-11-08 | 15:30 - 17:00
curated and moderated by Christopher Sauder Engeler
Perceptual sweet area of 2-channel stereo playback
AC-3-1 | Start 15:30 | Duration 30 min. | Matthias Frank |
Johanna Kristl | Talk (English)
Panning and recording techniques for 2-channel stereo were developed for perfect playback conditions, i.e. a single listener sitting at equidistant position to the loudspeaker pair. However, in practice, these conditions are not met when playing back to a larger audience resulting in large localization errors. This contribution evaluates the perceptual sweet area in a larger loudspeaker system for typcial 2-channel recording techniques and amplitude/delay panning strategies. The evaluation also investigates the effect of different audio signals on the size of the sweet area (stationary noise and speech). Our results can be predicted by a simple localization model and are furthermore applicable to 3D audio playback on loudspeakers that is based on stereophonic principles.
Simulating physically accurate very-high-order spatial room impulse responses
AC-3-2 | Start 16:00 | Duration 30 min. | Finnur Pind | Talk (English)
Spatial (ambisonics) room impulse responses can be used for room acoustics analysis and audio rendering, e.g., binaural and multichannel loudspeaker rendering. Spatial room impulse responses can be measured, but a measurement-based approach is only possible in existing real-world spaces. Moreover, measuring spatial IRs can be time consuming and expensive, especially when going to very high ambisonics order. In this talk we present a method to simulate very high order spatial room impulse responses with a high level of physical accuracy. The method leverages a wave-based room acoustic simulation framework and spherical harmonics post-processing methods, the details of which will be presented in the talk. The simulation and post-processing system is entirely cloud-based, making the whole process efficient and scalable. We demonstrate through various benchmarks that the method yields spatial impulse responses which closely resemble measured spatial impulse responses. The ambisonics order can essentially be arbitrarily high, but we present benchmarks with up to 16th order impulse responses.
Immersive Wave Experimentation and The Matrix
AC-3-3 | Start 16:30 | Duration 30 min. | Dirk-Jan van Manen |
Johan Robertsson | Talk (English)
Immersive Wave Experimentation is a technology, originally developed in Geophysics, that makes it possible to immerse a physical wave propagation experiment in a larger desired numerical environment in real-time. Thus, the waves propagating in the physical experiment drive the numerical simulation and, vice-versa, the waves propagating in the simulation drive the experiment. This is made possible through a combination of active wave control on the boundaries of the experiment and by exchanging boundary conditions in real-time through a low latency high-performance compute system. Not only are reflections from the physical boundaries suppressed while reflections from scatterers in the virtual numerical environment are included, but our exact, full-wavefield approach facilitates multiple scattering between any real and virtual scatterers. Examples of succesful applications of the technology in Geophysics will be presented, demonstrating, e.g., the ability to turn a circular wave propagation domain into a square propagation domain. Another application of the technology is to acoustically cloak objects, while a third application is to clone them. Our two-way technology is focused on immersing entire spaces (of arbitrary shape and dimensions) characterized by complex wavefields originating from a multitude of sound sources (i.e., active environments). We speculate on the applications of our technology to the human scale and liken our technology to a "Matrix" for acoustic waves.