Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

An attention-guided algorithm for improving the performance of acoustic simulations

Autio, Hanna LU orcid and Bard, Delphine LU (2019) p.2619-2626
Abstract
When performing acoustic simulations with the purpose of auralization, there is a trade-off between accuracyand speed. In real-time simulations of virtual reality, finding the balance of this trade-off is paramount to achiev-ing the desired result. If successful, the simulation speed is sufficient to provide a seamless acoustic experienceas the agent moves around the space, while still being accurate enough to be realistic. It is generally acceptedthat a 20ms update interval for the impulse response is sufficient for achieving proper interactivity in mostapplications. However, reaching this threshold without degrading the quality of simulation too badly can bechallenging indeed for complex scenes.In this paper, a... (More)
When performing acoustic simulations with the purpose of auralization, there is a trade-off between accuracyand speed. In real-time simulations of virtual reality, finding the balance of this trade-off is paramount to achiev-ing the desired result. If successful, the simulation speed is sufficient to provide a seamless acoustic experienceas the agent moves around the space, while still being accurate enough to be realistic. It is generally acceptedthat a 20ms update interval for the impulse response is sufficient for achieving proper interactivity in mostapplications. However, reaching this threshold without degrading the quality of simulation too badly can bechallenging indeed for complex scenes.In this paper, a compromise between interactivity (response time) and accuracy is suggested for raytracing simu-lations. This compromise mimics the behaviour of the listener or the agent, prioritizing speed or accuracy basedon how the agent behaves. When the listener is actively moving around the space, interactivity is prioritized.When the listener stands still, fully immersing in the experience, accuracy is improved. This is achieved byexploiting a fundamental truth of Monte Carlo simulations: Convergence improves with more samples (Less)
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
Proceedings of the 23rd International Congress on Acoustics : integrating 4th EAA Euroregio 2019 : 9-13 September 2019 in Aachen, Germany
pages
2619 - 2626
publisher
Deutsche Gesellschaft für Akustik
external identifiers
  • scopus:85099329244
ISBN
978-3-939296-15-7
DOI
10.18154/RWTH-CONV-239586
language
English
LU publication?
yes
id
0b603bda-1376-48fe-837f-3c93630dfa3c
date added to LUP
2021-01-26 10:10:46
date last changed
2022-04-19 04:30:09
@inproceedings{0b603bda-1376-48fe-837f-3c93630dfa3c,
  abstract     = {{When performing acoustic simulations with the purpose of auralization, there is a trade-off between accuracyand speed. In real-time simulations of virtual reality, finding the balance of this trade-off is paramount to achiev-ing the desired result. If successful, the simulation speed is sufficient to provide a seamless acoustic experienceas the agent moves around the space, while still being accurate enough to be realistic.  It is generally acceptedthat  a  20ms  update  interval  for  the  impulse  response  is  sufficient  for  achieving  proper  interactivity  in  mostapplications.  However, reaching this threshold without degrading the quality of simulation too badly can bechallenging indeed for complex scenes.In this paper, a compromise between interactivity (response time) and accuracy is suggested for raytracing simu-lations. This compromise mimics the behaviour of the listener or the agent, prioritizing speed or accuracy basedon how the agent behaves.  When the listener is actively moving around the space, interactivity is prioritized.When the listener stands still, fully immersing in the experience, accuracy is improved.  This is achieved byexploiting a fundamental truth of Monte Carlo simulations: Convergence improves with more samples}},
  author       = {{Autio, Hanna and Bard, Delphine}},
  booktitle    = {{Proceedings of the 23rd International Congress on Acoustics : integrating 4th EAA Euroregio 2019 : 9-13 September 2019 in Aachen, Germany}},
  isbn         = {{978-3-939296-15-7}},
  language     = {{eng}},
  pages        = {{2619--2626}},
  publisher    = {{Deutsche Gesellschaft für Akustik}},
  title        = {{An attention-guided algorithm for improving the performance of acoustic simulations}},
  url          = {{http://dx.doi.org/10.18154/RWTH-CONV-239586}},
  doi          = {{10.18154/RWTH-CONV-239586}},
  year         = {{2019}},
}