Measurement-based fuzzy interpolation of room impulse responses

C. Huszty, B. Németh, P. Baranyi, F. Augusztinovicz

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Application of room impulse responses (RIRs) to acoustic evaluation and auralization often requires many measurements to get enough information about the hall, or to provide enough flexibility for virtual sound source placements in convolution reverberation. In this paper we propose a measurement-based fuzzy modeling method to approximate the RIR function at an arbitrary location between available measured points, without apriori information on the hall geometry or wall reflection parameters. For the fuzzy model identification we define an accuracy indicator of the spatial density of the source positions and predict the required number of them in a selected hall. This indicator quantifies the relationship of the early reflections, determined for various measured positions. This paper also proposes a method that treats non-uniform spatial sampling of the measurement positions, and its implementation for 2D cases is shown. Non-uniform spatial sampling can be useful when RIRs at some source positions - e.g. positions of musicians on a stage of a concert hall - are known or have to be measured precisely, but RIRs at locations in between require an approximation only. The proposed fuzzy model of RIRs actually transforms the measured information into a uniform and tensor product form, enabling the analyst to use further matrix and tensor algebra based numerical methods.

Original languageEnglish
Title of host publicationProceedings - European Conference on Noise Control
Pages5827-5832
Number of pages6
Publication statusPublished - 2008
Event7th European Conference on Noise Control 2008, EURONOISE 2008 - Paris, France
Duration: Jun 29 2008Jul 4 2008

Other

Other7th European Conference on Noise Control 2008, EURONOISE 2008
CountryFrance
CityParis
Period6/29/087/4/08

Fingerprint

Impulse response
rooms
interpolation
impulses
Interpolation
Tensors
Acoustics
Sampling
Position measurement
Reverberation
sampling
tensors
Convolution
Algebra
Numerical methods
acoustics
Identification (control systems)
reverberation
Acoustic waves
convolution integrals

ASJC Scopus subject areas

  • Acoustics and Ultrasonics
  • Public Health, Environmental and Occupational Health
  • Building and Construction
  • Mechanical Engineering
  • Industrial and Manufacturing Engineering
  • Automotive Engineering
  • Aerospace Engineering

Cite this

Huszty, C., Németh, B., Baranyi, P., & Augusztinovicz, F. (2008). Measurement-based fuzzy interpolation of room impulse responses. In Proceedings - European Conference on Noise Control (pp. 5827-5832)

Measurement-based fuzzy interpolation of room impulse responses. / Huszty, C.; Németh, B.; Baranyi, P.; Augusztinovicz, F.

Proceedings - European Conference on Noise Control. 2008. p. 5827-5832.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Huszty, C, Németh, B, Baranyi, P & Augusztinovicz, F 2008, Measurement-based fuzzy interpolation of room impulse responses. in Proceedings - European Conference on Noise Control. pp. 5827-5832, 7th European Conference on Noise Control 2008, EURONOISE 2008, Paris, France, 6/29/08.
Huszty C, Németh B, Baranyi P, Augusztinovicz F. Measurement-based fuzzy interpolation of room impulse responses. In Proceedings - European Conference on Noise Control. 2008. p. 5827-5832
Huszty, C. ; Németh, B. ; Baranyi, P. ; Augusztinovicz, F. / Measurement-based fuzzy interpolation of room impulse responses. Proceedings - European Conference on Noise Control. 2008. pp. 5827-5832
@inproceedings{216723659ea943a48000e267c3e4060f,
title = "Measurement-based fuzzy interpolation of room impulse responses",
abstract = "Application of room impulse responses (RIRs) to acoustic evaluation and auralization often requires many measurements to get enough information about the hall, or to provide enough flexibility for virtual sound source placements in convolution reverberation. In this paper we propose a measurement-based fuzzy modeling method to approximate the RIR function at an arbitrary location between available measured points, without apriori information on the hall geometry or wall reflection parameters. For the fuzzy model identification we define an accuracy indicator of the spatial density of the source positions and predict the required number of them in a selected hall. This indicator quantifies the relationship of the early reflections, determined for various measured positions. This paper also proposes a method that treats non-uniform spatial sampling of the measurement positions, and its implementation for 2D cases is shown. Non-uniform spatial sampling can be useful when RIRs at some source positions - e.g. positions of musicians on a stage of a concert hall - are known or have to be measured precisely, but RIRs at locations in between require an approximation only. The proposed fuzzy model of RIRs actually transforms the measured information into a uniform and tensor product form, enabling the analyst to use further matrix and tensor algebra based numerical methods.",
author = "C. Huszty and B. N{\'e}meth and P. Baranyi and F. Augusztinovicz",
year = "2008",
language = "English",
pages = "5827--5832",
booktitle = "Proceedings - European Conference on Noise Control",

}

TY - GEN

T1 - Measurement-based fuzzy interpolation of room impulse responses

AU - Huszty, C.

AU - Németh, B.

AU - Baranyi, P.

AU - Augusztinovicz, F.

PY - 2008

Y1 - 2008

N2 - Application of room impulse responses (RIRs) to acoustic evaluation and auralization often requires many measurements to get enough information about the hall, or to provide enough flexibility for virtual sound source placements in convolution reverberation. In this paper we propose a measurement-based fuzzy modeling method to approximate the RIR function at an arbitrary location between available measured points, without apriori information on the hall geometry or wall reflection parameters. For the fuzzy model identification we define an accuracy indicator of the spatial density of the source positions and predict the required number of them in a selected hall. This indicator quantifies the relationship of the early reflections, determined for various measured positions. This paper also proposes a method that treats non-uniform spatial sampling of the measurement positions, and its implementation for 2D cases is shown. Non-uniform spatial sampling can be useful when RIRs at some source positions - e.g. positions of musicians on a stage of a concert hall - are known or have to be measured precisely, but RIRs at locations in between require an approximation only. The proposed fuzzy model of RIRs actually transforms the measured information into a uniform and tensor product form, enabling the analyst to use further matrix and tensor algebra based numerical methods.

AB - Application of room impulse responses (RIRs) to acoustic evaluation and auralization often requires many measurements to get enough information about the hall, or to provide enough flexibility for virtual sound source placements in convolution reverberation. In this paper we propose a measurement-based fuzzy modeling method to approximate the RIR function at an arbitrary location between available measured points, without apriori information on the hall geometry or wall reflection parameters. For the fuzzy model identification we define an accuracy indicator of the spatial density of the source positions and predict the required number of them in a selected hall. This indicator quantifies the relationship of the early reflections, determined for various measured positions. This paper also proposes a method that treats non-uniform spatial sampling of the measurement positions, and its implementation for 2D cases is shown. Non-uniform spatial sampling can be useful when RIRs at some source positions - e.g. positions of musicians on a stage of a concert hall - are known or have to be measured precisely, but RIRs at locations in between require an approximation only. The proposed fuzzy model of RIRs actually transforms the measured information into a uniform and tensor product form, enabling the analyst to use further matrix and tensor algebra based numerical methods.

UR - http://www.scopus.com/inward/record.url?scp=84874852018&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84874852018&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84874852018

SP - 5827

EP - 5832

BT - Proceedings - European Conference on Noise Control

ER -