Measured Room Impulse Responses Research Articles

Dynamic measurement of room impulse responses using a moving microphone

A technique for the recording of large sets of room impulse responses or head-related transfer functions is presented. The technique uses a microphone moving with constant speed. Given a setup (e.g., length of the room impulse response), a careful choice of the recording parameters (excitation signal, speed of movement) leads to the reconstruction of all impulse responses along the trajectory. In the case of a moving microphone along a circle, the maximal angular speed is given as a function of the length of the impulse response, its maximal temporal frequency, the speed of sound propagation, and the radius of the circle. As a result of the presented algorithm, head-related transfer functions sampled at 44.1 kHz can be measured at all angular positions along the horizontal plane in less than 1 s. The presented theory is compared with a real system implementation using a precision moving microphone holder. The practical setup is discussed together with its limitations.

Robust Speech Dereverberation Using Multichannel Blind Deconvolution With Spectral Subtraction

A robust dereverberation method is presented for speech enhancement in a situation requiring adaptation where a speaker shifts his/her head under reverberant conditions causing the impulse responses to change frequently. We combine correlation-based blind deconvolution with modified spectral subtraction to improve the quality of inverse-filtered speech degraded by the estimation error of inverse filters obtained in practice. Our method computes inverse filters by using the correlation matrix between input signals that can be observed without measuring room impulse responses. Inverse filtering reduces early reflection, which has most of the power of the reverberation, and then, spectral subtraction suppresses the tail of the inverse-filtered reverberation. The performance of our method in adaptation is demonstrated by experiments using measured room impulse responses. The subjective results indicated that this method provides superior speech quality to each of the individual methods: blind deconvolution and spectral subtraction.

Segmented swept sine technique for room impulse response estimation

The swept sine technique is very popular for the measurement of room impulse response due to its large signal-to-noise ratio (SNR) and immunity against subtle time-variance and nonlinearity of the room under test. However, the acoustical measurements in the presence of high background noise levels leads usually to unsatisfactory results. For situations of high level nonstationary noise, the mean square of the overall sequence must be minimized in order to increase the SNR. The new technique named segmented swept sine consists in exciting the room with a set of M swept sine signals with application of a dedicated weighting average method and a noise pattern recognition procedure. Each swept sine signal is segmented followed by the estimation of the mean square (MS) value of the respective segment. A weighting procedure is applied to each segment according to the MS value and the likelihood between them. The average technique is then applied. This procedure ensures that the resulting swept sine signal has the highest SNR value. Several examples are presented to compare the standard swept sine and the segmented swept sine techniques, giving advantages and disadvantages of each technique.

Modeling room impulse response by incorporating speaker polar response into image source method

Simple computer modeling of impulse responses for small rectangular rooms is typically based on the image source method, which results in an impulse response with very high time resolution. The image source method is easy to implement, but the simulated impulse responses are often a poor match to measured impulse responses because the description of the source is often too idealized to match the real measurement conditions. For example, the basic image source method has often assumed the sound source to be an omni‐directional point source for ease of implementation, but a real loudspeaker may include multiple drivers and exhibit an irregular polar response in both the horizontal and vertical directions. In this paper, an improved room impulse response computer modeling technique is developed by incorporating the measured horizontal and vertical polar responses of the speaker into the basic image source method. Results show that compared with the basic image source method, the modeled room impulse response using this method is a better match to the measured room impulse response.

An automated high spatial-resolution scanning system for room-acoustic scale models

Experimental study of acoustically complex environments is a lengthy and troublesome process, in part due to the great deal of effort in acquiring data. A custom Cartesian scanning system has been developed to perform automated room impulse response measurements in scale models. High spatial‐resolution measurements are performed within a planar grid and mapped to observe the spatial distribution of the sound field. Phase relationship across the scanning grid is maintained between measurements using continuous pseudo‐random signals. The efficiency and repeatability of the high spatial‐resolution scanning system is demonstrated in acquiring a large number of room impulse responses over a predefined Cartesian grid, while maintaining a wide bandwidth in accordance with the spatial Nyquist theorem. This paper discusses some preliminary results obtained in an eighth scale model of a coupled‐volume system.

Permissible number of synchronous averaging times to obtain reverberation time from impulse response under time-variance conditions

In the measurement of room impulse response, the synchronous averaging technique and such new methods as the MLS and the swept-sine methods are being widely used to improve the signal-to-noise ratio. In actual measurement conditions, however, the air in a room is continuously moving and the temperature is changing to some degree. The measured value of the reverberation time in such a room tends to be shorter at higher frequencies when applying the synchronous averaging. Therefore, the assumption of a time invariant has to be carefully considered, and, on this point, some research has been conducted to date. We also have reported various research results concerning the impulse response measurement under the time-variance conditions. In this paper, the permissible number of synchronous averaging times for reverberation measurement is studied through some field experiments. In each field, many time impulse response measurements were taken between a fixed pair of sound source and receiving positions by the swept-sine method, without averaging. After the measurements, the characteristics and the extent of the time-variance under measuring were estimated by a short-term running cross-correlation function between each impulse response. The influence of the time variance on the synchronous averaging result was studied based on the estimated time variance.

Room volume estimation from diffuse field theory

Among the parameters relevant in room acoustics, the room volume is one of the most important. The general course in room acoustics research is to use the room volume in the prediction of room acoustic parameters such as reverberation time or total relative sound pressure level. Contrary to this, it has been investigated to what extent the room volume can be retrieved from a measured room impulse response. The approach followed is based on room acoustic diffuse field theory and requires correctly measured room impulse responses with the initial time delay corresponding to the source to receiver distance. A total of ten rooms of varying size and acoustic characteristics have been included. The results in three rooms were unreliable, which was explained by the particular acoustic characteristics. In the remaining rooms the results were numerically useful and consistent between different positions within the same room (relative standard deviation around 20%). The influence of source and receiver directivity is also considered.

Verifying two commercial software implementations of impulse-response-based speech intelligibility measurements

A number of objective evaluation methods are currently used to quantify the speech intelligibility in a built environment, including the speech transmission index (STI), rapid speech transmission index (RASTI), articulation index (AI), and the percent articulation loss of consonants (%ALCons). Certain software programs can quickly evaluate STI, RASTI, and %ALCons from a measured room impulse response. In this project, two impulse-response-based software packages (WinMLS and SIA-Smaart Acoustic Tools) were evaluated for their ability to determine intelligibility accurately. In four different spaces with background noise levels less than NC 45, speech intelligibility was measured via three methods: (1) with WinMLS 2000; (2) with SIA-Smaart Acoustic Tools (v4.0.2); and (3) from listening tests with humans. The study found that WinMLS measurements of speech intelligibility based on STI, RASTI, and %ALCons corresponded well with performance on the listening tests. SIA-Smaart results were correlated to human responses, but tended to under-predict intelligibility based on STI and RASTI, and over-predict intelligibility based on %ALCons.

An efficient algorithm for Bayesian energy-decay analysis in architectural acoustics

Sound energy-decay analysis, including experimental determination of reverberation times in single-space enclosures and decay-time estimation in acoustically coupled spaces, has relied primarily on Schroeder backward integration of room impulse responses. Recent work has demonstrated that model-based Bayesian approaches [Xiang et al., J. Acoust. Soc. Am. 110, 1415–1424 (2001); 113, 2685–2697 (2003); 117, 3705–3715 (2005)] can be very useful for such analysis in architectural acoustics. In order to remove any concern of a high computational load within the Bayesian framework, this paper describes a search algorithm for the model-based Bayesian approaches which leads to an efficient analysis method, not only for multiple decay rates in acoustically coupled spaces but also for classical reverberation time estimation. Implementation of the algorithm within the Bayesian framework is discussed, and its performance with respect to other existing algorithms is analyzed using experimentally measured room impulse responses.

Evaluation of decay times in coupled spaces: Reliability analysis of Bayeisan decay time estimation

This paper discusses quantitative tools to evaluate the reliability of "decay time estimates" and inter-relationships between multiple decay times for estimates made within a Bayesian framework. Previous works [Xiang and Goggans, J. Acoust. Soc. Am. 110, 1415-1424 (2001); 113, 2685-2697 (2003)] have applied Bayesian framework to cope with the demanding tasks in estimating multiple decay times from Schroeder decay functions measured in acoustically coupled spaces. A parametric model of Schroeder decay function [Xiang, J. Acoust. Soc. Am. 98, 2112-2121 (1995)] has been used for the Bayesian model-based analysis. The relevance of this work is that architectural acousticians need to know how well determined are the estimated decay times calculated within Bayesian framework using Schroeder decay function data. This paper will first address the estimation of global variance of the residual errors between the Schroeder function data and its model. Moreover, this paper discusses how the "landscape" shape of the posterior probability density function over the decay parameter space influences the individual decay time estimates, their associated variances, and their inter-relationships. This paper uses experimental results from measured room impulse responses in real halls to describe a model-based sampling method for an efficient estimation of decay times, and their individual variances. These parameters along with decay times are relevant decay parameters for evaluation and understanding of acoustically coupled spaces.

Aurally-adequate time-frequency analysis for scattered sound in auditoria

The goal of this work was to apply an aurally-adequate time-frequency analysis technique to the analysis of sound scattering effects in auditoria. Time-frequency representations were developed as a motivated effort that takes into account binaural hearing, with a specific implementation of interaural cross-correlation process. A model of the human auditory system was implemented in the MATLAB platform based on two previous models [A. Härmä and K. Palomäki, HUTear, Espoo, Finland; and M. A. Akeroyd, A. Binaural Cross-correlogram Toolbox for MATLAB (2001), University of Sussex, Brighton]. These stages include proper frequency selectivity, the conversion of the mechanical motion of the basilar membrane to neural impulses, and binaural hearing effects. The model was then used in the analysis of room impulse responses with varying scattering characteristics. This paper discusses the analysis results using simulated and measured room impulse responses. [Work supported by the Frank H. and Eva B. Buck Foundation.]

Decay time uncertainty analysis in acoustically coupled spaces

In order to guide acoustic design and to verify the design goals, evaluations of multiple decay times in coupled spaces are of practical significance when designing coupled volumes in performing art spaces. Traditionally, however, identification of double- or multiple-sloped decay in room impulse response measurement has been considered very challenging. Previous works [Xiang and Goggans, J. Acoust. Soc. Am., 110, 1415–1424 (2001); 113, 2685–2697 (2003)] have applied Bayesian framework to cope with the demanding tasks in estimating multiple decay times from Schroeder decay functions, evaluated from measured room impulse responses. The Bayesian approaches have also been applied in systematic investigation of double-slope decay in coupled volumes with respect to volume ratio, absorption ratio, and aperture size [Bradley and Wang, J. Acoust. Soc. Am. 113, 2188 (2003) (A)]. These recent applications of Bayesian approaches stimulate further study of uncertainties associated with decay time estimation. This paper discusses the estimation uncertainties and interrelationships between multiple decay times within Bayesian framework.

The digital waveguide mesh: Applications in acoustic modeling and current research directions

The digital waveguide mesh (DWM) was first introduced as a modeling technique for musical acoustics applications in 1993 as a natural extension to the 1D digital waveguide that is widely used in physical modeling sound synthesis. DWM related work at York has focused on a number of areas with notable results including reverberation simulation using the 2D triangular mesh, modeling enclosed spaces with the 3D tetrahedral mesh, improved anechoic absorbing boundary conditions (ABCs) and the analysis and validation of spatial properties of DWM-based room impulse response measurements. This paper presents recent results from three current areas of research. RoomWeaver is a development environment that facilitates research into the application of DWM-based models for virtual acoustic spaces, incorporating new hybrid mesh-types in 2D and 3D through the use of KW-pipes. The second parallel research area involves the development of new boundary formulations that extend the anechoic ABC to the more general frequency dependent case with variable diffuse reflection control. Finally, Digitract is a speech synthesis research tool based on a 2D DWM model of the human vocal tract that offers improved control over formant bandwidths and greater naturalness in the production of speech sounds.

Acoustical analysis and multiple source auralizations of charismatic worship spaces

Because of the spontaneity and high level of call and response, many charismatic churches have verbal and musical communication problems that stem from highly reverberant sound fields, poor speech intelligibility, and muddy music. This research looks at the subjective dimensions of room acoustics perception that affect a charismatic worship space, which is summarized using the acronym RISCS (reverberation, intimacy, strength, coloration, and spaciousness). The method of research is to obtain acoustical measurements for three worship spaces in order to analyze the objective parameters associated with the RISCS subjective dimensions. For the same spaces, binaural room impulse response (BRIR) measurements are done for different receiver positions in order to create an auralization for each position. The subjective descriptors of RISCS are analyzed through the use of listening tests of the three auralized spaces. The results from the measurements and listening tests are analyzed to determine if listeners’ perceptions correlate with the objective parameter results, the appropriateness of the subjective parameters for the use of the space, and which parameters seem to take precedent. A comparison of the multi-source auralization to a conventional single-source auralization was done with the mixed down version of the synchronized multi-track anechoic signals.

Bayesian decay time analysis in coupled spaces using a proper decay model

Acoustically coupled spaces have recently been drawing more and more attention in the architectural acoustics community. Determination of decay times in these coupled spaces from measured Schroeders decay function has become increasingly significant. In architectural acoustics practice, identification of decay times from measurements of room impulse responses in these coupled spaces can often require considerable effort. Recent works [N. Xiang, P. M. Goggans, and D. Li, J. Acoust. Soc. Am. 109, 2283 (2001); P. M. Goggans and N. Xiang, ibid. 109, 2383 (2001); N. Xiang and P. M. Goggans, ibid. 110, 1415–1424 (2001)] have demonstrated that Bayesian parameter estimation and model selection prove to be powerful tools for evaluating decay times in coupled spaces. Following a brief introduction to Bayesian model selection, this paper will discuss further results from decay-time evaluations of measured room impulse responses. Emphasis will be given to the behavior of the decay model selection algorithm.

Influence of time variance on room impulse response measurement

The measurement of room impulse response is made under the assumption that the sound propagation system is time invariant. Actually, however, the air in a room is usually moving and the temperature is changing to a greater or lesser extent. In order to examine the influence of such time variance on the measurement of room impulse response, experimental investigations were performed in a reverberation room in which the air was excited by fans and in a concert hall in which the air-conditioning system was operated. Impulse response measurements were performed by the maximum-length sequence (MLS) method and the sweep pulse (SP) method and these results were compared. From the results, it has been found that the reverberation decay tends to become steeper by repeating the impulse response measurement to get high signal-to-noise ratio and the SP method is more robust than the MLS method against the influence of time variance of the atmospheric condition in a room.

Auditory perception of reverberant surroundings

Three experiments on perception of reverberant surroundings were performed. In the first experiment, listeners were presented with binaural recordings of a speech signal and a set of photographs taken in different rooms, and were asked to point out where each recording was made. Most listeners performed very well in this task (average number of correct answers was 70%), showing that reverberation contains information on room properties that the listeners are able to extract. In the second experiment, listeners were asked to judge the relative size of different rooms by listening to the recorded signals. The results show substantial individual differences, with some listeners being close to veridical performance while others had only a few correct answers. A scaling analysis suggests that some listeners use the direct/reverberant energy ratio, while others use reverberation time as a cue for room size. In the third experiment, artificial reverberation generated by a geometric room model (including individual reflections) and from a stochastic model (controlling only energy measures) were compared with signals generated directly from measured room impulse responses. The results show comparable performance, suggesting that energy measures are the most important for assessing information on room parameters. [Work supported by ONR Grant No. N00014-96-1-0675.]