Deriving content-specific measures of room acoustic perception using a binaural, nonlinear auditory model

A comparative study of eight human auditory models of monaural processing

A number of auditory models have been developed using diverging approaches, either physiological or perceptual, but they share comparable stages of signal processing, as they are inspired by the same constitutive parts of the auditory system. We compare eight monaural models that are openly accessible in the Auditory Modelling Toolbox. We discuss the considerations required to make the model outputs comparable to each other, as well as the results for the following model processing stages or their equivalents: Outer and middle ear, cochlear filter bank, inner hair cell, auditory nerve synapse, cochlear nucleus, and inferior colliculus. The discussion includes a list of recommendations for future applications of auditory models.

Perceptual Matching of Room Acoustics for Auditory Augmented Reality in Small Rooms - Literature Review and Theoretical Framework

For the realization of auditory augmented reality (AAR), it is important that the room acoustical properties of the virtual elements are perceived in agreement with the acoustics of the actual environment. This perceptual matching of room acoustics is the subject reviewed in this paper. Realizations of AAR that fulfill the listeners' expectations were achieved based on pre-characterization of the room acoustics, for example, by measuring acoustic impulse responses or creating detailed room models for acoustic simulations. For future applications, the goal is to realize an online adaptation in (close to) real-time. Perfect physical matching is hard to achieve with these practical constraints. For this reason, an understanding of the essential psychoacoustic cues is of interest and will help to explore options for simplifications. This paper reviews a broad selection of previous studies and derives a theoretical framework to examine possibilities for psychoacoustical optimization of room acoustical matching.

Perceptual similarity between piano notes: Simulations with a template-based perception model

In this paper, the auditory model developed by Dau, Kollmeier, and Kohlrausch [(1997). J. Acoust. Soc. Am. 102, 2892-2905] was used to simulate the perceptual similarity between complex sounds. As complex sounds, a set of piano recordings was used, whose perceptual similarity has recently been measured by Osses, Kohlrausch, and Chaigne [(2019). J. Acoust. Soc. Am. 146, 1024-1035] using a three-alternative forced-choice discrimination task in noise. To simulate this discrimination task, the auditory model required a new back-end stage, the central processor, which is preceded by several processing stages that are to a greater or lesser extent inspired by physiological aspects of the normal-hearing system. Therefore, a comprehensive review of the model parameters as used in the literature is given, indicating the fixed set of parameter values that is used in all simulations. Due to the perceptual relevance of the piano note onsets, this review includes an in-depth description of the auditory adaptation stage, the adaptation loops. A moderate to high correlation was found between the simulation results and existing experimental data.

Assessing the perceived reverberation in different rooms for a set of musical instrument sounds

Previous research has shown that the perceived reverberation in a room, or reverberance, depends on the sound source that is being listened to. In a study by Osses Vecchi, Kohlrausch, Lachenmayr, and Mommertz [(2017). J. Acoust. Soc. Am. 141(4), EL381-EL387], reverberance estimates obtained from an auditory model for 23 musical instrument sounds in 8 rooms predicted a sound-source dependency. As a follow-up to that study, a listening experiment with 24 participants was conducted using a subset of the original sounds with the purpose of mapping each test sound onto a reverberance scale. Consistent with the literature, the experimental reverberance estimates were significantly dependent on the instrument sound being listened to, but on the top of that, the estimates were significantly correlated with simulated reverberance estimates for the test stimuli as well as for the previously reported long-duration sounds.

Development of a clarity parameter using a time-varying loudness model

The perceived sound clarity is often estimated with the clarity index, which is calculated on the basis of physical acoustic measures that can correlate weakly to the way humans perceive sound for certain test conditions. Therefore, this study proposes a clarity parameter based on a binaural room impulse response processed with a time-varying loudness model. The proposed parameter is validated by calculating the correlation coefficient with subject responses collected from previous listening experiments. Results show that the parameter outperforms the clarity index in most of the tested conditions, but its performance is less robust than parameter for clarity (P_CLA).

Effects of test method and participant musical training on preference ratings of stimuli with different reverberation times

Selecting an appropriate listening test design for concert hall research depends on several factors, including listening test method and participant critical-listening experience. Although expert listeners afford more reliable data, their perceptions may not be broadly representative. The present paper contains two studies that examined the validity and reliability of the data obtained from two listening test methods, a successive and a comparative method, and two types of participants, musicians and non-musicians. Participants rated their overall preference of auralizations generated from eight concert hall conditions with a range of reverberation times (0.0-7.2 s). Study 1, with 34 participants, assessed the two methods. The comparative method yielded similar results and reliability as the successive method. Additionally, the comparative method was rated as less difficult and more preferable. For study 2, an additional 37 participants rated the stimuli using the comparative method only. An analysis of variance of the responses from both studies revealed that musicians are better than non-musicians at discerning their preferences across stimuli. This result was confirmed with a k-means clustering analysis on the entire dataset that revealed five preference groups. Four groups exhibited clear preferences to the stimuli, while the fifth group, predominantly comprising non-musicians, demonstrated no clear preference.

Comparison of psychoacoustic-based reverberance parameters

This study compared psychoacoustic reverberance parameters to each other, as well as to reverberation time (RT) and early decay time (EDT) under various acoustic conditions. The psychoacoustic parameters were loudness-based RT (T_N), loudness-based EDT [EDT_N; Lee, Cabrera, and Martens, J. Acoust. Soc. Am. 131, 1194-1205 (2012a)], and parameter for reverberance [P_REV; van Dorp Schuitman, de Vries, and Lindau., J. Acoust. Soc. Am. 133, 1572-1585 (2013)]. For the comparisons, a wide range of sound pressure levels (SPLs) from 20 dB to 100 dB and RTs from 0.5 s to 5.0 s were evaluated, and two sets of subjective data from the previous studies were used for the cross-validation and comparison. Results of the comparisons show that the psychoacoustic reverberance parameters provided better matches to reverberance than RT and EDT; however, the performance of these psychoacoustic reverberance parameters varied with the SPL range, the type of audio sample, and the reverberation conditions. This study reveals that P_REV is the most relevant for estimating a relative change in reverberance between samples when the SPL range is small, while EDT_N is useful in estimating the absolute reverberance. This study also suggests the use of P_REV and EDT_N for speech and music samples, respectively.

Predicting the perceived reverberation in different room acoustic environments using a binaural auditory model

In this paper a binaural auditory model was used to compute reverberance estimates in four simulated halls. For three of the halls different absorption conditions were evaluated. The model estimates (pRev) were obtained using music excerpts of an orchestra consisting of 23 instrument sections and then compared with the room acoustic parameters of reverberation time (T₃₀) and early decay time (EDT) at mid frequencies. Although the results showed that pRev has a higher correlation with EDT rather than with T₃₀, this relationship depends on the properties of the instruments. The simulations show that pRev depends on the presentation level and that for instruments with similar critical-band spectrum, pRev follows a similar trend across acoustic conditions. A computational framework and sound stimuli are provided to encourage the search of experimental evidence of the aspects addressed in this study.