Non-verbal Perception of Room Acoustics using Multi Dimensional Scaling Method
Leonie Böhlke-Grabe, Tim Ziemer, Rolf Bader · 2025
Read Non-verbal Perception of Room Acoustics using Multi Dimensional Scaling Method by Leonie Böhlke-Grabe, Tim Ziemer, Rolf Bader (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-09029.txt.
Introduction
Leonie Böhlke and two co-authors map how listeners perceive concert room acoustics without asking them for words. Instead of rating scales, music is convolved with measured binaural room impulse responses, and listeners judge similarities between the results, letting multidimensional scaling reveal the hidden axes of perception.
Five perceptual dimensions emerge from the scaling. Each one lines up with a measurable acoustic quantity: echo density, fractal correlation dimension, roughness, loudness, and early decay time. The method sidesteps two older habits, matching measures to experts' memories of halls and hunting correlates for bipolar rating scales. Because the dimensions come from the listeners' own similarity judgments, the authors present the approach as a cleaner route to what audiences actually hear in a room.
Catalog record
| CREATOR | Leonie Böhlke-Grabe, Tim Ziemer, Rolf Bader |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-09029
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Non-verbal Perception of Room Acoustics using Multi Dimensional Scaling Method Preprint: AIP/123-QED Leonie Böhlke-Grabe, Tim Ziemer, Rolf Bader Abstract Subjective room acoustics impressions play an important role for the performance and reception of music in concert venues and auralizations. Therefore, room acoustics since the 20th century dealt with the relationship between objective, acoustic parameters and subjective impressions of room acoustics. One common approach is to correlate acoustic measures with experts’ subjective ratings of rooms as recalled from their long-term memory, and explain them using acoustical measures. Another approach is to let listeners rate auralized room acoustics on bipolar scales and find objective correlates. In this study, we present an alternative approach to characterizing the subjective impressions of room acoustics. We concolve music with binaural room impulse response measurements and utilize Multi Dimensional Scaling (MDS) to identify the perceptual dimensions of room acoustics. Results show that the perception of room acoustics has 5 5 dimensions that can be explained by the (psycho-)acoustical measures echo density, fractal correlation dimension, roughness, loudness, and early decay time. I Introduction Room acoustics, particularly important for music venues Beranek 2004 , must be related to music perception in music spaces. Such subjective room acoustics refer to perceptual listening impressions ( Ziemer 2020 , chap. 6) . As with timbre, these impressions are multidimensional, where most studies find subjects to discriminate timbre in three perceptual dimensions (see, e.g., Grey 1978 ; for a review, see ( Bader 2013 , Chap. 11) ). Books on architectural acoustics and spatial audio use various terms to characterize subjective room acoustics, such as reverberance , spaciousness , liveness , spatial impression , listener envelopment , intimacy , or apparent source width Beranek 2004 ( Ziemer 2020 , chap. 6) . Several approaches have been developed to capture subjective room acoustics systematically. The Spatial Audio Quality Inventory (SAQI) Lindau et al. 2014 provides a comprehensive vocabulary for room acoustics and virtual acoustics that includes three to ten terms assigned to the categories timbre , tonalness , geometry , room , time behavior , dynamics , artifacts , and general . Its room acoustical counterpart, the Room Acoustical Quality Inventory (RAQI), suggests items or factors, such as quality , strength , reverberance , brilliance , irregular decay , coloration , clarity , liveliness , or intimacy , plus four additional terms. Beranek suggested 16 attributes and respective measurement parameters to estimate the ’quality’ of concert halls Beranek 1962 . Criticizing the amount of rooms used in listening test, a minimum of 60 rooms was proposed, along with new methods of virtual room acoustics and 3D recordings Kalkandjiev and Weinzierl 2015 . Additionally, the influence of subjective preferences or the music played in the respective rooms was critically discussed, and more objective means are demanded. In a Multidimensional Scaling (MDS) experiment Hawkes and Douglas 1971 , six dimensions were found and associated with balance and blend , resonance , intimacy , brilliance , proximity , and definition . Additionally, a factor analysis was conducted, yielding similar results Cerdá et al. 2009 . The stimuli were taken from different artificial reverberations used in the Royal Festival Hall in London. The subject ratings were done on bipolar scales like ’brilliant – dull’ or ’distant – close’. Also, a pairwise comparison was used to identify latent classes Lachenmayr et al. 2023 . Here, it was found that one group prefers sounds that are proximate, enveloping, and warm, while another prefers clarity. All these subjective attributes can be associated with one or more room acoustical attributes, like Early Decay Time (EDT), RT 20 , strength G, lateral sound level L_{J} , and early lateral energy J_{LF} . Figure 1: Room Impulse Response (IR) measurement setup, using an alarm pistol, shot on a music venue stage. Most of these previous studies used mainly verbalizations to characterize room acoustic, most often adjectives. Such verbalizations were additionally most often presented as bipolar scales. Thereby, the relation between the physics of the rooms and the subjective ratings makes a detour via language, which introduces an additional bias. The present study avoids such verbalizations by only asking for similarities between musical stimuli, performing a Multidimensional Scaling analysis, and correlating the perceptual dimensions with psychoacoustic and room acoustic parameters. Although parameters with high correlation can then be verbalized in a last step, like associating a spectral centroid with perceived brightness or a calculated roughness with perceived roughness, this last step of verbalization is not part of the listening test or the analysis part and might be omitted entirely. The MDS method used in the present study was derived from listening tests in timbre analysis (for a review see ( Bader 2013 , Chap. 11) ) and has several advantages. Firstly, it omits the blurredness of language with the subjective interpretation of adjectives, as discussed above. Secondly, it omits the bias of a finite list of adjectives presented by the experiment beforehand, where some important adjectives might not be present. Thirdly, it omits bipolar scaling, which enforces a subject’s judgment in a bipolar way. Bipolar scales are a bias, as subjects might find both poles to be strongly or weakly present in the sound at the same time. So, e.g., a subject might find a musical sound to be bright and dark at the same time when, e.g., a strong bass instrument is accompanying a bright lead instrument. The cause of such judgments is found in auditory scene analysis, where listeners extract several sonic objects in a sound, making judgments on each of them. In musical terms, this can be a polyphonic texture where each melodic line is perceived separately; it might also be a lead melody accompanied by instruments playing chords. Yet, a fourth advantage is omitting terms like quality from a rating. Although ratings might sometimes be important for listeners, such ratings strongly depend on the musical pieces played and the venue built for each piece. Rooms with variable acoustics point to such complex uses. Previous studies often used clusters to extract main features for room acoustic perception. This includes the search for orthonormal parameters. Indeed, in psychology and psychoacoustics all perceived dimensions are confounded and never orthonormal. The best-known example might be the perceived pitch of a sinusoidal frequency rising when increasing its amplitude while keeping its frequency constant. Therefore, none of the psychoacoustic parameters are expected to be perfectly orthonormal. In cases where more than one parameter is correlating significantly with a perceptual dimension, there is no explicit need to decide on one of them, and additional reasoning needs to be taken into consideration, as discussed below in the results section in detail. One disadvantage of the proposed MDS method is the restriction of the number of sounds possibly used. As proximity estimations need each sound to be compared to all other sounds, already nine sounds used in this paper need 36 pair judgments. Increasing the amount of sounds is very soon overburdening subjects’ concentration. Also, the choice of the dry signal the room impulse responses are convolved with is expected to lead to different results. Therefore, this study is a first step towards non-verbal-based investigations of perceived room acoustics, which already leads to unexpected results with additional psychoacoustic parameters fitting