TY - JOUR A1 - Versümer, Siegbert A1 - Steffens, Jochen A1 - Blättermann, Patrick A1 - Becker-Schweitzer, Jörg T1 - Modelling evaluations of low-level sounds in everyday situations using linear machine learning for variable selection JF - Frontiers in Psychology KW - DOAJ Y1 - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:hbz:due62-opus-23779 SN - 1664-1078 N1 - Supplementary Material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpsyg.2020.570761/full#supplementary-material VL - 11 PB - Frontiers ER - TY - CHAP A1 - Rosenthal, Fabian A1 - Versümer, Siegbert A1 - Steffens, Jochen T1 - Audioinhaltsanalyse und Multilevelmodellierung zur Vorhersage der Bewertung von Indoor Soundscapes T2 - Deutsche Jahrestagung für Akustik (DAGA) Y1 - 2022 PB - DAGA CY - Stuttgart ER - TY - CHAP A1 - Versümer, Siegbert A1 - Steffens, Jochen A1 - Blättermann, Patrick T1 - Subjektive Lautheitsbewertung unter Einfluss situativer und personenbezogener Faktoren T2 - Deutsche Jahrestagung für Akustik (DAGA) Y1 - 2022 CY - Stuttgart ER - TY - GEN A1 - Versümer, Siegbert A1 - Steffens, Jochen A1 - Rosenthal, Fabian T1 - Extensive crowdsourced dataset of in-situ evaluated binaural soundscapes of private dwellings containing subjective sound-related and situational ratings along with person factors to study time-varying influences on sound perception - research data N2 - The soundscape approach highlights the role of situational factors in sound evaluations; however, only a few studies have applied a multi‐domain approach including sound‐related, person‐related, and time‐varying situational variables. Therefore, we conducted a study based on the Experience Sampling Method to measure the relative contribution of a broad range of potentially relevant acoustic and non‐auditory variables in predicting indoor soundscape evaluations. Here we present the comprehensive dataset for which 105 participants reported temporally (rather) stable trait variables such as noise sensitivity, trait affect, and quality of life. They rated 6.594 situations regarding the soundscape standard dimensions, perceived loudness, and the saliency of its sound components and evaluated situational variables such as state affect, perceived control, activity, and location. To complement these subject‐centered data, we additionally crowdsourced object‐centered data by having participants make binaural measurements of each indoor soundscape at their homes using a low‐(self‐)noise recorder. These recordings were used to compute (psycho‐)acoustical indices such as the energetically averaged loudness level, the A‐weighted energetically averaged equivalent continuous sound pressure level, and the A‐weighted five‐percent exceedance level. This complex hierarchical data can be used to investigate time‐varying non‐auditory influences on sound perception and to develop soundscape indicators based on the binaural recordings to predict soundscape evaluations. KW - Dataset KW - soundscape KW - person factors KW - indoor soundscape KW - acoustic environment KW - non-auditory factors KW - situational factors KW - binaural recording KW - Experience Sampling Method KW - acoustic descriptors Y1 - 2023 U6 - https://doi.org/10.5281/zenodo.7193937 N1 - This study was sponsored by the German Federal Ministry of Education and Research. “FHprofUnt” funding code: 13FH729IX6. Versions: Version V.01.1 10.5281/zenodo.7858848 Apr 25, 2023 Version V.01.0 10.5281/zenodo.7193938 Mar 7, 2023 CY - Zenodo ET - V.01.1 ER - TY - JOUR A1 - Versümer, Siegbert A1 - Blättermann, Patrick A1 - Rosenthal, Fabian A1 - Weinzierl, Stefan T1 - A comparison of methods for modeling soundscape dimensions based on different datasets JF - The Journal of the Acoustical Society of America N2 - Soundscape studies vary considerably in study design, statistical methods, and model fit metrics used. Due to this confounding of data and methods, it is difficult to assess the suitability of statistical modelling techniques used in the literature. Therefore, five different methods and two performance metrics were applied to three existing soundscape datasets to model soundscape Pleasantness and Eventfulness based on seven acoustic and three sociodemographic predictors. Datasets differed in soundscape type (urban outdoor vs indoor), experimental setting (field- vs lab-based), size, and study design (site- vs person-centered). The fixed-effects and mixed-effects methods ranged from linear to nonlinear regression based on advanced machine learning approaches. Results showed that models performed better for Eventfulness than for Pleasantness in most cases, while performance as measured by the out-of-sample R2 was dependent on the total variance of the target, especially in both field studies with imbalanced targets and groups. Nonlinear methods consistently outperformed linear regression, with random forest and extreme gradient boosting performing particularly well, while the performance levels of all nonlinear methods remained comparable. Mixed-effects models provided a more generalized, albeit slightly smaller prediction performance when tested on unknown groups. Finally, this study motivates the use of cross-validation with special splitting for analyzing small imbalanced datasets. KW - Acoustical properties KW - Audiometry KW - Psychological acoustics KW - Acoustic ecology KW - Artificial neural networks KW - Machine learning KW - Urban development KW - Finite-element analysis KW - Regression analysis KW - Statistical analysis KW - Akustische Eigenschaft KW - Audiometrie KW - Psychoakustik KW - Neuronales Netz KW - Maschinelles Lernen KW - Stadtentwicklung KW - Finite-Elemente-Methode KW - Regressionsanalyse KW - Statistische Analyse Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:hbz:due62-opus-48849 SN - 1520-8524 VL - 157 IS - 1 SP - 234 EP - 255 PB - AIP Publishing ER - TY - GEN A1 - Rosenthal, Fabian A1 - Versümer, Siegbert A1 - Blättermann, Patrick T1 - flexcv: Easy and flexible nested cross validation for tabular data in python N2 - Python package customizing nested cross validation for tabular data. KW - Software KW - Python (Programmiersprache) Y1 - 2024 U6 - https://doi.org/10.5281/zenodo.10160846 PB - Zenodo CY - Genf ET - Version v24.0-beta ER - TY - CHAP A1 - Rosenthal, Fabian A1 - Blättermann, Patrick A1 - Versümer, Siegbert T1 - flexcv: Python package for fitting, comparing, and logging multiple machine learning models using various cross-validation methods [Poster] T2 - Fortschritte der Akustik - DAGA 2024, 18. - 21. März 2024, Hannover N2 - The evaluation of listening experiments and studies on acoustics not only requires basic knowledge of statistics, but also poses challenges in the implementation of different methodologies. Especially for small sample sizes and peculiarities regarding hierarchical data structures, the need to individualize evaluation scripts arises. Therefore, we introduce flexcv, a powerful machine learning package for Python for evaluating various models on experimental tabular data, especially with small sample sizes. It supports random effects evaluation (including random slopes) for both linear and non-linear regressors, providing broad applicability to different experiments and research questions.flexcv quickly allows to perform nested cross-validation on a variety of models for comparison with each other. On the one hand, the implementation of a flexible interface simplifies the exchange of methods in the script, allowing researchers to change cross-validation methods without having to touch the actual cross-validation code. On the other hand, extensive online logging allows and simplifies the evaluation and experiment tracking along the process and different machines. Y1 - 2024 PB - Deutsche Gesellschaft für Akustik e.V. CY - Berlin ER - TY - CHAP A1 - Erdmann, Matthias A1 - Versümer, Siegbert A1 - Steffens, Jochen T1 - Advanced Soundscape Search: A Web Application for Interactive Exploration of ISO 12913 Datasets [Vortrag] T2 - Fortschritte der Akustik - DAGA 2024, 18. - 21. März 2024, Hannover N2 - The development of the ISO 12913 has led to the creation of standardized methodologies for collecting and analyzing individual responses regarding acoustic environments in context, with the aim of enabling comparability, replicability, and the development of modelling algorithms in soundscape prediction and design tasks. The development of methods to represent and explore soundscape data has the potential to facilitate deeper insights into different datasets. Although it is unlikely that any single method will sufficiently capture all information in empirical studies, attempts are worthwhile to facilitate advancements in this realm. Therefore, we propose a web application that fosters an interactive exploration of datasets based on the ISO 12913. The application facilitates a comprehensive soundscape search using the 8 items of the standard, 8 features characterizing the soundscape (e.g., intensities of natural, human, and technical sound sources heard), and 10 distinct acoustic features. The search encompasses all possible combinations of these components, thus facilitating detailed exploration, offering an in-depth dataset overview, and enabling the comparison of different datasets. Further functionalities include text searching and the option to download, save, and load selected soundscapes, enhancing control and flexibility in the design of empirical studies and supporting urban planners in understanding various soundscape items. Y1 - 2024 UR - https://pub.dega-akustik.de/DAGA_2024/files/upload/paper/336.pdf SN - 978-3-939296-22-5 SP - 1417 EP - 1420 PB - Deutsche Gesellschaft für Akustik e.V. CY - Berlin ER - TY - CHAP A1 - Versümer, Siegbert A1 - Steffens, Jochen T1 - Investigating real indoor soundscapes and their predictability for known and unknown subjects [Vortrag] T2 - Fortschritte der Akustik - DAGA 2024, 18. - 21. März 2024, Hannover N2 - Understanding people's perception of everyday sound involves a complex interplay of various acoustic and non-acoustic factors. Our objective is to uncover important predictors for predicting real indoor soundscapes. To this end, we re-analyzed soundscape data recorded and evaluated by individuals in their homes, aiming for the utmost ecological validity. Given the intricate nature of sound perception and potential non-linear effects of acoustic and psychoacoustic predictors, we utilized machine-learning methods combined with hierarchical linear modeling to analyze 6375 soundscapes from 96 participants including acoustical, psychoacoustical, as well as situational und person-related predictors. Based on this analysis, we measured both the variability within and between subjects and obtained model fits for the prediction of both soundscape pleasantness and eventfulness. In order to discuss cases where predictions for known individuals or places are of interest, we employed (nested) cross-validation using different data-splitting methods. By doing so, our study sheds light on the generalizability of empirical findings when predicting soundscape evaluations for unknown or known subjects and thus supports a deeper understanding of everyday sound perception. KW - Vortrag Y1 - 2024 PB - Deutsche Gesellschaft für Akustik e.V. CY - Berlin ER - TY - CHAP A1 - Versümer, Siegbert A1 - Blättermann, Patrick T1 - Predicting real indoor soundscapes based on auditory and non-auditory factors across different loudness ranges with linear and nonlinear models T2 - INTER-NOISE and NOISE-CON Congress and Conference Proceedings, Conference Proceeding 8, 4 October 2024 KW - Soundscape KW - Lautstärke KW - Psychoakustik Y1 - 2024 U6 - https://doi.org/10.3397/IN_2024_3386 SP - 3882 EP - 3892 PB - Institute of Noise Control Engineering ER - TY - CHAP A1 - Versümer, Siegbert A1 - Blättermann, Patrick A1 - Steffens, Jochen T1 - Nonlinearities in generalized models based on different soundscape datasets [Abstract] T2 - DAS|DAGA 2025: 51st Annual Meeting on Acoustics, March 17-20, 2025, Copenhagen KW - Soundscape KW - Datenanalyse KW - Regressionsanalyse KW - Lineares Regressionsmodell KW - Nichtlineares Regressionsmodell KW - Verallgemeinertes Regressionsmodell Y1 - 2025 UR - https://pub.dega-akustik.de/DAS-DAGA_2025/konferenz-2008.html?article=433 PB - Deutsche Gesellschaft für Akustik e.V. CY - Berlin ER - TY - CHAP A1 - Blättermann, Patrick A1 - Versümer, Siegbert A1 - Steffens, Jochen T1 - Loadings of Acoustical Metrics on Soundscape Items and their relation to Soundscape Dimensions across different datasets [Abstract] T2 - DAS|DAGA 2025: 51st Annual Meeting on Acoustics, March 17-20, 2025, Copenhagen KW - Soundscape KW - Psychoakustik KW - Datenanalyse KW - Maschinelles Lernen Y1 - 2025 UR - https://pub.dega-akustik.de/DAS-DAGA_2025/konferenz-2007.html?article=383 PB - Deutsche Gesellschaft für Akustik e.V. CY - Berlin ER - TY - JOUR A1 - Versümer, Siegbert A1 - Steffens, Jochen A1 - Weinzierl, Stefan T1 - Day-to-day loudness assessments of indoor soundscapes: Exploring the impact of loudness indicators, person, and situation JF - The Journal of the Acoustical Society of America N2 - This study investigates loudness perception in real-world contexts using predictors related to the sound, situation, or person. In the study, 105 participants recorded 6594 sound environments in their homes, which were then evaluated based on the Experience Sampling Method. Hierarchical linear regressions using a loudness level based on ISO 532-1 allowed for obtaining the best model fits for predicting perceived loudness and explaining the highest variance. LAeq and LAF5 provided comparable results and may require less computational effort. However, the analysis shows that only one-third of the variance explained by fixed effects was attributable to the loudness level. Sixteen percent stemmed from perceived properties of the soundscape; 1% were attributable to relatively temporally stable, person-related predictors like participants' age; non-auditory situational predictors made no additional contribution. The results thus did not confirm previous findings on loudness perception under laboratory conditions, emphasizing the importance of the situational context. Along with the current paper, a comprehensive dataset, including the assessed person-related, situational, and sound-related measures as well as LAeq time-series and third-octave spectrograms, is provided to enable further research on sound perception, indoor soundscapes, and emotion. KW - Lautwahrnehmung KW - Akustische Eigenschaft KW - Hörschädigung KW - Lautstärke KW - Audiometrie KW - Sonogramm KW - Regressionsanalyse KW - Gauß-Prozess KW - Auditory perception KW - Acoustical properties KW - Hearing impairment KW - Loudness perception KW - Audiometry KW - Acoustic noise control KW - Spectrograms KW - Acoustic ecology KW - Regression analysis KW - Gaussian processes Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:hbz:due62-opus-46127 SN - 0001-4966 VL - 153 IS - 5 SP - 2956 EP - 2972 PB - Acoustical Society of America (ASA) ER - TY - CHAP A1 - von Berg, Markus A1 - Bitta, Joshua A1 - Versümer, Siegbert A1 - Steffens, Jochen T1 - Investigation of auditory memory through loudness reproduction in the laboratory [Abstract] T2 - DAS|DAGA 2025: 51st Annual Meeting on Acoustics, March 17-20, 2025, Copenhagen KW - Gedächtnis KW - Lautwahrnehmung KW - Lautstärke Y1 - 2025 UR - https://pub.dega-akustik.de/DAS-DAGA_2025/konferenz-1497.html?article=132 PB - Deutsche Gesellschaft für Akustik e.V. CY - Berlin ER - TY - JOUR A1 - von Berg, Markus A1 - Versümer, Siegbert A1 - Bitta, Joshua A1 - Steffens, Jochen T1 - Reduced reproduction levels of outdoor soundscapes are deemed appropriate - even after real-world exposure JF - Acta Acustica N2 - Laboratory experiments in psychoacoustical and soundscape research indicate that participants perceive a reproduction sound level lowered by 8-10 dB as more plausible than the original level. This bias supposedly roots in an adaptation of perceptual loudness scaling to the laboratory environment, that is overall quieter than urban outdoor soundscapes. To gain further insights into the nature of such loudness bias, we conducted a listening experiment in both field and laboratory using a within-subjects design. Thirty-one participants visited a street and listened to the environmental sounds for one minute, while these sounds were also recorded using a dummy head. Thereafter, they listened to the recording in a quiet laboratory nearby and adjusted its level as they remembered it. About half of the sample did this immediately, the other half about 20 minutes after the recording. Results confirm a bias towards lower levels with a mean of about 8.9 dB, regardless of the time between the recording and the reproduction in the laboratory. Also, participants with higher musical abilities tended to select higher, more accurate levels, whereas noise-sensitive participants deemed lower levels appropriate. Results suggest that the hypothesized adaptation of perceptual scaling to the laboratory happens immediately and is affected by individual factors. KW - Psychoakustik KW - Lautwahrnehmung KW - Lautstärke Y1 - 2025 UR - https://doi.org/10.5281/zenodo.17663267 U6 - https://doi.org/10.1051/aacus/2025062 SN - 2681-4617 N1 - Acta Acustica, Forthcoming article, Accepted Manuscript PB - EDP Sciences ER - TY - GEN A1 - von Berg, Markus A1 - Versümer, Siegbert A1 - Steffens, Jochen A1 - Bitta, Joshua T1 - Reproduction levels of outdoor soundscapes in the lab after in-situ exposure: participant responses and recording level envelopes N2 - Supplemental material for https://doi.org/10.1051/aacus/2025062 To investigate loudness biases in the lab, where outdoor sounds are perceived as being too loud if reproduced with physically accurate levels in the lab, a listening experiment was conducted, where thirty-one participants visited a street and listened to the environmental sounds for one minute. These sounds were also recorded using a dummy head. Afterwards, participants listened to the recording in a quiet laboratory and adjusted its level as they remembered it. About half of the sample did this immediately, the other half about 20 min after the recording, during which the performed another listening tests. Afterwards, previous experience with acoustics, musical sophistication and noise sensitivity were assessed. The results revealed a mean gain adjustment of 8.9 dB below the original. The data uploaded here comprises participants gain adjustment, the recording levels averaged other A) the entire recording and B) the first and last 7.5s or 15s, respecitvely and all questionnaire responses as well as aggregate scores for musical sophistication and noise sensitivity, as well as a binary categorization of the respondent into laymen and experienced, based on their previous experience with acoustics. For sociodemographic information, participants age, sex and educational degree are reported as well. Content - 00_Description: - Description of the data files - Includes a list of explanation of the columns names in "data_questionnaires_aggregateLevels.csv - data_questionnaires_aggregateLevels.csv - Participants' adjusted recordings levels - Several level and loudness aggregates of each participant's recording that were analyzed in the paper - Participants' responses to the administered questionnaires - level_envelopes.zip - Archive with 31 CSV files containing level envelopes of each participant's recording. - A-weighted and unweighted level envelope for the dummy head's left and right ear - One-minute recordings with a temporal resolution of 125ms KW - Psychoakustik KW - Lautwahrnehmung KW - Lautstärke Y1 - 2025 UR - https://doi.org/10.1051/aacus/2025062 U6 - https://doi.org/10.5281/zenodo.17663267 PB - Zenodo ER - TY - THES A1 - Versümer, Siegbert T1 - Generalized prediction of quiet indoor soundscapes based on retrospective and in-situ judgements N2 - This dissertation aims to enhance the understanding of how people perceive and react to everyday sounds, particularly in their homes. The focus is laid on indoor soundscapes and the interplay between acoustic measures, individual differences, and contextual factors. A retrospective online study and a field study at peoples’ homes based on the Experience Sampling Method were the ecologically valid ground truth for four key publications. These address significant gaps in soundscape research by providing insights into the factors influencing sound perception in real-life contexts, the evaluation of low-level (i.e., quiet) sounds, and the consequences of (in)appropriate statistical analysis of imbalanced hierarchical soundscape data. The studies reveal that the sound source category is a critical predictor of annoyance and the pleasantness of soundscapes, with different types of sounds (e.g., natural, human, technical) having distinct impacts. Contrary, the effect of acoustic measures like perceived and calculated loudness on sound perception was expected to be high but was masked by contextual factors. These were the perceived control over the acoustic situation and the affective state of a person in the specific situation, proving context-related perceptual measures to be more important than the sound itself. The research consequently highlights the need for automatic sound source identification in complex polyphony everyday sound environments and suggests that individual preferences and liking of sounds could be potential factors in predicting individual sound perception. The dissertation also emphasizes the importance of using appropriate statistical methods, including mixed-effects models and nonlinear regression techniques, to generate more generalizable and more valid results based on hierarchical and imbalanced soundscape data. The findings underscore the limitations of traditional acoustic metrics and advocate for the use of time-series data to better capture the dynamic nature of everyday sound environments in contrast to stimuli typically used in laboratory studies. The author further developed a multi-objective function for avoiding both over- and underfitting during hyperparameter tuning, significantly improving generalized model fitting in soundscape research. Overall, this work contributes to the field of soundscape research by providing a comprehensive analysis of sound perception in indoor environments, highlighting the importance of context, individual differences, and advanced statistical methods in understanding and modeling soundscapes. Finally, future perspectives were discussed, such as the focus on individual preferences, experiences, and expectations for the improved prediction of individual sound perception. N2 - Diese Dissertation verbessert das Verständnis dafür, wie Menschen alltägliche Geräusche wahrnehmen. Der Schwerpunkt liegt dabei auf der Geräuschkulisse in Innenräumen und dem Zusammenspiel von (psycho-)akustischen Messgrößen, individuellen Unterschieden und kontextuellen Faktoren. Eine retrospektive Online-Studie und eine Feldstudie in Wohnräumen der Probanden bildeten die ökologisch valide Grundlage der vier Schlüsselpublikationen. Diese untersuchen Faktoren, die die Geräuschwahrnehmung in realen Kontexten beeinflussen, indem sie die Bewertung von Geräuschen und die Folgen einer (un-)angemessenen statistischen Analyse von unausgewogenen hierarchischen Soundscape-Daten betrachten. Die Studien zeigen, dass die Kategorie der Geräuschquelle ein entscheidender Prädiktor für die Belästigung und die Angenehmheit ist. Entgegen der Erwartung, dass akustische Messgrößen wie die wahrgenommene und berechnete Lautheit einen großen Einfluss auf die Geräuschwahrnehmung haben, zeigte sich jedoch, dass diese durch kontextbezogene Faktoren maskiert wurde. Dabei handelte es sich um die empfundene Kontrolle über die akustische Situation und den emotionalen Zustand einer Person in der spezifischen Situation, wobei sich kontextbezogene Wahrnehmungsmaße als wichtiger erwiesen als das Umgebungsgeräusch selbst. Die Untersuchung unterstreicht folglich die Notwendigkeit einer automatischen Identifizierung von Geräuschquellen in komplexen polyphonen alltäglichen Geräuschumgebungen und legt nahe, dass individuelle Vorlieben für bestimmte Klänge wichtige Faktoren bei der Vorhersage der individuellen Geräuschwahrnehmung sein könnten. Die Dissertation unterstreicht zusätzlich die Bedeutung der Verwendung geeigneter statistischer Methoden, einschließlich Modellen mit gemischten Effekten und nichtlinearen Regressionstechniken, um verallgemeinerungsfähigere Ergebnisse auf der Grundlage hierarchischer und unausgewogener Daten von Soundscape-Studien zu erzielen. Die Ergebnisse zeigen zudem die Grenzen akustischer und psychoakustischer Metriken auf und implizieren die Verwendung von Zeitreihendaten, um die dynamische Natur alltäglicher Geräuschumgebungen besser zu repräsentieren. Darüber hinaus entwickelte der Autor eine Mehrzieloptimierung zur Vermeidung von Über- und Unteranpassung beim Hyperparametertuning, wodurch die Anpassung verallgemeinerbarer Modelle erheblich verbessert wird. Insgesamt leistet diese Arbeit einen wesentlichen Beitrag, indem sie eine umfassende Analyse der Geräuschwahrnehmung in Innenräumen liefert und die Bedeutung von Kontext, individuellen Unterschieden und angemessenen statistischen Methoden für das Verständnis und die Modellierung von Soundscapes hervorhebt. Abschließend werden mögliche Zukunftsperspektiven erörtert, wie z. B. der Fokus auf individuelle Präferenzen, Erfahrungen und Erwartungen für eine verbesserte Vorhersage der individuellen Geräuschwahrnehmung. KW - Psychoakustik KW - Lautwahrnehmung KW - Lautstärke KW - Belästigung Y1 - 2025 U6 - https://doi.org/10.14279/depositonce-23962 PB - Berlin CY - Technische Universität Berlin ER -