Archives of Acoustics

This paper presents a method for safe retuning of fixed-pitch string instruments to alternative musical scales with fewer degrees than their original design. Our approach uses a systematic monotonic surjective mapping to assign the existing set of strings to a new, smaller set of pitch classes. The primary goal is to preserve the instrument’s timbre and structural integrity by keeping string tension changes within safe limits. We demonstrate the method on a grand piano and an upright piano retuned from 12-tone equal temperament (12-TET, 12EDO) to 10-tone equal temperament (10-TET, 10EDO). Presented approach may be generalized for retuning from N- to M-step scales (N > M) and to other fixed-pitch string instruments. A grand piano was safely retuned using the proposed method and successfully used in a professional concert.

This article explores the potential for modifying acoustic parameters within a theatre hall, using the Maria Zankovetska Theatre in Lviv, as a case study. The study used sensitivity analysis to evaluate how changes in the sound absorption of specific surfaces affect selected acoustic parameters in the horseshoe-shaped theatre hall. A numerical model of the hall is developed and calibrated based on in-situ acoustic measurements to assess the sensitivity of parameters such as reverberation time (T₂₀), early decay time (EDT), clarity (C₅₀), and early sound strength (G₈₀).
The analysis reveals that surfaces such as the stage tower ceiling, stage walls with curtains, audience walls, and seating have the most significant impact on the acoustic parameters. Modifying the sound absorption of these surfaces can affect T₂₀, EDT, C₅₀, and G₈₀. Notably, increasing the absorption of a single surface might not significantly alter C₅₀ and G₈₀ values, whereas reducing the absorption of surfaces such as seating can lead to noticeable changes in these parameters. These findings provide valuable insights for future renovations and acoustic adjustments, aiming to optimize the theatre’s acoustic performance while preserving its historical character.

Modern room acoustics employs a variety of objective measures to characterize the acoustical properties of interiors. Despite these advancements, the relationship between these parameters and subjective assessments of room acoustics remains unclear. Subjective perception, particularly listening effort (LE), plays a critical role in how individuals experience acoustic environments, even when speech intelligibility (SI) is high. This study aims to bridge the gap between objective acoustic measures and subjective listening experiences. We conducted experiments in three rooms equipped with reverberation enhancement systems, resulting in nine different acoustic settings. Objective parameters, including reverberation time (RT), early decay time (EDT), clarity (C50), and the speech transmission index (STI), were measured. Additionally, subjective SI was assessed, and LE was rated on a 7-step Likert scale by 180 volunteers with normal hearing. The analysis revealed a nonlinear relationship between LE and both RT20 and EDT (R² = 0.6), with an even weaker correlation for LE vs. C50 (R² = 0.46). The Pearson correlation coefficient for STI was 0.74, compared to 0.55 for SI. These findings indicate that the relationship between LE and objective parameters, as well as SI, is complex and not straightforward. Our results suggest the importance of incorporating LE into room acoustic design and evaluation. The disparity between objective measures and subjective experiences suggests that LE may be a crucial factor in accurately assessing acoustic environments. This approach sheds the light on a more holistic understanding of acoustic quality that prioritizes human perception.

The article presents the results of comparative studies concerning the efficiency of systems aimed at minimising the acoustic nuisance of noise generated by the railway vehicle movement. The issue of noise in railway traffic is a significant challenge, affecting both human health and the quality of life in the vicinity of railway lines. Prototype sound absorbing panels with varied surface geometry, a rubber slab, and ballast layer (stone aggregate, grain size 31.5/50mm) were examined. Experiments were conducted in a reverberation chamber, analysing the response to broadband noise excitation. The reverberation chamber allows for obtaining repeatable results, eliminating the influence of external sound sources. It enables the assessment of the sound absorption properties of various materials which makes it possible to determine their effectiveness in noise reduction. The research methodology included measurements of reverberation time in different frequency bands for an empty chamber and a chamber containing the tested materials. The obtained differences in reverberation times provide information on the influence of the tested material on the distribution of acoustic energy in individual frequency bands. The research results allow for a preliminary assessment of the effectiveness of the tested materials in the task of reducing railway line noise.

This work is a continuation of the author’s previous research on modeling a piezoelectric sensor–actuator hybrid. It presents the results of vibration and structural sound reduction for a plate with attached piezoelectric elements. The models consist of a steel plate with two piezoelectric actuators attached on one side and a hemispherical air volume on the other side. One of the actuators is used to excite the plate’s vibration and has the same shape and size in all models. The second actuator is used for vibration and structural sound reduction and varies between a standard square-based full actuator and a sensor–actuator hybrid with different sizes and shapes of the sensor component (either square- or disc-based). Harmonic analyses were performed for the first four mode shapes (skipping the third mode since it is a square plate). Optimization was performed using internal ANSYS functions, with the objective of minimizing the sum of displacement vectors at a number of nodes corresponding to either the full plate or the sensor placed on the said plate.

In this study, the impacts of the inclusion of two semi-circular baffles and their orientations on the acoustic performance of the side outlet muffler have been investigated. The side outlet muffler has a circular simple expansion chamber with an axial inlet and a side outlet and two semi-circular baffles that have been placed inside the expansion chamber at different orientations. The axis of the outlet is at the right angle to the axis of the inlet. The acoustical investigation of the side outlet muffler with two semi-circular baffles is done using the plane wave analysis, the finite element method (FEM), and the two-load technique. Based on the orientations of the two semi-circular baffles, three different models of side outlet muffler with semi-circular baffles have been investigated. The plane wave analysis, FEM, and two-load method are applied to all models and it is found that analytical, computational, and experimental transmission loss (TL) are in good agreement. The analytical modelling successfully predicts the presence of semi-circular baffles in the form of peaks and troughs in the TL of side outlet muffler with semi-circular baffles before the cut-off frequency and thus proves its effectiveness. Among all the models, model 2 gives 42% higher TL than model 1 and model 3 shows 16.20% higher TL than model 2. Hence, model 3 proves to be the best design for the side outlet muffler with semi-circular baffles in the attenuation of noise. The model 3 is effective for 1030 Hz–1480 Hz, 1500 Hz–1570 Hz, and 1640 Hz–2400 Hz frequency sound waves. The TL curve, sound pressure contours for model 3, and the band power variation in the 1/3 octave band for all the models have also been presented.

This paper investigates the overfitting problem in vowel classification task for automatic speech recognition (ASR). It utilizes a pitch synchronized human factor cepstral coefficients (PS-HFCC) as the parametrization method, which outperforms traditional methods like HFCC and mel-frequency cepstral coefficients (MFCC) in frame-level classification accuracy. While deep learning models are prevalent in contemporary ASR systems, they often lack explainability, a characteristic of classical classifiers. Therefore, this study examines overfitting phenomenon using a range of classifiers with well-understood properties. Specifically, it analyzes the impact of different training strategies on classifier performance, comparing the susceptibility to overfitting of several widely used classifiers, including the Gaussian mixture model (GMM), a standard approach in speech recognition. The analysis of training strategies considers various data splitting methods: random, speaker-based, and cluster-based. Our analysis of training strategies highlights the crucial role of data splitting methods: while random splitting is commonly used, it can lead to inflated accuracy due to overfitting. We demonstrate that speaker-independent splitting, where the classifier is trained on one set of speakers and tested on a separate, unseen set, is essential for robust evaluation and for accurately assessing generalization to new speakers. Potentially, the resulting insights may inform the future development and training of more reliable ASR systems.

This paper presents a thorough benchmarking analysis of a recently introduced realistic dataset for speech separation tasks. The dataset contains audio mixtures that replicate real-life scenarios and is accompanied by ground truths, making it a valuable resource for researchers. Although the dataset construction methodology was recently disclosed, its benchmarking and detailed performance analysis have not yet been conducted. In this study, we evaluate the performance of four speech separation models using two distinct testing sets, ensuring a robust evaluation. Our findings underscore the dataset’s efficacy to advance speech separation research within authentic environments. Furthermore, we propose a novel approach for assessing metrics in real-world speech separation systems, where ground truths are unavailable. This method aims to improve accuracy evaluations and refine models for practical applications.We make the dataset publicly available to encourage innovation and collaboration in the field

This paper focuses on ultrasonic wireless power transfer (UWPT), a form of underwater wireless power transfer (WPT) using acoustic (ultrasonic) waves. We propose an ultrasonic transducer designed based on the Langevin transducer model, along with its equivalent circuit model for underwater transmission, tailored for efficient medium-distance underwater WPT. The performance of this transducer is simulated and analyzed using the acoustic-piezoelectric structure module in COMSOL. The results demonstrate that the transducer exhibits excellent underwater transmission characteristics. Additionally, an equivalent circuit model of the underwater UWPT system is developed and analyzed to characterize its transmission properties. For validation, a prototype UWPT is fabricated. At a 35 cm separation between the transmitter and receiver, the transmitter power is 2.5 W. With a 1.2 kΩ load, the received root mean square (RMS) voltage is measured at 22.8 V, corresponding to a received power of 433mW and a transmission efficiency of 17 %. These results verify the proposed model, demonstrate a significant improvement in the transmission distance of underwater UWPT systems, and confirm the feasibility of medium-distance UWPT.

Sound source localisation is a fundamental ability for a listener’s functional interaction with the environment, yet it remains a significant challenge for hearing aid users. In general, they perform worse on spatial hearing tests than individuals with normal hearing. This narrative literature review examines the critical factors affecting sound source localisation when hearing aids are worn, with a focus on direction identification. We analysed peer-reviewed articles published over the past three decades to evaluate the impacts of the type of fitting, form factor, acoustic coupling, processing delay, bandwidth, directional microphones, and dynamic range compression on localisation ability. As a general conclusion, there is a consensus in the literature that binaural and open fittings, with microphones at the ear entrance and with extended bandwidth, significantly improve localisation performance. In addition, this review is intended to provide valuable information that can guide future innovations in hearing aid technology to improve users’ hearing experiences and their integration into the environment.