In this paper, we propose an element selection approach for speech enhancement using deep neural networks (DNNs) targeting small devices with complexity constraints, such as hearing aids. Element selection reduces the input dimensionality by selecting specific elements from the input vector. Unlike other dimensionality reduction algorithms such as principal component analysis, element selection does not require multiplications, making it suitable for low-complexity environments. To optimize which elements are selected, we propose two methods: 1) a linear-regression-based method minimizing the regression error in estimating a target vector from the dimensionality-reduced vector and 2) a pruning-based method that selects elements corresponding to the remaining weight coefficients in the first layer after applying structured pruning to a DNN. We evaluate their performance in a speech enhancement task under complexity constraints, assuming a simple fully-connected network, with no more than 4 × 10⁵ multiplications per inference and an algorithmic delay below 8 ms. Experiments show that the proposed approach under the complexity constraints achieves a scale-invariant source-to-distortion ratio (SI-SDR) improvement of 5.6 dB on average compared to non-processed noisy speech at signal-to-noise ratios -5, 0, and 5 dB, and 2.54 dB SI-SDR improvement compared to simply using only the latest frames.

To enhance the immersive experience of six-degree of freedom (6DoF) content, we previously proposed a distance attenuation model of the human voice - referred to as the head model - that incorporates direction-dependent head diffraction and radiation point on the basis of measurement of human utterances. In this study, we conducted a subjective evaluation experiment to verify the perceptual validity and practical applicability of the proposed head model in content production. The results suggested that the head model tends to represent the distance attenuation of the human voice under various directional and distance conditions more accurately than the conventional model. Furthermore, listeners generally did not report perceptible differences between the voice to which the head model was applied and the measured voice. These findings indicate that the head model is perceptually suitable for use in 6DoF content.

Sound-field intensity can be analyzed by measuring sound pressure and particle velocity. This study examines a previously proposed c–c method, in which the sound pressure and particle velocity can be measured using cardioid microphones. This method was validated for both far and near sound fields. However, to be considered as a viable alternative to the conventional p-p and p–u methods, the c–c method must also be validated in complex sound fields containing interfering sound waves. In our previous study, we took measurements by rotating a single microphone. This study aimed to further validate the c–c method using a pair of face-to-face cardioid microphones by conducting experiments in complex sound fields. The experimental results show that the sound pressure, particle velocity, complex intensity, active intensity, and reactive intensity measurements obtained in interference sound fields using the c–c method corresponded well with those obtained using the p–u probe. These results demonstrate that the c–c method can be applied to complex sound fields, including interference sound fields, and that a sound intensity probe can be constructed using a pair of cardioid microphones even if their cardioid directivities are not ideal.

The author has proposed a modified gap stiffness model incorporated into the Biot model, known as the BIMGS model. The model demonstrated that the experimental results could be reasonably explained. It describes acoustical relaxation caused by local flow in the gap between grains. In this study, the micro-geometric structure of a gap was investigated experimentally, and the results showed the grain diameter dependence of the contact radius and separation distance. Moreover, the compressional wave speeds and attenuation in water-saturated silica sands as a function of frequency were calculated and examined. The results demonstrated the existence of two attenuation peaks.

This paper proposes an audio information-hiding technique termed Audio Anagram, inspired by linguistic wordplay. The technique introduces a cryptographic mechanism analogous to an anagram, in which rearranged data remains perceptually plausible, thereby concealing the presence of the hidden message. The proposed method enables the extraction of secret audio data by reversing a mapping that rearranges the cover audio signal. This approach serves not only as a practically feasible method but also as a creative audio manipulation technique, allowing two distinct audio streams to coexist within a single signal.

The assessment of environmental noise in urban areas is challenged by the issue of effectively modelling complex noise propagation behaviors among buildings. As one of the solutions, the ASJ RTN-Model, a calculation model for road traffic noise, includes an effective calculation method for predicting noise propagation behind buildings, which enables prediction of noise distribution without heavy computational burden. However, the current model is only applicable to fixed frequency characteristic of the noise source with a fixed height at ground level, so further efforts are needed to extend its application range. In previous work, we established a modified model that is frequency-dependent and supports various frequency characteristics of the noise source. Based on the successful demonstrations, we further propose the attempt that extends the model to be applicable to elevated noise source. The effort mainly involves re-investigations of the determination of the geometric variables and constants in the prediction equations, which is then examined by comparing the prediction results to the measured results from scale model experiments.

This study examined how professional opera singers modulate vocal tract configurations to express “bright” and “dark” timbres. In the “bright” condition, F2–F3 frequencies increased with lip opening, whereas in the “dark” condition, F1 and F3 frequencies decreased with pharyngeal expansion and laryngeal lowering. These effects were more pronounced at lower pitches. An untrained participant showed minimal articulatory or acoustic variation across conditions. Simulations based on vocal-tract area functions confirmed F1–F2 changes but only weakly reflected F3, suggesting additional lateral or vertical adjustments beyond the midsagittal plane. These findings clarify articulatory–acoustic mechanisms of timbre control in operatic singing.

The Contrastive Analysis Hypothesis (CAH) and the Speech Learning Model (SLM) make distinct predictions about how L1-L2 phonological similarity affects second language acquisition. This study evaluates these models through an analysis of Mandarin affricate production by Thai learners. Acoustic measurements showed significant differences between learners and native speakers across all parameters examined. Native speaker transcriptions revealed that while Thai learners successfully produced Mandarin's aspiration contrasts, they exhibited difficulties with place and manner distinctions: For place contrasts, learners frequently substituted alveolar (/ts/) and retroflex (/tʂ/) affricates with alveolopalatal counterparts (/tɕ/), and produced unaspirated retroflex affricates (/tʂ/) as alveolar variants ([ts]); For manner distinctions, aspirated affricates (e.g., /tsʰ/, /tʂʰ/, /tɕʰ/) were often misproduced as their homorganic fricatives (/s/, /ʂ/, /tɕ/). While these findings partially support both CAH and SLM, they suggest the need for models to incorporate more detailed phonetic specifications to fully account for L2 production patterns.

This study introduced laser-induced acoustic pulses as sound sources for 1:10 scale model experiments in concert halls. These pulses exhibit high sound pressure, excellent reproducibility, short transient response, and broadband frequency range up to 100 kHz, including instrumental harmonics, making them suitable as sound sources for auralization. To measure three-dimensional sound fields in the scale model, a single microphone was sequentially moved to multiple locations on virtual spherical surfaces with radii of a few millimeters, centered at the receiving point, and room impulse responses were measured at each location. To improve both sound quality and spatial resolution during playback, measurements were performed on three concentric virtual spherical surfaces with different radii, creating fourth-order ambisonics signals covering the audible range from 50 Hz to 10 kHz. This approach ensured sound reproduction and clear sound localization compared to conventional methods, thus enabling high-quality auralization of orchestral performances using 1:10 scale models.

A conventional feedforward active noise control system suffers from degraded performance of noise reduction due to the causality constraint caused by processing and propagation delays. To address this issue, we propose a multichannel feedforward active noise control system that combines an optical laser microphone and an air-conduction microphone. The proposed system relaxes the causality constraint by acquiring vibration information at the speed of light using the optical laser microphone and enhances coherence using the signal of the air-conduction microphone. Experiments using a laser Doppler vibrometer and an electret condenser microphone demonstrate superior noise reduction compared to the conventional system.

The motion of a polystyrene foam sphere in a Kundt’s tube is investigated and confirmed to be driven by the primary acoustic radiation force. When spheres are placed in a straight line in the direction of the tube axis, they are aligned with the tube axis at certain spaces. This phenomenon is investigated using the balance between the primary and secondary acoustic radiation forces acting on the spheres and also using the equation of motion for the spheres. When spheres are close, they exert secondary acoustic radiation forces and adhere to each other, forming lines in the direction perpendicular to the tube axis. These lines are found to be arranged at regular spaces in the direction of the tube axis by the two types of acoustic radiation forces, forming stable striae. The spacing between adjacent striae depends on factors such as sound frequency and sphere size.

An ISO standard on the reference equivalent threshold sound pressure levels for circumaural earphones was established in 2004. This international standard enabled the use of circumaural earphones for pure-tone audiometry, along with supra-aural earphones, which had long been used. However, some reports describe that the hearing threshold levels obtained using supra-aural earphones and circumaural earphones did not always agree even when the same subjects were tested under identical measurement conditions. Clinicians might become confused by this disagreement when the use of circumaural earphones becomes popular. In this study, we examined subjects by pure-tone audiometry using several earphones of both types and an identical measurement procedure. On the basis of the measurement results, the expected amount of threshold difference relative to reference circumaural earphones was calculated for supra-aural earphones of each model.

High-Intensity Focused Ultrasound (HIFU) is a potent treatment for solid tumors. The Energy Efficiency Factor (EEF), defined as the acoustic energy needed to ablate a unit volume of tissue, is pivotal for assessing HIFU's efficiency. However, determining the optimal transducer configuration and irradiation parameters remains challenging. A hybrid approach combining Machine Learning (ML) and Genetic Algorithm (GA) was utilized to minimize the EEF. Multi-Layer Perceptron (MLP), Radial Basis Function (RBF), and Extreme Gradient Boosting (XGB) models were employed to predict EEF across various HIFU transducer settings and irradiation conditions. The most accurate model was chosen for further refinement using GAs to identify minimal EEF values. The MLP model demonstrated superior performance in EEF prediction over RBF and XGB. The MLP-GA model effectively determined the optimal parameters for HIFU ablation, achieving minimal EEF values of 0.68 J/mm³ for continuous wave and 0.70 J/mm³ for pulsed wave modes. The differences between the MLP-GA model's findings and those of a traditional grid search were approximately 3.0%. The integrated ML-GA approach is effective for optimizing HIFU transducer design and irradiation settings to achieve maximal efficiency, defined as the minimal EEF, in simulated coagulative ablation scenarios.

It has been reported that the effects of noise interventions are more clearly observed in terms of annoyance than in activity interferences such as sleep disturbance. This study confirmed these findings by reanalyzing socio-acoustic survey data on community responses to road traffic noise along an arterial road with a tree belt in Tomakomai. The results indicate that the tree belt did not lead to a significant change effect in sleep disturbance. This finding is expected to contribute to future systematic reviews on noise interventions and public health.

The authors had conducted a series of questionnaire surveys on the acoustical quality of concert halls in the Tokyo metropolitan area approximately 30 years ago, and now they have reexamined these data in light of current knowledge. In the surveys, eight to ten musical experts served as listeners at 12 orchestral concerts, judging the subjective impression in terms of reverberance, clarity, perceived sound strength, spatial impression (LEV), and overall quality. The surveys consisted of two rounds, each with slightly different sets of subjective attributes. Regression analyses were conducted to investigate the relationships between the subjective attributes and their corresponding physical parameters. In the first round, the relationships between subjective attributes and physical parameters differed from those observed today, when relatively low reverberant spaces were evaluated more favorably. In the second round, the relationships approached those currently recognized as such. This shift is likely attributable to historical context. That is, during the first round, concerts were typically held in multipurpose halls, while around the time of the second round, music-dedicated concert halls began to open.

Optical music recognition technology has significantly enhanced the efficiency and accuracy of computational score transcription through deep learning methodologies. Although current techniques demonstrate strong performance in processing monophonic and single-voice scores, they struggle to achieve comparable accuracy when handling complex scores containing harmonic intervals, chords, polyphony, or multivoice compositions. In this paper, we propose ConvNeXt-Transformer Fusion (CNTF), an autoregressive end-to-end neural network framework employing an image-to-sequence architecture specifically optimized for automated transcription of intricate musical scores. The model integrates a ConvNeXt-based encoder for sheet music feature extraction and a Transformer-based decoder that generates transcription sequences through autoregressive prediction. To address class imbalance during training, we implement Focal Loss optimization. Experimental results demonstrate that the CNTF model achieves state-of-the-art performance in polyphony-rich score recognition, exhibiting superior character, symbol, and line error rates to existing baseline systems.

Previous studies reported that, when two identical sounds are presented from different directions in the median plane, the sound image perception differs from that in the horizontal plane. The current study performed a psychoacoustical experiment to explore the summing localization occurring in the median plane. The results coincide with the literature, showing weaker summing localization in the median plane than in the horizontal plane. Furthermore, the results indicate an upper-lower asymmetricity and inter-individual variability in the summing localization in the median plane. These results also hold for the case without time delay between the two sounds, namely, the amplitude panning.

This study proposes a non-destructive testing method that combines air-coupled ultrasound and nonlinear harmonic techniques with frequency compounding. The frequency compounding process utilizes nonlinear phenomena observed in air-coupled ultrasound to enhance defect detection capability. The objective of this research is to achieve accurate localization and sizing of defects.To validate the proposed method, experiments were conducted using a 400 mm × 400 mm × 3 mm aluminum plate with a 5 mm square artificial thinning area to simulate wall-thinning defects. The effectiveness of the method was evaluated in terms of the accuracy of defect localization and sizing. As a result, it was confirmed that the proposed frequency compounding method, which selectively combines harmonic components up to the 10th order, maintained localization accuracy comparable to that of conventional methods while achieving an improvement in sizing accuracy of up to 50% in the wave propagation direction and 40% in the direction perpendicular to wave propagation. These findings demonstrate that the proposed method enables reliable estimation of both the position and size of defects.

We previously proposed a noninvasive and quantitative method for estimating the aggregate size of red blood cells (RBCs) based on ultrasound backscatter characteristics. In this study, we improved the estimation accuracy by appropriately setting the focal-region width of the ultrasound beam used to calculate the reference scattering power spectra. Phantom experiments using blood-mimicking particles demonstrated that a wider focal region yielded an estimate closer to the true value and reduced the root-mean-square error (RMSE) of the spectral fitting compared with the conventional width. This result indicates that an appropriate focal-region size enhances the accuracy of RBC aggregate size estimation.

When a planar object with a small through-hole at the center was placed near a vibrating surface, a jet flow from the hole was observed. The objective of this study is to clarify the flow velocity distribution and characteristics of the jet phenomenon through measurement and analysis. The airflow with the jet was visualized using particle smoke, a light sheet, and a high-speed camera. The flow velocity fields produced by the jet were calculated by particle image velocimetry. The jet flow from the hole was considered to occur with suction flow around the hole. The maximum velocity of the jet flow was twenty times higher than that of the airflow without the planar object. The sound pressure fields were analyzed by finite element analysis, and high sound pressure was found in the hole. This pressure was higher than that in the air gap between the vibrating surface and the object. Furthermore, it was confirmed that the airflow around the hole fluctuated with driving frequency. Thus, a breathing-like jet was formed in the hole. The flow velocity was increased by increasing the vibration amplitude and decreasing the air gap.

In this study, the validity of the detection of multiple defects in a billet by time-of-flight (TOF) profile and transmission method with linear scanning are evaluated through simulation and experimentation. Simulation results show that TOF measured using the transmission method has higher lateral resolution than using echoes from the defects. Although distinguishing each defect from the TOF profiles becomes difficult when defects are close together, the size of the defect should not be underestimated. As the TOF profiles are consistent between the simulation and experimental results, it is believed that the verification results of the TOF-linear method by simulation are valid.

Additive manufacturing (AM) has emerged as a novel fabrication method capable of producing complex geometries. Among metal AM techniques, laser-based powder bed fusion of metals (PBF-LB/M) is particularly effective for creating intricate structures. However, the mechanical properties of AM parts often exhibit anisotropy and spatial variation. In this study, Inconel 718 components were fabricated using PBF-LB/M, and their mechanical properties—specifically, Young’s modulus—were nondestructively evaluated using laser ultrasonics (LU). The results revealed pronounced anisotropy and location-dependent variations in mechanical properties. Because LU enables nondestructive evaluation at the millimeter scale, it is highly effective for investigating the physical characteristics of AM parts. This study demonstrates the potential of LU as a powerful tool for advancing the understanding of AM material behavior and optimizing AM processes.

The introduction of intense pressure waves such as laser-induced stress waves (LISWs) during electroporation is expected to solve the problem encountered in electroporation by either promoting cell membrane permeability or cell growth. In this study, the repeated LISW introduction exceeding 10 MPa at peak intensity into a versatile electroporation cuvette was experimentally demonstrated. For generating the LISWs, circulating black ink and a transparent soft tube was used as an optical absorber and a plasma-confined medium, respectively. From experimental results, the author found that the peak positive pressure of LISWs in the cuvette was dominant; however, the effect of cavitation was relatively weak because of a small negative pressure. Moreover, the stable repeated generation of LISWs was observed in laser irradiation for more than 30 cycles. The effects of repeated LISW introduction on yeast cell growth were also investigated using the present system. The reduction of yeast growth after LISW introduction of 100 cycles was suggested.

Pulsed pressure waves propagating through a narrow air gap surrounded by spherical walls of different diameters were computationally and experimentally investigated to generate an impact force for noncontact and nondestructive testing. It was confirmed that with the proposed model, an impulse-like wave with peak positive pressure exceeding 10 kPa can be obtained with a gap width of 0.9 mm at a laser energy of 35 mJ.

A cellular polypropylene film metallized with gold electrodes is directly glued to the back plate of a commercially available condenser microphone cartridge whose metallic diaphragm is removed, composed of a microphone set with an existing pre-amplifier and power module. Lack of vibration resonance of the diaphragm makes it possible to widen frequency ranges in pressure sensitivity compared with those of commonly used condenser microphones. In fact, it has been experimentally verified that the prototyped microphone has good response to sound waves over the extremely wide range of frequencies from 200 Hz to 400 kHz and is highly acceptable to intense sound pressures of up to several tens of kilo-pascals.

This study investigates strategies to enhance the loading and dispensing capabilities of a droplet in ultrasonic levitation systems through acoustic field optimization. Using a 28.58 kHz transducer, two approaches were evaluated: (1) horizontal standing waves with a 5° angled reflector at first- to third-order resonances, and (2) inclined standing waves at different angles (first-order resonance) under a fixed surface vibration velocity of 0.8 m/s. Results show that the horizontal configurations with the angled reflector required up to 112.5% higher surface vibration velocity to reach comparable levitation performance to that of parallel reflectors, revealing inefficiencies in reflector-angle adjustments. In contrast, tilting the standing wave angle to 45° significantly enhanced stability, enabling reliable levitation of an averaged 1.0 µL droplet with reduced energy input. The inclined-wave method outperformed reflector-angle modifications, achieving precise droplet insertion and dispensing while minimizing acoustic energy consumption.

Liquid crystal (LC) varifocal lenses are characterized by their need for compactness and high-speed response, rendering them well-suited for next-generation optical devices. The focal length can be modulated by reorienting LC molecules through acoustic radiation force. In this study, the influence of the geometrical structure of ultrasonic LC lenses on the optical performance was examined. Two LC lenses with distinct glass substrate thicknesses were fabricated, and their optical characteristics were evaluated. The electro-mechanical parameters were found to be altered by the thickness of the glass substrate, which consequently led to an improvement in the power consumption in focus tunability.

In general, the resonance frequencies of ultrasonic emitters that use bolt-clamped Langevin transducers differ slightly. However, when using these emitters in an arrayed device, the resonance frequencies of each emitter must be matched. In this paper, the arbitrary reduction of the resonance frequency by adding a small amount of mass after fabrication is examined. The resonance frequency was lowered by adding more mass, demonstrating the utility of this method.

The effect of the cylinder with a cavity around the target location on underwater acoustic streaming at 28.2 kHz was investigated. Finite element analysis was performed to optimize the dimensions of the transparent acrylic cylinder by the resonance frequency analysis of the whole structure to increase sound pressure in the cavity. Simulation methods ignoring cavitation bubbles and considering bubbles were used to obtain distributions of sound pressure and acoustic streaming at the initial period and stable state separately. For comparison, particle image velocimetry experiments were conducted using the adjusted and original cylinders. The results showed that when the gap was smaller than 25 mm, the cavity had an obvious enhancement effect on streaming velocity, increasing it to twice the maximum value.