To better understand human behavior, it is essential to investigate the speech features that contribute the most to emotional expressions. In this study, we investigated how different emotions affect the acoustic properties of speech. This study explored a new set of widely utilized acoustic features to recognize emotions from audios. Experimental investigation using the Bangla and English emotional datasets were conducted using SVM, Random forest, and XGBoost algorithm. We used the Grid Search method with five-fold cross-validation to select the optimal parameters for obtaining the best results from the models. Again a five-fold cross-validation was applied to evaluate the models' effectiveness in emotion perception. The XGBoost analysis was employed to calculate the feature importance of speech emotion identification from the datasets. We found that selecting the most important features allows a high level of accuracy in using ML models that is competitive with deep learning models' performance while utilizing less computational complexity.

Acoustic scene classification (ASC) and sound event detection (SED) are major topics in environmental sound analysis. Considering that acoustic scenes and sound events are closely related to each other, the joint analysis of acoustic scenes and sound events using multitask learning (MTL)-based neural networks was proposed in some previous works. Conventional methods train MTL-based models using a linear combination of ASC and SED loss functions with constant weights. However, the performance of conventional MTL-based methods depends strongly on the weights of the ASC and SED losses, and it is difficult to determine the appropriate balance between the constant weights of the losses of MTL of ASC and SED. In this paper, we thus propose dynamic weight adaptation methods for MTL of ASC and SED based on dynamic weight average (DWA) and multi-focal loss (MFL) to adjust the learning weights automatically. By comparing the two methods, we then clarify how the dynamic adaptation of the loss weights, rather than specific methods of DWA and MFL, generally benefits the joint analysis of ASC and SED based on MTL. Moreover, we investigate how the training of the joint ASC and SED model dynamically progresses and disclose how the loss weights affect their performance.

The road traffic noise prediction model "ASJ RTN-Model," which has been developing by the Research Committee in the Acoustical Society of Japan, is used widely in environmental impact assessment in Japan. In the latest model, ASJ RTN-Model 2018, several coefficients of the calculation equation for the A-weighted sound power level of road vehicles were updated. In European countries as well, various prediction models such as Nord, Harmonoise, and CNOSSOS-EU have been published. Since 2018, the specific noise emission, which can be incorporated into CNOSSOS-EU, was proposed in several countries: CNOSSOS-NL, -FR, and sonROAD. Also, the emission values in CNOSSOS-EU were replaced in 2021. Thus, the values of noise emission and its frequency characteristics were compared between these prediction models using the immission values calculated considering the effects of a barrier and the ground surface. The results demonstrated that a valid relationship exists between the emission values in ASJ RTN-Model 2018 and those in European models. Additionally, in the prediction assuming a reflective ground such as in urban areas, the differences in frequency characteristics of noise emission between the models will be negligible.

Musical string vibration has been the subject of scientific study for centuries. Recent increases in computational power have allowed the exploration of increasingly detailed features of perceptual significance through simulation approaches. The starting point for any simulation is a well-defined model, usually framed as a system of differential equations, with parameters determined by measurement and experiment. This review article is intended to take the reader through models of string vibration progressively, beginning with well-known and well-studied linear models, and then introducing new features that form the basis for the modern study of realistic musical string vibration. These include, first, nonlinear excitation mechanisms, such as the hammer-string and bow-string interaction, and then the collision mechanism, both for pointwise obstructions and over a distributed region. Finally, the linear model of string vibration is generalized to include geometric nonlinear effects, leading to typical nonlinear behaviour such as pitch glides and the appearance of so-called phantom partials due to nonlinear mixing of modes. The article concludes with a general overview of numerical simulation techniques for string vibration.

Our previous study described a procedure for calculating fluctuation strength (FS) from tremolo produced with a mandolin. However, the forms of the amplitude envelope on sounds played on the mandolin vary so little that how the forms affect the perception of fluctuation has not yet been investigated. Therefore, we present a procedure for calculating FS by considering the form of the amplitude envelope on a tremolo produced with a marimba, whose forms of the acoustic waveform on the tremolo vary due to the stiffness of the head of the mallet. We obtained acoustic features of a marimba tremolo as a set of parameters and calculated FS using the parameters. We found that the procedure observing three features concerning constant fluctuation, the deviation for time and amplitude, and deformation on the envelope can calculate FS correctly.

We report on the effects of internal losses in strings on the attenuation characteristics of tones and vibrations in guitars with nylon, fluorocarbon, and phosphor bronze strings. Radiated sounds and vibrations of bridges generated by individually plucking the 1st, 2nd, 3rd strings pressed down against the 1st or 13th fret were studied. Attenuation coefficients of sound and vibration acceleration from the 1st frequency mode to the 10th frequency mode were determined on the basis of time variations of sound pressure and vibration acceleration levels. Storage and loss elastic moduli and elastic loss factors of nylon, fluorocarbon, and phosphor bronze strings were measured using a dynamic mechanical analyzer. The motion equation, which includes the complex elastic modulus, provided the mode frequency relating the inharmonicity and attenuation coefficient depending on the loss elastic modulus. Nylon and fluorocarbon strings demonstrated further increases in attenuation coefficient when they were pressed down against the 13th fret. The motion equation suggests that the results are caused by the loss elastic modulus of the string material. Attenuation coefficients of fluorocarbon strings were smaller than those of nylon strings because of their higher tension and density and smaller diameter, even though their loss elastic moduli were the same.

In this study, we develop a physical model of a grand piano string involving a high-fidelity expression of hammer shank dynamics. A hammer shank is modeled as a general beam, and its governing equation is analyzed using the finite element method (FEM). In addition, the rotational motion of the hammer shank is explicitly expressed and incorporated into the formulation of the FEM. The developed method also involves a more general physical model of piano strings, which is numerically analyzed using the finite difference method (FDM). The proposed method is validated by comparing numerical results with previously reported measurement results.

The previously proposed phantom silhouette method is promising for converting ordinary speech into whispered speech. It is a simple parametric method that uses high-quality vocoder-type speech analysis and synthesis. An ordinary speech sample is first analyzed using the WORLD vocoder. Then, based on the extracted spectral envelope, spectral features are manipulated so that the voice sounds like a whisper. The target speech is synthesized by driving it with white noise instead of the vocal source signal to make the whole speech sound voiceless. In this study, this method was applied to singing voices to generate whisper voices. In addition to actual singing voices, virtual singers' voices were generated using a Vocaloid voice synthesizer, and AI singers' voices synthesized using a NEUTRINO neural singing synthesizer were also tested to generate whisper voices from singing voices.

For the Schumacher model and Adachi–Sato model, which are effective physical models of wind instruments, the present study introduces a more precise sound field expression using a three-dimensional wave equation, instead of one-dimensional analytical solutions of a wind instrument bore. The sound field around a bell section is expressed by the normally differentiated Kirchhoff–Huygens formula, then the boundary element method is applied to numerically analyze the formula. The remaining sound field inside the tube is analyzed with a conventional transfer matrix approach, and the two sound fields are coupled correctly, thereby modeling the total acoustic behavior of the wind instrument. The improved acoustic analysis method is combined with the Schumacher model and Adachi–Sato model to accomplish the physical modeling sound synthesis. Trial numerical calculations demonstrate that the developed method is effective in the sound synthesis of a clarinet and trumpet, involving three-dimensional characteristics of their radiated sound.

In this paper, an improved vibro-acoustic analysis method for cellos is presented. The soundpost and endpin are physically modeled as ordinary beams. The cello box, bassbar, bridge, and neck are physically modeled as shells. Their governing equations are numerically analyzed by the finite element method, yielding an equation of motion of the total multi-degree-of-freedom system. The radiated sound field around the cello box is dealt with using the normally differentiated Kirchhoff–Huygens integral equation, which is numerically analyzed by the boundary element method. All vibration and sound fields are fully coupled. The proposed method is validated through measurements of a cello. As an application of the developed numerical method, the authors investigate the effects of endpin properties.

Interactions between the airflow, elastic body of the lips, and acoustic resonator of the instrument cause self-sustained oscillation of the lips when generating sound using brass instruments, and the steady-state oscillation of the instrument can be expected to be periodic. However, quasi-periodic oscillation or period doubling can also occur, and a cascade of period doublings may further introduce chaos. Therefore, given a set of dynamic equations representing the acoustic behaviors of the airflow, lips, and instrument, a method for detecting and obtaining the periodic solution by adopting a shooting method that relies on the match between the initial and terminal states after the time corresponding to the oscillation period has passed is presented in this paper. Experiments were performed for a trumpet model, where the resonance frequency of the lips and the blowing pressure were used as the main control parameters. The minimum blowing pressure was estimated using a linear stability analysis. The method could capture the corresponding changes in the periodic solution very finely when a small perturbation was successively applied to the control parameters; however, it was less effective when the acoustic load of the instrument was capacitive at the oscillation frequency.

In this paper, the author extends existing analysis methods for bowed string vibration to that involving the dynamics of a soundbox and neck of a bowed string instrument. Almost all the components of the instrument such as strings, nut, tuning pegs, neck, soundbox, and bridge are physically modeled and numerically analyzed by the finite element method, thereby yielding an equation of motion of the total multi-degree-of-freedom (MDOF) system. Then, the MDOF system is introduced into an existing equivalent circuit representation of the vibrations of a bowed string. The developed method is applied to the physical modeling sound synthesis of the erhu, a Chinese traditional bowed string instrument. Trial numerical calculations demonstrate that the developed method is effective in the sound synthesis of the erhu and also has the potential to analyze the three-dimensional spatial vibration of a bowed string and the effect of the interaction between the string and soundbox vibrations.

In this paper, we present a physical model of Japanese temple bells based on the elasticity theory of thin cylindrical shells. The proposed model involves the dynamics of the clapper, called shumoku, thereby constructing a total physical system of a temple bell. The governing equation of a temple bell is solved semi-analytically using the Fourier series expansion in the circumferential direction. The finite difference method (FDM) is applied to the governing equation in the axial direction. These modeling and numerical treatment are an attempt to reduce computational costs while retaining the physical essence of temple bells. Shumoku is physically modeled as a 3D beam, and numerically analyzed by the finite element method (FEM). Simulation results show that the fundamental frequencies calculated by the proposed method are close to those of existing FDM calculations and measured data, while differences in higher-order partial frequencies are observed, indicating that the differences are attributed to the details of the structural features of the bell omitted in the proposed model. Changing the material of shumoku affects the spectrogram of the vibration of the bell, suggesting the importance of physical modeling of clappers to precisely simulate bell systems.

Scordatura, a method of changing the tone color by changing the tuning interval, is sometimes used for playing string instruments. Resonance is one of the important factors in playing stringed instruments. In this study, the frequency response was demonstrated to depend on the difference between tuning temperaments such as Pythagorean tuning and equal temperament. In addition, the decay time was found to increase or decrease with changes in tuning conditions. It was also observed that the inharmonicity of higher-order harmonics was due to the tuning frequency.