A numerical production of /s/-like sound is dealt with by means of computational fluid dynamics in a two-dimensional model of oral front cavity. The basic hypothesis is that the target sound is the outcome of the lowest or first-mode resonance to the pressure field inside the front cavity, whose resonance frequency is in the range 5∼8 kHz. The model domain is composed of a small semi-closed area representing the front cavity, the upstream half-space, the downstream half-space and two channels that connect the upstream/downstream half-spaces to the cavity. Computation was carried out with the method of direct numerical simulation on the two-dimensional Navier-Stokes equations. The pressure wave emitted into the downstream half-space has a continuous spectrum with a single spectral peak around the assumed resonance frequency. The spatial/temporal characteristics of the turbulence and pressure fields are studied in relation to the so-called quadrupole field.
Conventional ultrasonic welding uses one-dimensional linear vibration, which is performed in only one direction and yields low weld strength. A problem with this method is that it may be difficult to perform depending on the direction of installation. In previous studies, we showed that stable welding independent of installation direction can be achieved by using two-dimensional planar vibration. However, the reason for this was unknown. In addition, the vibration of the welding sample in welding using planar vibration has not been reported. In this paper, we investigated the effect of the installation direction on the vibration characteristics of the welding tip and the welding sample. We explained why welding using linear vibration depends on the installation direction of the welding sample, and why welding using planar vibration can achieve stable welding independent of the installation direction.
The purpose of this study is to emphasize sound from a specific place by using an acoustic lens in sound recording. Signal processing related to sound source separation using a microphone array has been studied as a robust sound recording technique against ambient noise; however, real-time processing is still difficult to implement because of its calculation cost. On the other hand, a directional microphone can realize high sensitivity in a specific direction; but, it is unable to show such selectivity for the distance. This paper has proposed a method of removing the arrival time difference of sound waves from a specific position by utilizing an acoustic lens composed of multiple horns. The distance selectivity and speech intelligibility were evaluated for a prototype acoustic lens made by a 3D printer. As a result, it was confirmed that sound recorded with the acoustic lens had a higher sound pressure level than that recorded with a conventional microphone for a desired position, and high speech intelligibility was constantly maintained up to the desired sound source distance.
In this paper, the results of the experiment for young normal-hearing listeners were presented, which is a part of the whole project, the perception experiment for young normal-hearing and elderly listeners. The perception experiment was conducted by using geminate and non-geminate words with /shu/ and /shi/ (referred to as /∫u/ and /∫i/), `shukan,' `shukkan,' `shikan' and `shikkan,' having devoiced and non-devoiced vowels to investigate if the lower limit frequency (FL) and its falling affect to discriminate /∫u/ from /∫i/ with devoiced vowels by young normal-hearing listeners. The synthesised fricative sound was used for C1, and the frication has three types of FL, 1,440 Hz, 2,760 Hz and 4,080 Hz. Moreover, there are two patterns of the frication with each FL: 1) FL of the frication was kept in the same frequency until the end of the frication in each FL, and 2) the frequencies fall from each FL to 480 Hz in the middle of the frication. The rate which they perceive as /∫u/ (referred to as the response rate of /∫u/) for stimuli with devoiced vowels in FL falling condition was higher than in FL non-falling condition. The response rates of /∫u/ in the both devoiced/non-devoiced condition were also higher in case that the FL was 1,440 Hz regardless of FL falling. Hence, these results implied that FL itself and FL falling affect to distinguish /∫u/ from /∫i/ with devoiced vowels by young normal-hearing listeners. In the results of geminate perception, the tendency was observed that the response rates of geminate for the stimuli with devoiced vowels were almost same as the response rates for the stimuli with non-devoiced vowels, suggesting it was possible for young normal-hearing listeners to discriminate geminate from non-geminate even when vowels were devoiced.
We present an auditory scaling method to generate reverberant sounds that more appropriately match the expected auditory impression of a space in a 2D image. Since the conventional method uses linear scale for the regression parameters of reverberation characteristics, correspondence with the human sense scale has been not considered. We have incorporated concepts from psychoacoustics into the reverberate parameters to improve regression performance in an actual environment, including the sound-masking effect, equal-loudness curves, and subjectively-equal reverberation time. Estimation errors in our scaling method were significantly lower than in comparison with previously presented results. The proposed reverb synthesis method was then evaluated in tests, using several scenes to demonstrate its benefits. Our reverb synthesis method can reproduce plausible reverberant sounds from 2D images, which can be used in mixed and augmented-reality applications.