Parametric speakers generate narrow directive sound field using modulated ultrasonic. Small size parametric speakers have a problem in sound quality. Demodulated audible sounds by parametric speakers have poor sound pressure in low and middle frequency bands due to the theory of the finite amplitude sound. The sound pressure of a demodulated sound monotonically increases as the frequency of the demodulated sound increases. To solve these problems, an ultrasonic transducer having a wide single peak or more than two peaks must be fabricated. Our previous paper proposed to use two resonant peaks which are close each other. Two close resonant peaks were obtained by two same sized diaphragms linked each other by a specially designed rod. In this paper, a transducer was developed to better structure which has durability against strong mechanical forces and two electrical inputs to control resonant modes of vibration. Durability was improved by optimization of the junction structure. The resonant mode control was enabled by a phase difference between two electrical signals applied to piezoelectric elements on two diaphragms. We succeeded in boosting mid-range sounds, by the newly developed transducer.
This paper investigates possible mechanisms of cochlear two-tone suppression (2TS) in models of the hair cell transducer and the cochlea. The hair cell transducer model can be represented by a saturation function. To simulate cochlear mechanics, a nonlinear transmission line cochlear model is used. The mechano-electric transducer curve of the outer hair cell (OHC) is regarded as the source of nonlinearity in cochlear mechanics. The saturation function approximated by a power series can explain 2TS in the OHC transducer model. However, this mathematical formulation cannot account for cochlear 2TS because the cochlear mechanics is more complicated than the saturation in the OHC transducer. To clarify two-tone interference graphically, it is expressed as a residual vector, the entries of which are the frequencies of the probe and suppressor. In this construct, the stronger of two tones introduced to the system nonlinearly reduces the output by vector subtraction. The model accomplished 2TS and displayed similar horizontal and vertical residual vectors. These analytical results suggest that 2TS is obtained from simple interference between the probe and the suppressor in the cochlear mechanics with nonlinear variation of the OHC transducer current.
For speech communication, it is useful to emphasise the sound coming from a target direction. A large-scale microphone array enables a flexible design of the directivity pattern. However, it is difficult to implement a large-sale microphone array onto an ear-type hearing aid system. In this paper, we propose a hat-type hearing aid system composed of 48 MEMS microphones that can increase the flexibility of the sound directivity pattern. We found that it is possible to reduce the number of microphones to be used by considering the hat effect, without affecting the system performance. By compensating the time delays and reducing the number of microphones based on impulse response measurements, the directivity range of the sound directivity pattern was improved by 5 dB. In addition, we conducted a qualitative evaluation that resulted in an intelligibility improvement when using our hearing aid system.
The accuracy of synchronous tapping with a sequence of fixed-interval light flashes (light-synchronized tapping, LST) is impaired by the presence of a sound sequence depending on its temporal relationship with the light flashes. The present study tested the possibility that the LST task can be used as an objective method of estimating auditory detection thresholds without requiring the listener to report directly his/her sensation as in standard audiometry. The experiment used tone bursts as distractor sequences and varied the frequency and level of the tones. The tone level had a statistically significant effect on the distraction level, but the effect of frequency was not significant. Significant distraction was observed for a tone level of, on average, as low as 15 dB above the detection threshold. In other words, once the lowest tone level of the distraction effect is identified, one can expect that the participant's detection threshold would lie at around 15 dB below the level, regardless of the frequency. The results indicate that in principle, the LST task could be used to estimate auditory detection thresholds, although the reliability of the threshold estimation still has to be improved for it to be applicable to audiometry.