In a voltammetry employing an electrode with a diameter of 2 mm, vibrating at ultrasonic frequecy of 38 kHz and at an amplitude of 0.5 through 3μm_p-p, acoustic streamings were investigated by the flow visualizing experiment and the analysis of a two-dimensional finite element method. The experiment revealed that a gas bubble was generated on the electrode surface at the amplitude above 0.7μm_p-p, and the vibrating mode of the bubble played an important role in the acoustic streaming. These mechanisms were successfully simulated by the analysis assuming a vortex pair located on the electrode.

View full abstract

Over-easy characteristic of compressor/limiter is widely used due to the fact that the ratio around a threshold level is changed slowly, and the compression process is smooth and almost inaudible even for a large compression ratio. This paper describes a new design method of over-easy characteristic of compressor/limiter on a DSP for digital audio systems. For the standard characteristic of compressor/limit, we have proposed an efficient method on DSP by calculating the polynomial expression to approximate the power transformation. Based on the approximation algorithm, we devide an over-easy characteristic curve into two parts, i. e. over-easy and standard compression ranges, then apply two polynomial expressions to realize the curves, respectively. By this method, the ideal input-output characteristic can be achieved using two seven-order polynomial expressions for a compression range control of 50dB. The digital signal processing on DSP is about 90 steps for the compressor/limiter. Standard characteristic with higher quality can be obtained by using the same signal processing structure.

View full abstract

It is well known that voice timbre in autophonic production (hearing one's own voice during phonation) is different from that in extraphonic production (hearing one's voice after recording). The differences between the two productions, however, have not been evaluated quantitatively. This results naturally from the fact that there is no objective way to evaluate the voice timbre in autophonic production. The purpose of the present study was to show a method, named the delayed feedback method created by the author, in which voice timbre in autophonic production can be examined objectively, as a function of frequency, and then to evaluate the timbre in comparison with that in extraphonic production quantitatively, when male subjects phonated five Japanese vowels in the spoken mode. The experiment was conducted so as to simulate the sound through a loudspeaker as accurately as possible in its timbre and loudness to that perceived during phonation. It was found that, for all the vowels, the vocal sound was perceived relatively louder in the low frequency region (about 5 dB at 100 Hz) and softer in the high frequency region (about -5 dB at 4 kHz) than the reproduced sound after recording. This may result from the bone conduction elements during phonation. The findings obtained do also, in fact, coincide with experiences in everyday life.

View full abstract

Speaker individualities in fundamental frequency (F₀) contours are investigated through analyses of several speakers'uttered speech and psychoacoustic experiments. The analyses are performed to extract significant physical characteristics of F₀ by using Fujisaki and Hirose's analysis method and the F-ratio of each physical characteristic. The experiments are performed to clarify the relationship between these physical characteristics and the perception of speaker's speech. The stimuli used in the experiments are re-synthesized with manipulated Fo contours and spectral envelopes averaged overall for all speakers by using the Log Magnitude Approximation analysis-synthesis system. The analysis and experimental results indicate that (1) there is speaker individuality in the Fo contours, (2) some specific parameters related to the dynamics of F₀ contours have many speaker individuality features and speaker individuality can be controlled by manipulating these parameters, and (3) although there are speaker individuality features in the time-averaged F₀, they help improve speaker identification less than the dynamics of the F₀ contours.

View full abstract

Automatic speaker identification is the other aspect in speech science, vis-a-vis spoken word recognition. In its traditional realization as a pattern classification task, there are some difficulties, such as the uncertainty of personalities, the dynamic variation of features, and so on. Nevertheless, vector quantization based pattern classifier still gains remarkable interest, because of its simplicity and the recent development in parallel computation technology. There are several reports on the combination technique of LVQ and other method for improvement of the conventional classifier. However, this paper concentrates on a hybrid algorithm of LVQ and DP-matching. The feature normalization capability in both thetime and frequency domains of such a method can decrease the incorrect speaker identification which is caused by the variation of feature vectors in a short-term or a long-term. The experiment of 10 speakers identification shows that the proposed method can produce the better reference vectors, and hence it can promote the correct identification.

View full abstract