For measuring reverberation decay curve in a room using an impulsive sound source, the following two points, in digital signal processing, are examined. First, theerror caused during the FFT proces, is examined. The next step concerns, the different methods using FFT, the "frequency domain division method" and two "time domain division method", which are then compared. An attempt is made to clarify the problem arising in these two methods on signal processing. As a practical example, an investigation of the fast room acoustic analyzer, which had been developed by our corporation, is shown.
A method of representing Japanese vowels is proposed based on the Principal Component Analysis of the estimated vocal tract area functions. The vocal tract area functions are estimated by an adaptive inverse filtering method and normalized assuming that the volume of the vocal tract is constant. The experimental results are summarized as follows. In case the analysis is limited to single vowels from a group of male speakers, the cumulative contribution rate for the first 3 vectors reaches 92. 3%. in this dimensionality each axis corresponds to the articulatory structures of the vowels and the separation among vowel groups is completed. Then the variations of the area functions of single vowels can be represented by a linear model of 3-dimensions. The linearity of the variations holds also for the vowels uttered continuously. When a difference in the vocal tract length between males and females is taken into account, the cumulative contribution rate of vocal tract variations reaches 91. 2% in the first four principal axis, and the separation of vowels is completed also in this dimensionality. When the child's vocal tract functions are considered together with those adult materials, the dimensionality of the variations is the same as those in the case of the group of males and females. Then the variations of the vocal tract functions with extreme differences in the vocal tract length can be represented by a linear model of 4-dimensions.
For the construction of a bolt-clamped Langevin type longitudinal transducer, the resonance conditions, design method for the clamping-bolt and the statics-stress distributions in the PZT element clamped by the bolt have to be studied. Resonance frequencies are successfully estimated with more than an accuracy of 95%, by using the resonance conditions obtained from the analogy if transmission line analysis. Available diameters of the clamping-bolt are also obtained with regard to both the length of the metal block and the diameter of the transducer, by calculating the stress distribution. These diameter calculations have been experimentally proved resonable.
The annoyance of some fluctuaing noises with extreme temporal features was measured by means of the magnitude estimation method. The noises judged by the subjects examined were traffic noises simulated on the basis of the equally spaced vehicles model, convex type sounds, concave type sounds, increasing sounds, and decreasing sounds. The experimental results of the simulated traffic noise show that L_<eq> is a better predictor of annoyance than L_<10> and L+σ. L_<eq>, however, does not predict the ESL (Effective Sound Level) of the convex type sound. The increasing sound was judged to be more annoying than the decreasing sound when L_<eq> of both sounds were equivalent. L_<eq> does not predict ESL of the increasing or decreasing sound either. From the results obtained for the convex type sounds, increasing sounds, and decreasing sounds, it can be said that the equal energy hypothesis is not valid. Despite the invalidity of the equal energy hypothesis, L_<eq> gives an approximate prediction of the annoyance value as a whole.
A procedure for predicting ESL (Effective Sound Level ) of fluctuating noises on the basis of the additivity of annoyance was proposed. The ESL was deduced from the probability density function of the sound level of fluctuating noise. The procedure succeeded in predicting the ESL of convex type sound, which L_<eq> failed to prodict. But the ESL of increasing sound could not yet be predicted by the procedure. Good agreement was found, however, between calculated ESL and observed ESL as a whole. It was discussed that the duration of aircraft noise should be defined as the duration between the points of 30dB below the peak level from the point of view of the procedure.
Recently, the finite element method has been applied to various acoustic problems and this has made it possible to apply theoretical investigation to problems which had not been analyzed because of the difficulty of analytical treatment or experiment. The finite element method is usually formulated by minimizing the functional characteristics approximately utilizing the variation principle. However, the energy of the system is variable for such a dissipative system as the acoustic system with sound absorption. In such a case, one has carried out calculation for dissipative systems as though they were constant systems by introducing an adjunct system. As another approach to problem of this type, we introduce the approximate equation of the finite element directly from the wave equation by means of the Galerkin method, which provides the best approximation of those methods known as the method of weighted residuals, and carry out numerical calculation for some acoustic systems. We take rectangular rooms with flat walls and with triangular wedge-sharped walls as examples of the two dimentional sound fields in a room and ducts with a rectangular bend as examples of the calculation of the transmission loss, and then we discuss the differences of the result due to the absorption treatment.
The Chinese character is one of important units in pronounciation and recognition of Japanese. Many of words can be anticipated, if we can know, at least, one of Chinese characters composing the words. Experiments were carried out firstly on the transformation of strings of phonemic symbols obtained from an input speech into strings containing Chinese characters, and secondly on the determination of the input speech using the strings.