The main goal of this work is to provide statistical data related to the acoustical syllabic organization of Spanish speech sounds. Results can be applied to speech recognition and other systems based on probabilistic information. The syllabic spectral patterns of a sample of 68, 141 units are inventoried. Based on their gross spectral shapes, acous ticsyllabic components were grouped into four sets: vowels, non-vocalic periodic sounds, bursts and noise bands. For syllabic type consonant-vowel (55.8%) the preferred combinations were burst-vowel (26.8%) and periodic sound-vowel (18.6%). For the consonant-vowel-consonant type the combination of burst-vowel-periodic sound ranked first (5.5%). The occurrence of acoustic syllabic components at word-initial and word finalpositions of frequently used words was also counted. Vowels either alone or in vowel-consonant pairs predominate at the onset of words. The opposite is true for the combination of a non-vocalic periodic sound with a vowel. Noise bands in con sonant-vowelsyllabic type show a tendency to occur at the offset of words. Burst vowel, which is the more frequently repeated combination, is evenly distributed at word margins. This type of spectral pattern has a high functional load in the acoustical structure of Spanish speech.

View full abstract

Optical beam deflection method is used for real-time measurements of ultrasonic absorp tionin the frequency range 3 to 40 MHz in liquids. A laser beam is deflected by broad bandultrasonic pulse excited with a 1 MHz-quartz transducer, and the deflection signal is spectral analyzed. The dependence of the spectrum on sound-propagation distance gives absorption coefficients at several harmonic frequencies of the quartz. Absorp tionspectrum is obtained in benzene, o-xylene and carbon disulfide.

View full abstract

This paper presents a new method of measuring blood velocities and the Doppler fre quencydistribution that is not limited by the Nyquist frequency. A non-equally-spaced pulse transmission (a staggered pulse repetition time) method is used to calculate the second-order phase difference (velocity) between successive echoes from the blood flow, and the noise-induced fluctuation in the phase difference is suppressed by averaging the phase difference vectors. With conventional pulsed-Doppler equipment, Nyquist's sampling theorem and the relationship between frequency and distance limit the maxi mummeasurable velocity to the reciprocal of twice the pulse repetition time (sampling time). We show theoretically that the proposed method has a larger maximum velocity. When the non-equally-spaced pulse repetition intervals are T and T+T_s, the maximum measurable velocity can be increased to P times the conventional value by selecting T_s equal to T/P (P>1). The results of computer simulation with a high signal-to-noise ratio are in good agreement with the results measured using the proposed method. In vivohigh-speed blood flow (at over the Nyquist-frequency) in the carotid artery and heart is unambiguously measured without aliasing. To illustrate the feasibility of the proposed method, we also compare the spectrum with that obtained using conventional pulsed-Doppler equipment.

View full abstract

A finite element approach to the calculation of nonlinear sound propagation is pro posed.The equation is discretized in space, and is then solved for time by using New mark-β integration scheme. Some numerical demonstrations are made for the non linearsound propagation of a single-shot pulse in air. Under the assumption of a weak nonlinearity, a linearized one-dimensional equation is first considered for verifying the validity of the approach. It is shown that the nonlinear wave propagation including shock formation is stably and accurately simulated by the introduction of the artificial viscosity.

View full abstract

The relationship between alternating current (AC)-induced otoacoustic emissions (OAEs) and the cochlear outer hair cells (OHCs) was examined. Injection of an AC into the scala media of the guinea pig cochlea, together with simultaneous stimulation by a pure tone, induced sound emissions in the ear canal at the frequencies of the AC and intermodulation distortion products. Amplitude of these emissions changed depending on the frequency of the injected AC or stimulating tone. The amplitude and its fre quencydependence were decreased by selective damage to OHCs by ototoxic agents, and the decrease depended on the extent of the OHC damage, especially at the AC injection site. It is suggested that OHC acts as an electromechanical transducer in the cochlea and is responsible for the generation of the AC-induced OAEs. A nonlinearity to produce the intermodulation distortion was more susceptible to the OHC damage than the electromechanical transducing process.

View full abstract

The performance of Cross Correlation Method for identifying aircraft noise was im provedby adding a pre-whitening procedure in advance of the computation of the cross correlation function. The pre-whitening procedure was realized using one or both of pre-emphasis with first order differentiation of signals and inverse filtering based upon linear prediction. In the result of examination with about half of the conventional microphone spacing, the cross correlation function was successfully modified into im pulselike shape by the effect of pre-whitening. It was able to detect changes in time delay of correlation peak, even if the maximum elevation angle of sound arrival was only about 10°. From the result of automatic detection of correlation peak, the best rate of correct peak detection was attained when both pre-emphasis and inverse filtering were applied. The greater part of errors was mis-detection to peaks due to ground reflection, especially in the case of taking-off noise beneath the flight path. Then, a means using a shielding disk and inverted microphone technique was invented for de creasingdetection errors. Finally, a prototype device implementing pre-whitening was successfully developed, resulting in the confirmation of real time operation and improved ability of correlation peak detection.

View full abstract

In IEC 1043 (Instruments for the Measurement of Sound Intensity, 2nd Committee Draft), the pressure-residual intensity index is defined for the evaluation of the phase mismatch between two channels of a sound intensity measuring instrument. However, an evaluation method of the intensity probe is not described clearly in the draft. This may cause a confusion when they are standardized and actually applied for the evalua tionof intensity measuring instruments. In this technical report, we propose a method that is carried out in an anechoic room for the evaluation of an intensity probe or an instrument as a whole. Some technically important points are also mentioned. The accuracy of this method is confirmed by comparing its result with that obtained by a calibrator.

View full abstract

Horizontal axis wind turbine (HAWT) with downwind rotor location is well known to generate an impulsive low-frequency noise. Large-scale downwind machines caused several low-frequency noise problems in overseas countries. This paper describes firstly measured results of the noise generated by a 15m diameter HAWT “WINDMEL.” Generation of an impulsive low-frequency noise by the downwind WINDMEL has been ascertained in the current measurement, though it was of a rather low level. In case of the upwind WINDMEL, generation of this noise could not be ascertained. Next, the predictive calculation of an impulsive low-frequency noise is carried out based on an available method. Namely, assumptions and cautions required for the predic tivecalculation are presented and discussed. Furthermore, applicability of the cal culationmethod is discussed based on the comparisons between the calculated results and the measured data.

View full abstract