Oxidation of hydroquinone to quinone under ultrasonic irradiation has been reported in acidic solutions (Sharma and Prasad). In neutral aqueous solution (pH=6-7), however, we found that this type of reaction does not take place, but another type ultra sonicreaction, production of 2, 5-dihydroxyquinone takes place instead. We inves tigatedthis latter type of reaction and obtained the following results:(i) ultrasonic irradiation produced hydrogen peroxide in solution, which led to the formation of 2, 5-dihydroxyquinone, and the concentration of H₂O₂ formed increased with the ultrasonic level;(ii) the formation reaction of 2, 5-dihydroxyquinone was accelerated by the ultrasonic local thermal action;(iii) the observed reaction was also accelerated by the vigorous ultrasonic stirring.

抄録全体を表示

A new voice conversion technique through vector quantization and spectrum mapping is proposed. This technique is based on mapping codebooks which represent the cor respondencebetween different speakers' codebooks. The mapping codebooks for spectrum parameters, power values, and pitch frequencies are separately generated using training utterances. This technique makes it possible to precisely control voice individuality. The performance of this technique is confirmed by spectrum distortion and pitch frequency difference. To evaluate the overall performance of this technique, listening tests are carried out on two kinds of voice conversions: one between male and female speakers, the other between male speakers. In the male-to-female conversion experiment, all converted utterances are judged as female, and in the male-to-male conversion, 57% of them are identified as the target speaker.

抄録全体を表示

In various kinds of sound and vibration measurements, impulsive excitation is often used as well as stationary random excitation. Concerning such kinds of measurements using impulsive sound or vibration source, it has been theoretically cleared that, by squaring and time-integrating the measured responses of signals like sound pressure and vibration acceleration and by time-integrating the responses of sound intensity, equivalent results corresponding to those measured by stationary random excitation can be obtained. To examine the theoretical relations, three kinds of basic experiments relating to spatial distributions of sound pressure level, vibration acceleration level and sound intensity level were conducted and their validities have been confirmed. Further, as an application of impulsive excitation to practical sound radiation measurement, an experiment performed to investigate the noise reduction of an iron railway bridge made by hammering is introduced.

抄録全体を表示

Segmental-phonological rules are constructed for the phonological component of a semantics-to-speech system of Japanese in a quest for phonetic form generation process to be incorporated into indispensable knowledge for attaining high quality speech syn thesisfrom concept. Through observation of nonmoraic lexical base forms obtaining moraicity on the phonological surface, it is shown that the quintessential mission of the segmental phonology of Japanese lies in preservation of mora structure. The seg mentalrules are classified into three groups: cyclic and noncyclic segmental-phonological rules, and noncyclic phonetic rules, among which mora-preserving rules are concentrated in the cyclic rule class.

抄録全体を表示

A phonetic form generation experiment is performed for the purpose of examining validity of the phonological rules and their application algorithms. The rules reside in the phonological component of a semantics-to-speech system implemented in a computer, and are tiered firstly based on the cyclic rule application principle, and secondly on the two different dimensions of phonological processing: segmental and accentual. Success fulintegration of application of the segmental and accentual rules is achieved through conversion of tree-structured sentences to phonetic form. The rules are installed in the system in a human-editable format, translated by a rule compiler into executable Lisp functions, tested and found to be valid.

抄録全体を表示

Composite complex sinusoidal modeling (CCSM) offered a better bearing characteristic than other methods using auto regressive modeling (ARM). However, neither method can estimate the waveforms. Therefore, a new CCSM to estimate both the directions and waveforms of incident sounds has been devised. From the results of computer simula tionusing 16 sensors, two sound sources with the same power in the 40°and 75°direc tionswere separated, and the waveform of the sound from the 40°direction was estimated to decrease by 34 dB the effect of the sound from the 75°direction.

抄録全体を表示

Pitch detection is an essential technique in the various speech processing systems. The pitch period is useful for applications such as the speaker recognition, the speech synthesis and so on. A new frequency-time domain method in order to obtain a fundamental fre quencyof speech is presented in this paper. The paper also describes the results of the computer simulations and a real time pitch period detector using a conventional Digital Signal Processor. The algorithm uses the Discrete Running Fourier Series (DRFS) for emphasis of the peaks of the speech signal corresponding to the pitch and a new simple rule for the determination of the pitch period.

抄録全体を表示