This paper describes a method for separating each of progressive (direct) and retrogressive (reflected) sound waves in real time which are traveling between the source and the reflector and are overlapped mutually. In this method, the sound wave is expressed and measured in the form of a correlation function obtained from the source and receiver signals in the case of scanning the time delay (variable) and fixing the receiving position (parameter). In order to realize the real time separation, three microphones closely set at the receiving position are used. From them, time delay differential and space differential values are derived. Then, by addition and subtraction of these two differential values, each of progressive and retrogressive sound waves can be obtained separately. Some experiments based on the theory in the case of normal incidence are carried out. Then, considering about the measured sound waves concerning the attenuation by distance and the traveling time, it is shown that the theory and the procedure of separating are valid. But, it is not yet about the case of oblique incidence.
In order to achieve quantitative measurement of spatial impressions of sound image, a method using masking pattern, spatial difference of masking level, is proposed. Masking patterns of real and phantom sound image are measured as a lowings:(1)the masking patterns correspond closely to the spatial impressions of the sound image, (2) the maximum of the masking patterns coincides with the azimuth of the sound image, (3) the width of the sound image can be interpreted by the masking patterns. The phantom sound image for the white noise whose cross-correlation (γ) is equal to 1 is wider than the real sound image localized in front of the subject. The phantom sound image for noise with γ=0 is almost as wide as the distance between the two loudspeakers, (4) the phantom sound image for the white noise with γ=1 is localized higher than the real sound image localized at the same height as the subject's head, (5) the masking patterns of the sound image reproduced through headphones correspond to the spatial impressions fairly well.
This paper attempts to find optimally phase-fluctuation and doppler tolerant waveforms, i. e. , signals such that magnitude of correlator output is relatively insensitive for phase-fluctuation and doppler. So, theoretical consideration is made to derive the signal maximizing the wide band ambiguity function, in which both phase-fluctuation and doppler effects are included. An analysis is made by using techniques of variational calculus, with appropriate constraint conditions for making physically realizable waveforms. The derived signals are approximately the liner period modulated ones and found to be very similar to some of the waveforms used by dolphins and bats for echolocation. It is also shown that these waveforms are obtained by changing only the coefficient of the same differential equation derived under the same assumptions (the correlation process for detection and the constraints). This makes us expect the similarity of their signal generating mechanism and signal processing system.
Some experiments concerning the generation of positive and negative stress pulses have been performed for the purpose of investigating the response of material under positive and negative shock stress. The electro-magnetic induction method was used for the generation of positive and negative stress pulses. Strip plates made of photoelastic material(DAP) were prepared as specimen. Stress pulses were observed by dynamic photoelasticity method. The stress pulses were confirmed to be positive or negative by using a semiconductor strain gauge. The stress pulses were generated with a good reproducibility and their stress amplitudes and pulse width were easily controlled.
Measurements of daily noise exposures of 315 residences in Nagoya City are made and associated informations for each residence and inhabitant are obtained at the same time. Response of inhabitants to questionnaire concerning loudness, noisiness, annoyance etc. of noise environments in and around residence are analyzed using quantification theory, which explains contributions of noise exposure and other factors to response of inhabitants. The results are summarized as follows. 1) Main factor is read or noise exposure in all responses mentioned above. 2) L_<eq>M, L^(50)_<eq1/6>, and L^(90)_<eq1/6> among noise ratings based on L_<eq> show good correspondence to the reactions of inhabitants. 3) contributions of land use, type of house, train, airplane and season are detected in the responses around residence. In addition to the above factors attributes of subject(i. e. age and family member) have some influences to the responses inside the door. 4) The reactions in and around residence are highly correlated each other. The correlation, however, between the reactions and sleep disturbance is low. 5) Efficiency of analysis and contribution of noise rating depend on the psychological level of reaction.