日本音響学会誌
Online ISSN : 2432-2040
Print ISSN : 0369-4232
29 巻, 6 号
選択された号の論文の7件中1~7を表示しています
  • 駒村 光弥, 曽根 敏夫, 二村 忠元
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 331-338
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
    It is known that there are two kinds of cue in the pitch perception of complex tones, i. e. the temporal cue and spectral one. At a lower pitch the former is said to be dominant while at a higher pitch the latter is dominant. The influence of jitter on the pitch, therefore, may differ according to which cue dominants over the other. In this paper the effects of jitter on the pitch of pulse trains are investigated. Method of pitch-matching is adopted and the periodic pulse trains are used as the matching stimulus. The results obtained are as follows:(1) Accuracy of pitch-matching as a function of relative jitter shows different tendencies in accordance with whether the pulse-rate is in the lower pitch region or in the higher one. For stimuli with the pulse-rate below about 100 pps, even a small jitter deteriorates the accuracy of pitch perception. For higher pulse-rate stimuli with a relative jitter below 0. 1, however, the pitch of the stimulus of mono-polarity pattern brings about the same accuracy as for the periodic pulse train (Fig. 4). For an alternate-polarity pattern, a fairly accurate pitch-maching is obtained up to the condition of a larger amount of jitter(Fig. 5). (2) The accuracy in pitch-matching changes its tendency at a certain pulse-rate around 200〜250 pps for a mono-polarity pattern and at about 400 pps for an alternate-polarity pattern (Fig. 4 and Fig. 5). Since the fundamental frequency of an alternate-polarity pulse train is a half of the pulse-rate, the change of its tendency may occur at the fundamental frequency of about 200 Hz. (3) Judging from the power spectra of jittered pulse trains(Fig. 2 and Fig. 3), it may be concluded that the difference in the tendency comes from the difference in cues in pitch perception. In the higher frequency where the spectral peak is the cue in pitch perception, a fairly accurate pitch-matching is attained as long as the spectral peaks appear in the power spectrum. But in the lower frequency where the temporal information is a dominant cue, even a small jitter deteriorates the accuracy of pitch-matching regardless of the existence of spectral peaks. (4) For a random-polarity pattern whose power spectrum is nearly flat, the error in matching increases above the pulse-rate of about 200 pps(Fig. 8). (5) When the spectral information of jittered pulse trains is reduced by high-pass-filtering, the temporal information becomes a dominant cue in pitch perception up to a higher pulse-rate than in the unfiltered case(Fig. 9). (6) Accuracy of pitch-matching for a high-pass-filtered jittered pulse train becomes worse above 800 pps, and this value corresponds to the upper frequency limit of periodicity pitch perception(Fig. 10).
  • 氏原 淳一, 境 久雄
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 339-346
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
    Spatial response patterns of the secondary neurons for two-tone stimulus are simulated by an electronic model of the auditory system and neurophysiological mechanism for the critical band is discussed on the basis of change of the response pattern to frequency intervals of two components. The model consists of a basilar membrane, hair cells, primary and secondary neurons, and each neuron model has a response area and inhibitory areas by lateral inhibitory connection between neurons. When two-tone stimulus is applied to such a model, spatial response patterns of the secondary neuron are observed as shown in Fig. 7. Response pattern for two-tone (solid line) is not always the sum of the response pattern for single-tone stimulus (broken line) on account of the lateral inhibition and becomes unimodal from bimodal as the frequency interval of two components decreases. By comparing Fig. 8 which shows the response area and the inhibitory areas of the single neuron with the CF of 1300Hz to the transition from bimodal to unimodal response pattern in Fig. 7, we can see that the transition occurs when the frequency interval of two-components is identical with the frequency interval between the response characteristic frequency and the inhibitory characteristic frequency. In masking experiments with two-tone complexes performed by Greenwood, it has been found that the distribution of the masking effect with respect to sound frequency is also unimodal as long as the spacing between the components is less than one critical band (Fig. 1). Above two results presumably represent a similar phenomenon in different ways. We consider that, therefore, the critical band is caused by the lateral inhibition in the auditory neurons and that the critical bandwidth corresponds to the frequency interval between the response characteristic frequency and the inhibitory characteristic frequency. Furthermore, a neural connection function with an equivalent resolution to the critical band is designed for the model.
  • 大槻 茂雄, 奥島 基良
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 347-355
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
    The usual Doppler velocity meter using ultrasonic continuous wave is sensitive to every suspended particle within the intersection of the directional beams of transmitter and receiver. In order to detect selectively the reflected wave from the particles suspended at a given distance in the intersection, an M-sequence signal modulation method was devised and an ultrasonic Doppler velocity meter by this method was constructed for trial. The transmitted signal in this method is continuous wave modulated by M-sequence signal. The selectivity of signal was tested, and the velocity of the mitral valve in human heart was measured. This method was compared with pulsed Doppler method. The block diagram of the M-sequence modulation method is shown in Fig. 4. The signal from a particle to be detected selected by adjusting the delay time τ_d of M-sequence signal to the traveling time of ultrasonic wave via the particle. This method is superior to the pulsed Doppler method(Fig. 6)because larger received signal and greater S-N ratio are available when the same carrier signal and the same clock pulse are used. The signal selectivity to distance of this method is shown in Fig. 10 and Fig. 11. Slightly confronted semi-circular transducers(Fig. 9) were used as transmitter and receiver. An example of the block diagram of the ultrasonic Doppler velocity meter by the M-sequence modulation method is shown in Fig. 12. In order to test this velocity meter, we measured the velocities of heart motion. An example measured on a 34-year-old man of mitral stenosis after operation is shown in Fig. 13. Transducers were held manually on the breast to catch the echo of mitral valve optimally. The ultrasonic cardiogram and the electrocardiogram are shown in Fig. 13(a), the sound spectrogram without the M-sequence modulation method in Fig. 13(b) and those with the M-sequence modulation method in Fig. 13(c) and Fig. 13(d). The depths of the measured positions in the cases of Fig. 13(c) and Fig. 13(d) were 7cm and 8. 2cm respectively.
  • 青島 伸治
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 356-362
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
    M-sequence correlation method is applied to the measurement of flexural wave propagation in a plate. Transfer function G(s) between two points in a wave propagating system is considered. If G(s) has the form of (5), the relation between input and output signals is derived as (6), and group delay τ_&ltgj&gt and amplitude coefficient κ_j can be measured by M-sequence correlation method as (14). Experiments were performed to measure the flexural wave on a long thin steel strip and two wave components were obtained as shown in Fig. 2. The group velocity of the fast wave component is plotted in Fig. 3, which can be explained by conventional one dimensional beam theory. To explain the slow wave component, flexural wave propagation in a two dimensional plate was analysed. General solution (17) of the wave equation was assumed and parameters in it were determined from free edge boundary condition. Characteristic equation (25) was derived by assuming the presence of non-zero solution and it was solved by an electronic computer. There exist two kinds of waves, one is the zeroth order wave which is shown in Fig. 7 and the other is the first order shown in Fig. 10. The zeroth order wave consists from infinite components 01, 02, 03・・・, and the first order wave is very similar to the wave obtained by one dimensional conventional beam theory. From group velocities, mode shapes and cutoff frequencies it is supposed that the fast wave component measured in experiments is the first order wave and the slow wave component is the 02 order wave.
  • 中村 昭
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 363-372
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
  • 西宮 元
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 373-378
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
  • 山本 研二
    原稿種別: 本文
    1973 年 29 巻 6 号 p. 379-387
    発行日: 1973/06/01
    公開日: 2017/06/02
    ジャーナル フリー
feedback
Top