lated by simple circuits that are easy to design when vibrato sounds are required. First, the analysis of performed musical sounds in order to define the location of principal pitch of vibrato sounds was attempted. But, a definite answer has not been obtained from this approach. As Shonle and Horan2) state, the results of various experiments have variously indicated that principal pitch is localized in the higher side of the extent of vibrato,3) another in the middle,4) and yet another in the lower side.5) This discrepancy results from the fact that musical sounds are not played similar to notes because of the Abraham effect6) or artistic deviation.7) Consequently, it is difficult to locate the principal pitch of vibrato sounds through this approach.

Principal pitch" is defined as somewhat steady pitch perceived from periodically frequency modulated tones, like vibrato tones. Principal pitch of vibrato tones whose frequency alone modulated periodically is indicated to be located around their carrier frequencies (the middle of the extent of FM). We studied where the principal pitch of FM-AM tones is located by the psychoacoustical experiment using a method of adjustment. FM-AM tone is defined as vibrato tone whose frequency and amplitude are periodically modulated by an identical modulation waveform simultaneously. The principal pitch of FM-AM tones is shifted higher than their carrier frequencies when the phase difference between FM and AM is in-phase, and lower when it is anti-phase. The shifted interval from the carrier frequency to the localized principal pitch of FM-AM tones is proportional to the extent of FM and the square of the degree of AM. To interpret these experimental results, we suppose the psychoacoustical model which is the expansion of the pitch averaging mechanism to perceive principal pitch of FM tones. In this model, continuously changing pitch of FM-AM tones is sampled, weighted with its amplitude fluctuation, registered in sensory store, and averaged every certain moment. The weighting process in this model is constructed with loudness perception mechanism and the mechanism to perceive fluctuation strength. We also discussed how the effect of intensity on pitch influences our experimental results.

The definition of timbre has been established by standards such as the American National Standards. According to the results of factor analytical research on sound color, the content of sound color can be well expressed by such descriptive adjectives as soft, rich, clear, etc. This aspect of sound color can be called the “impression expressed by a descriptive adjective of sound color.” Furthermore, there exists another aspect of sound that identifies the sound source such as a piano, violin, etc. Generally speaking, sound impressions are considered to occur in the following order: (1) identification of the sound source, followed by recognition of detailed sound properties such as (2) the impression expressed by the descriptive adjective of the sound color, (3) loudness, (4) pitch, and (5) sound duration. The two aspects of sound, (1) and (2), are usually recognized as sound color or timbre. Accordingly, it is desirable that the content of timbre should be understood more widely as mentioned above.

This paper reports and discusses the results of the second stage of experiments of a round robin test on the evaluation of impact sounds. Two psychological attributes, loudness and noisiness, were dealt with in this study. The PSE's for both attributes were obtained by using the same experimental procedure.In this study, a repeated impulsive sound whose duration is 3 s at the longest was presented in a diotic listening condition. The results of the study are summarized as follows: 1) The coefficients for linear regressions of PSE on LpE (the frequency-unweighted sound exposure level) were 0.62-0.63 for loudness, while they were 0.93-1.11 for noisiness. This seems to indicate that the time constant for integration of sensation is longer for noisiness than for loudness. 2) For both loudness and noisiness, LpE showed a high correlation with PSE obtained by our experiments, and thus loudness of repeated impulsive sounds might be described by LpE with some modifications, and noisiness can be expressed by LpE with minor modifications. 3) The difference between the results of carrier signals used in this study has little effects on the experimental data.

No abstract