The effects of leading and trailing tones (interference tones) on frequency discrimination of a brief tone are investigated. Changes in frequency discrimination are measured as a function of frequency difference between the interference and signal tone, and interval between the two tones (Δt). The results obtained are as follows : (1) The frequency discrimination for a tone with the leading tone is almost unchanged, even if Δt is increased. On the other hand, frequency DL for a tone with the trailing tone increases gradually as Δt decreases. (2) In the case where the frequency of the interference tone is changed as a parameter, frequency DL takes its minimum value when the interference tone and the test tone has the same frequency. Frequency discrimination becomes worse, however, in the presence of interference tones with the frequency slightly different from the test tone. The effect of the interference tone is maximum for the frequency difference of 50 Hz. In theses experiments it is found that the interference effect of trailing tone on frequency discrimination is greater than that of the leading tone. Finally a model of the mechanism for frequency analysis in hearing was proposed and it can explain the results fairly well.
The question, to which extend the ear is able to discriminate and follow random temporal changes of sound pressure level (SPL) is studied. For this purpose experiments are designed to determine the dynamic difference limen (DDL) for intensity, that is, intensity discrimination for sounds whose SPLs change with time. The psychophysical methods which have been used commonly in previous studies cannot be applied to measure DDL. The first problem of the present study, therefore, is to develop an appropriate method for DDL measurement. Three experiments are performed and larger values of DDLs are obtained than DLs for steady-state stimulus.
The detectability of clicks perceived at onset or offset of a brief sinusoidal signal is measured as a function of the Sensation Level (SL) of the tone ranging from 30 to 70 dB SL. The tone frequencies are 400 and 2000 Hz, while the two types of amplitude envelopes (linear and exponential) are employed. The results show that the critical rise or decay time (tc_1, tc_2) required to achieve clickless signals depend on the level, the onset of offset slope and the signal frequency, while does not depend on the initial phase of the tone. As the SL of the tone is decreased, it is necessary to decrease the onset or offset slope in order to achieve clickless signals. The critical decay time is about 1/3 times as small as the critical rise time. This means that the transient clicks are less audible at offset than at onset of the tone. It is shown that the average data can be well fitted with the equation (I/I_0)・tc^a_jj=constant, where I and I_0 are the tone intensity and the threshold intensity respectively, and a _j is a consonant depending on the signal frequency, while tc_j is the critical rise (j=1) or the decay (j=2) time. A running average model is shown to be effective to account for the above results.
In order to investigate physical and psychological factors governing timbre of complex tones, three-frequency complex tones are synthesized on a digital computer. Subjects make similarity judgments of pairs of the complex tones and the experimental data are analyzed by Kruskal's multidimensional scaling program. The analysis of the experimental data shows the followings : (1) the absolute values of the lowest and the highest frequencies are the most important factors governing timbre, (2) from a different point of view, the mean of the logarithmic values of the lowest and the highest frequencies are also important, and these values correspond to tone height (kan-dakasa) of complex tones, and the mean of adjacent frequency ratios (or absolute dissonance) is another important physical factor governing timbre and this physical factor is closely related to consonance of complex tones.