Interstimulus interval dependence of the loudness difference limen obtained by taking into account the presentation order effect

Abstract

This paper describes the interstimulus interval dependence of the loudness difference limen obtained by taking into account the presentation order effect of sound stimuli. A paired comparison experiment of pure-tone sounds was carried out at different interstimulus intervals, and the experimental data were analyzed, taking into account the presentation order effect. The following two characteristic effects were obtained. First, the difference limen changed depending on the interstimulus interval. A logarithmic relation was found between the difference limen and the interstimulus interval. The difference limen was about 0.6 to 1.6 dB at an interval of 0.5 to 64 s. Second, the order effect also changed depending on the interstimulus interval. The sound presented first was perceived as being louder than the sound presented second at an interval of 0.5 to 4 s, and the sound presented second was perceived as being louder than the sound presented first at an interval of 16 to 64 s, although the sounds were of the same sound pressure level. When the interval was 8 s, the sound presented first was perceived as being as loud as the sound presented second. Based on the above findings, we estimated the region where the difference in loudness could not be detected. The obtained results show that the region is not symmetrical for the upper and lower boundary levels. BACKGROUNDSubjective evaluation tests of loudness are frequently performed not only in psychoacoustics studies but also in various product development tests. However, if the accuracy of a subjective evaluation test is low, the reliability of the data will be compromised, and the results will be inaccurate. It is, therefore, very important that evaluation tests be carried out accurately. To do that, we must find a suitable method and environment. It is widely recognized that the constant method [1][2][3][4], in which sound stimuli are presented randomly and there is a minimal load on the subjects, is one of the most accurate methods. The constant method is used for paired comparisons and it has two important restrictions: do not change any parameters except for the evaluation target, and do not spend a lot of time for the comparison interval to do not forget the first sound. However, changing the target parameters for a very complex or big evaluation target is still a very timeconsuming process. In this situation, we need to know how long the subject can retain information about the sound presented in the experiment, and how accurately subjects can detect changes in the sound.A number of studies on the loudness difference limen have been carried out. For example, Riesz (1928) undertook basic research on the difference limen by using modulation noises [5]. Miller (1947) investigated the difference limen for white noises [6]. Jesteadt (1977) showed the frequency and loudness dependence of the difference limen for pure tones [7]. Clement et al. (1999) studied the relationship between the difference limen and the interstimulus interval [8]. However, few eva...

Previous sound effects on loudness in paired comparison experiments

“…In this study, the experimental conditions were almost the same as those for a standard experimental method regarding the loudness difference limen [19]. On the basis of this, the presentation order effect was calculated according to the calculation method for difference limens taking into account the presentation order effect [19][20][21][22].…”

“…In this study, the experimental conditions were almost the same as those for a standard experimental method regarding the loudness difference limen [19]. On the basis of this, the presentation order effect was calculated according to the calculation method for difference limens taking into account the presentation order effect [19][20][21][22]. The order effect was calculated by fitting Gaussian distributions to the response probabilities of the comparison evaluations at each ISI and each presentation type (the second sound stimulus was presented in the contralateral or ipsilateral ear).…”

“…This experimental procedure was almost the same as the method of evaluating the loudness difference limen [19]. The first of a comparison pair of sounds was presented in one ear and the second was presented in the contralateral or ipsilateral ear.…”

“…The interval of SPLs was set at 1 dB from 77 to 83 dB SPLs to obtain a distribution of the evaluation. The stimuli at 74 and 86 dB SPLs, where most people can detect the difference in loudness compared with that of the base sound (80 dB SPL) [19], were used to confirm the hearing acuity of the subjects. The ISIs between the first and second sounds were set at 0.5, 1, 2, 4, and 8 s. All pairs of the base and comparison sound stimuli were recorded on digital audiotapes at a sampling frequency of 44.1 kHz and presented to the subjects monaurally via headphones (Sony closed-type headphones, MDR-Z600).…”