Mechanisms underlying the frequency discrimination of pulsed tones and the detection of frequency modulation

Publisher Website
Google Scholar
Abstract
This paper describes two experiments intended to test excitation-pattern models of frequency discrimination. Inexperiment I, frequency DLs (DLFs) for pulsed sinusoids were measured in four conditions: (1) in quiet with the level fixed at 70 dB SPL; (2) in quiet with the level of each tone varied randomly over a 6-dB range around 70 dB SPL; (3) in the presence of a bandpass noise designed to mask the upper side of the excitation pattern, with the level fixed; and (4) with bandpass noise and with the level of the signal randomized over a 6-dB range. Center frequencies of 0.5, 1.0, 2.0, 4.0 and 6.5 kHz were used. DLFs for conditions 3 and 4 were somewhat larger than DLFs for conditions 1 and 2, indicating that frequency discrimination is not based solely on information from the low-frequency side of the excitation pattern. DLFs for condition 4 were, on average, 1.65 times as large as those for condition 1 but, except at 6.5 kHz, ware considerably smaller than the values predicted on the assumption that the DLFs were based on the detection of changes in excitation level on the low-frequency side of the excitation pattern. It is suggested that phase-locking information plays a role in determining DLSs at 4.0 kHz and below. Experiment II was similar to experiment I, except that thresholds (FMDLs) for detecting frequency modulation (FM) at a 4-Hz rate were measured, and in conditions 2 and 4, a 4-Hz amplitude modulation (AM) with a peak-to-valley ratio of 4 dB was imposed on all signals; the phase of the AM relative to the FM was random. The FMDLs were affected more by the noise and by the variation in level than were the DLFs; FMDLs for condition 4 were, on average, 2.36 times as large as those for condition 1. Furthermore, the data for condition 4 were not inconsistent with the assumption that the FMDLs were based on the detectionof changes in excitation level on the low-frequency side of the excitation pattern. however, neither a single-band nor a multiple-band excitation-pattern model was able to account for the effect of the bandpas noise on the FMDLs.

This publication has 15 references indexed in Scilit: