Discrimination of the fundamental frequency of complex tones with fixed and shifting spectral envelopes by normally hearing and hearing-impaired subjects

Abstract

Difference limens for the fundamental frequency (F0) of complex tones (DLCs) were measured for four normally hearing subjects and three subjects with cochlear hearing loss. The nominal F0 was 100, 200 or 400 Hz. The two tones to be compared contained either low resolved harmonics (RES), harmonics with intermediate resolvability (INT) or high unresolved harmonics (UNRES). For one set of stimuli (fixed harmonics, FH), the tones to be compared contained three harmonics of fixed number, so the changes in F0 were associated with spectral cues. For a second set of stimuli (Shaped), spectral cues were minimized by filtering stimuli through a fixed passband. For the INT and UNRES conditions, the excitation patterns evoked by the Shaped stimuli hardly changed when F0 was altered. To prevent subjects from comparing the frequencies of individual harmonics in the RES condition, subjects were required to detect F0 differences between two tones with non-overlapping harmonics. It was not possible to obtain repeatable results for the hearing-impaired subjects in this condition. The normally hearing subjects had smaller DLCs for the FH than for the Shaped stimuli for the RES condition, and the UNRES condition at the two higher F0s. However, DLCs were similar for the FH and Shaped stimuli for the INT condition and the UNRES condition at the 100-Hz F0, suggesting that spectral cues were not used in these conditions. Except for one subject with F0=400 Hz, the hearing-impaired subjects had smaller DLCs for the FH than for the Shaped stimuli, for both INT and UNRES conditions (although the difference was small for F0=200 Hz in the UNRES condition), suggesting that they used spectral cues for the FH stimuli. For the Shaped stimuli, DLCs were similar in the INT and UNRES conditions for the hearing-impaired subjects, but were smaller in the INT than the RES condition for the normally hearing subjects. We suggest that, in the INT condition with Shaped stimuli, normally hearing subjects used temporal fine structure cues to perform the task. The hearing-impaired subjects appeared to use only temporal envelope cues.