Relation between measures of psychoacoustic abilities and masking release for unprocessed, envelope, and temporal fine-structure speech in listeners with normal and impaired hearing

Abstract

Léger et al. [J. Acoust. Soc. Am. 135, 2160, 2014] studied masking release in normal-hearing and hearing-impaired listeners using unprocessed speech and speech processed to convey envelope (ENV) or temporal fine-structure (TFS) cues. Consonant identification scores in speech-shaped background noise (continuous and square-wave interrupted at a rate of 10 Hz) indicated a substantial masking release (MR, better performance in the interrupted noise) for all speech types for normal-hearing listeners but only for the TFS speech in most hearing-impaired listeners. The current study was concerned with relating consonant identification scores and masking release to SRTs for sentences in both types of noise as well as to measurements of basic psychoacoustic abilities. These measures included absolute pure-tone detection thresholds, estimates of auditory bandwidth using a notched-noise procedure, and estimates of cochlear compression obtained using a forward-masking paradigm. For both unprocessed and ENV speech, intelligibility scores and MR were correlated with absolute thresholds and auditory bandwidths in the mid-frequency region (but not at a lower frequency). These correlations were less strong for TFS speech intelligibility and did not predict MR. Cochlear compression was not related to speech intelligibility when controlling for the effect of absolute thresholds. [Work supported by NIH R01 DC000117.]