Phoneme recognition in modulated maskers by normal-hearing and aided hearing-impaired listeners

Abstract

This study measured the influence of masker fluctuations on phoneme recognition. The first part of the study compared the benefit of masker modulations for consonant and vowel recognition in normal-hearing (NH) listeners. Recognition scores were measured in steady-state and sinusoidally amplitude-modulated noise maskers (100% modulation depth) at several modulation rates and signal-to-noise ratios. Masker modulation rates were 4, 8, 16, and 32 Hz for the consonant recognition task and 2, 4, 12, and 32 Hz for the vowel recognition task. Vowel recognition scores showed more modulation benefit and a more pronounced effect of masker modulation rate than consonant scores. The modulation benefit for word recognition from other studies was found to be more similar to the benefit for vowel recognition than that for consonant recognition. The second part of the study measured the effect of modulation rate on the benefit of masker modulations for vowel recognition in aided hearing-impaired (HI) listeners. HI listeners achieved as much modulation benefit as NH listeners for slower masker modulation rates (2, 4, and 12 Hz), but showed a reduced benefit for the fast masker modulation rate of 32 Hz.

Phoneme recognition in vocoded maskers by normal-hearing and aided hearing-impaired listeners

“…For a typical sloping mild to moderate hearing loss, standardized gain targets in hearing prostheses do not always raise the entire speech spectrum above audiometric thresholds, especially for frequencies above 4 kHz (Humes, 2007;Phatak and Grant, 2012). This hypothesis is further supported by studies that report reduced modulation benefit in NH listeners with simulated hearing loss using a thresholdmatching noise masker (Bacon et al, 1998).…”

“…However, speech perception in such environments is aided by the continuous fluctuations in the level of background noise (Miller and Licklider, 1950;Festen and Plomp, 1990;Bacon et al, 1998;Dubno et al, 2003;Lorenzi et al, 2006;Jin and Nelson, 2006;Bernstein and Grant, 2009;Phatak and Grant, 2012). The benefit due to masker modulations is often thought to be related to audible glimpses of the target speech during the low-intensity portions of the masker envelope (Cooke, 2006), resulting in less masking compared to a steady-state (SS) noise.…”

“…This study was concurrently run with a previously published study (Phatak and Grant, 2012), using the same group of listeners and the same set of nonsense syllables. Phatak and Grant (2012) measured modulation benefit for NH and aided HI listeners using sinusoidally amplitude modulated (SAM) maskers (100% modulation depth, which is typical for SAM maskers) at specific modulation rates, viz., 4, 8, 16, 32 Hz for consonants and 2, 4, 12, 32 Hz for vowels.…”

The near non-existence of “pure” energetic masking release for speech: Extension to spectro-temporal modulation and glimpsing

“…Originally the extrinsic modulations were square-wave in shape (Miller, 1947;Miller and Licklider, 1950;Nelson et al, 2003). More recent work has reported the effects of sinusoidal variation (F€ ullgrabe et al, 2006;Phatak and Grant, 2012). These works point to the general conclusion that a masker with low-rate temporal modulations (2-32 Hz) enables an FMB to be observed in normal-hearing listeners, although the variation with modulation rate depends on the phoneme category; flatter for consonants and more band-pass in shape for vowels (Phatak and Grant, 2012).…”

“…More recent work has reported the effects of sinusoidal variation (F€ ullgrabe et al, 2006;Phatak and Grant, 2012). These works point to the general conclusion that a masker with low-rate temporal modulations (2-32 Hz) enables an FMB to be observed in normal-hearing listeners, although the variation with modulation rate depends on the phoneme category; flatter for consonants and more band-pass in shape for vowels (Phatak and Grant, 2012). Although detailed S-T modulation transfer functions have been reported in humans for (i) maskers, (Chi et al, 1999), or (ii) speech (Elliott and Theunissen, 2009), the information provided does not lead to an ability to predict where in S-T space an FMB might be expected to exist.…”