Interactions among talker sex, masker number, and masker intelligibility in speech-on-speech recognition

Abstract

In competing speech, recognition of target speech may be limited by the number and characteristics of maskers, which produce energetic, envelope, and/or informational masking. In this study, speech recognition thresholds (SRTs) were measured with one, two, or four maskers. The target and masker sex was the same or different, and SRTs were measured with time-forward or time-reversed maskers. SRTs were significantly affected by target-masker sex differences with time-forward maskers, but not with time-reversed maskers. The multi-masker penalty was much greater with time-reversed maskers than with time-forward maskers when there were more than two talkers.

Susceptibility to Steady Noise Largely Explains Susceptibility to Dynamic Maskers in Cochlear Implant Users, but not in Normal-Hearing Listeners

“…Mean NH SRTs were 9.9 dB poorer with the FOR than with the REV masker, and 11.2 dB poorer with the FOR-FOR than with the REV-REV masker, suggesting a similar effect of lexical interference on SRTs for both 1-talker and 2-talker speech maskers. The data are generally consistent with Englishspeaking NH listeners in Thomas et al (2021), who reported that mean SRTs were 8.8 dB poorer with 1-talker forward than reversed speech maskers, and 11.8 dB poorer with 2-talker forward than reversed speech maskers. Interestingly, the slope for the maskers with both F0 and lexical cues (FOR, FOR-REV, FOR-FOR; −1.22 dB per % glimpse proportion) was nearly twice as steep as that for the maskers with no F0 or lexical cues (SSN, 4HZ, ENV, BAB; −0.59 dB per % glimpse proportion).…”

“…The effectiveness of speech maskers may also depend on the number of maskers and the vocal characteristics of the maskers ( Chen et al, 2020 ; Cullington & Zeng, 2008 ; Thomas et al, 2021 ). For example, NH listeners may experience considerable masking release (compared to SSN) with a 1-talker masker due to dip-listening.…”

“…NH listeners are highly susceptible to an increasing number of speech maskers. Previous studies have shown that NH listeners' speech reception thresholds (SRTS) can worsen by approximately 10 dB as the number of speech maskers increases from 1 to 2, beyond which SRTs do not change or even slightly improve ( Chen et al, 2020 ; Cullington & Zeng 2008 ; Thomas et al, 2021 ). Contrary to NH listeners, no significant effect of the number of masker talkers has been observed on CI users' SRTs ( Cullington & Zeng 2008 ).…”

“…However, only one speech masker was used in that study (1-talker speech with a different sex from that of the target). Previous studies have shown that SRTs may depend on the acoustic and linguistic characteristics in both NH listeners and CI users, and the susceptibility to these speech maskers varies differently between NH listeners and CI users (e.g., Chen et al, 2020 ; Cullington & Zeng 2008 ; Fu & Nogaki, 2005 ; Nelson et al, 2003 ; Stickney et al, 2004 ; Tao et al, 2018 ; Thomas et al, 2021 ). Due to differences in test materials, it can be difficult to compare masker effectiveness of various masker types across studies.…”

Cochlear Implant Facilitates the Use of Talker Sex and Spatial Cues to Segregate Competing Speech in Unilaterally Deaf Listeners

“…In this study, informational masking refers to lexical interference (i.e., both the target and masker speech are intelligible) and talker sex characteristics. Previous studies have shown that recognition of target speech is more difficult with two competing talkers than with one competing talker, presumably due to informational masking (e.g., Cullington and Zeng 2008;Calandruccio et al 2017;Chen et al 2020;Thomas et al 2021). Informational masking tends to be largest when the target and masker talkers are colocated and perceptually similar (e.g., when target and masker talkers are the same sex).…”