15th Speech in Noise Workshop, 11-12 January 2024, Potsdam, Germany 15th Speech in Noise Workshop, 11-12 January 2024, Potsdam, Germany

P17Session 1 (Thursday 11 January 2024, 15:35-18:00)
The role of periodicity in speech-on-speech understanding in normal-hearing and hearing-impaired listeners

Paolo A. Mesiano, Hamish Innes-Brown
Eriksholm Research Centre, Snekkersten, Denmark

Tobias May
Technical University of Denmark, Lyngby, Denmark

Johannes Zaar
Eriksholm Research Centre, Snekkersten, Denmark

Background: Understanding speech in the presence of one or multiple competing talkers is a challenging auditory task that occurs often in daily life. While normal-hearing (NH) listeners can perform this task successfully, hearing-impaired (HI) individuals encounter severe difficulties in understanding speech in such auditory scenarios. The periodicity information of the competing speech signals, which is connected to the characteristics of their fundamental frequency, can provide useful auditory cues for target-speech understanding, but it is unclear how hearing deficits interfere with the access to such cues. This study investigated how the periodicity information in target and interfering speech contributes to speech intelligibility in young NH and older HI listeners.

Methods: 10 NH and 30 HI listeners participated in a two-competing-voices experiment. The HI group was divided into two subgroups: 15 listeners affected by high-frequency hearing loss (HI1) and 15 listeners affected by both low- and high-frequency hearing loss (HI2). In the experimental stimuli, the periodicity information of target and/or masker signals was either fully available (natural speech) or removed using noise vocoding (vocoded speech). The stimuli were played through two frontal loudspeakers (one for each competing signal). HI listeners were provided with linear-gain amplification.

Results: NH listeners performed best when natural speech was masked by natural speech. Vocoding the target or the masking speech separately did not affect intelligibility significantly, but vocoding both target and masker signals reduced the performance, producing the highest speech reception thresholds (SRTs) overall. Compared to NH listeners, HI listeners showed overall worse performances (with HI2 being worse than HI1 in all experimental conditions) and larger variability across listeners. For both HI1 and HI2 listeners, with natural target speech, vocoding the masker negatively affected speech intelligibility, but not significantly. In contrast, vocoding the target worsened speech intelligibility significantly, with the highest SRTs measured when target and masker signals were both vocoded.

Conclusions: The obtained findings suggest that (i) the severity of (low-frequency) hearing loss is a predictor of speech-on-speech understanding, (ii) NH listeners are challenged only when the periodicity information is removed from both target and masker signals, and (iii) HI listeners rely mostly on the periodicity information in the target speech, while the presence of periodicity information in the masker is useful to them only when no target periodicity information is available. Further research may be directed at exploring potential strategies for enhancing the relevant periodicity information to improve speech intelligibility for HI listeners.

Last modified 2024-01-16 10:49:05