The role of periodicity in perceiving speech in quiet and in background noise

Abstract

The ability of normal-hearing listeners to perceive sentences in quiet and in background noise was investigated in a variety of conditions mixing the presence and absence of periodicity (i.e., voicing) in both target and masker. Experiment 1 showed that in quiet, aperiodic noise-vocoded speech and speech with a natural amount of periodicity were equally intelligible, while fully periodic speech was much harder to understand. In Experiments 2 and 3, speech reception thresholds for these targets were measured in the presence of four different maskers: speech-shaped noise, harmonic complexes with a dynamically varying F0 contour, and 10 Hz amplitude-modulated versions of both. For experiment 2, results of experiment 1 were used to identify conditions with equal intelligibility in quiet, while in experiment 3 target intelligibility in quiet was near ceiling. In the presence of a masker, periodicity in the target speech mattered little, but listeners strongly benefited from periodicity in the masker. Substantial fluctuating-masker benefits required the target speech to be almost perfectly intelligible in quiet. In summary, results suggest that the ability to exploit periodicity cues may be an even more important factor when attempting to understand speech embedded in noise than the ability to benefit from masker fluctuations.

Effects of acoustic periodicity, intelligibility, and pre-stimulus alpha power on the event-related potentials in response to speech

“…noisevocoded, henceforth referred to as the aperiodic condition), preserves the natural mix of voiced and voicelessness (henceforth the mixed condition; Dudley, 1939), or is completely voiced (henceforth the periodic condition). Previous behavioural work (Steinmetzger & Rosen, 2015) has shown that the intelligibility of the aperiodic and mixed conditions is very similar, while the unnatural-sounding fully periodic condition was found to be considerably less intelligible. In order to analyse effects of acoustic periodicity while controlling for differences in intelligibility, the individual trials in the current study were sorted according to the listeners' spoken responses (i.e.…”

Effects of acoustic periodicity and intelligibility on the neural oscillations in response to speech

“…At the same time, using noise as speech source results in a loss of the natural mix of voiced and voicelessness, and consequently also any voice pitch information, making it resemble whispered speech. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 4 Nevertheless, our previous behavioural work (Steinmetzger & Rosen, 2015) has shown that preserving periodicity information in a vocoder (henceforth the mixed condition) does not lead to improved intelligibility rates. This suggests that periodicity information, despite its salience, is a redundant cue, at least in non-tonal languages and quiet listening conditions.…”

“…Consequently, periodic speech has substantially lower intelligibility rates than the other two conditions (Ardoint, et al, 2014;Steinmetzger & Rosen, 2015).…”

The role of envelope periodicity in the perception of masked speech with simulated and real cochlear implants

“…Yet, none of the studies mentioned so far measured speech intelligibility in CI users and CI simulations with non-speech maskers specifically designed to vary regarding the presence or absence of F0 cues, which would enable a more direct investigation of the role of temporal periodicity. The current study seeks to do so by re-using materials introduced in Steinmetzger and Rosen (2015), where it was investigated whether periodicity cues in both target speech and masker affect the ability of normal-hearing listeners to understand spoken sentences. Specifically, periodic maskers based on harmonic complex tones with dynamically varying F0 contours derived from real speech were contrasted with aperiodic speech-shaped noise maskers.…”

“…Figs. 5 & 6 in Steinmetzger and Rosen, 2015). However, CI simulation studies have usually found hardly any benefit from masker envelope fluctuations (Cullington and Zeng, 2008; P. B.…”