Phonemic restoration of interrupted locally time-reversed speech

Kazuo Ueda,Valter Ciocca

doi:10.3758/s13414-021-02292-3

Kazuo Ueda, Valter Ciocca

Open Access

https://doi.org/10.3758/s13414-021-02292-3

Copy DOI

Abstract

Intelligibility of temporally degraded speech was investigated with locally time-reversed speech (LTR) and its interrupted version (ILTR). Control stimuli comprising interrupted speech (I) were also included. Speech stimuli consisted of 200 Japanese meaningful sentences. In interrupted stimuli, speech segments were alternated with either silent gaps or pink noise bursts. The noise bursts had a level of − 10, 0 or + 10 dB relative to the speech level. Segment duration varied from 20 to 160 ms for ILTR sentences, but was fixed at 160 ms for I sentences. At segment durations between 40 and 80 ms, severe reductions in intelligibility were observed for ILTR sentences, compared with LTR sentences. A substantial improvement in intelligibility (30–33%) was observed when 40-ms silent gaps in ILTR were replaced with 0- and + 10-dB noise. Noise with a level of − 10 dB had no effect on the intelligibility. These findings show that the combined effects of interruptions and temporal reversal of speech segments on intelligibility are greater than the sum of each individual effect. The results also support the idea that illusory continuity induced by high-level noise bursts improves the intelligibility of ILTR and I sentences.

Highlights

With segment duration is longer than 40 ms and up to 70 ms, segmental information is degraded but the intelligibility of locally time-reversed speech (LTR) speech remained relatively high likely because envelope features such as manner, segmental/syllable duration, and speech rhythm can still be retrieved by temporal integration over long time windows
By replacing silent gaps with pink noise samples, this study investigated the role of linguistic context by comparing interrupted sentences with and without noise bursts so that in both conditions the available speech signal was not masked by a simultaneous noise burst
A reduction in intelligibility was observed for ILTR sentences, compared to LTR sentences, with segment durations of 40 to 80 ms

Summary

Introduction

In describing the temporal characteristic of speech signals, Rosen (1992) proposed a framework according to which the temporal structure of speech is organized into threePortions of this work were presented in “Perceptual restoration of interrupted locally time-reversed speech: Effects of noise levels,” the 83rd Annual Convention of the Japanese Psychological Association, Ibaraki, Japan, September 2019, and “Perceptual restoration of interrupted locally time-reversed speech: Effects of noise levels and segment duration” Fechner Day 2019: the 35th Annual Meeting of the International Society for Psychophysics, Antalya, Turkey, October 2019.Atten Percept Psychophys (2021) 83:1928–1934A useful paradigm for testing the predictions of the 2TWI model for speech processing is the so-called “locally timereversed” (LTR) speech (Steffen & Werani, 1994; Saberi & Perrott, 1999; Greenberg & Arai, 2004; Kiss, Cristescu, Fink, & Wittmann, 2008; Stilp, Kiefte, Alexander, & Kluender, 2010; Remez et al, 2013; Ishida, Samuel, & Arai, 2016; Ueda, Nakajima, Ellermeier, & Kattner, 2017; Ishida, Arai, & Kashino, 2018; Nakajima, Matsuda, Ueda, & Remijn, 2018; Teng, Cogan, & Poeppel, 2019). Ueda, Nakajima, Ellermeier, and Kattner (2017) replicated this finding with LTR sentences in four languages (English, German, Japanese, Mandarin) when speech rates were normalized. They reported that intelligibility was at ceiling for all languages with 20-ms segments, remained high (about 90% correct or higher) when segment duration was shorter than 45 ms, and declined to about 40%50% as segment duration was increased further to 70 ms. With segment duration is longer than 40 ms and up to 70 ms, segmental information is degraded but the intelligibility of LTR speech remained relatively high likely because envelope features such as manner, segmental/syllable duration, and speech rhythm can still be retrieved by temporal integration over long time windows

Methods

Results

Conclusion