Speech signal modification to increase intelligibility in noisy environments

Abstract

The role of transient speech components on speech intelligibility was investigated. Speech was decomposed into two components--quasi-steady-state (QSS) and transient--using a set of time-varying filters whose center frequencies and bandwidths were controlled to identify the strongest formant components in speech. The relative energy and intelligibility of the QSS and transient components were compared to original speech. Most of the speech energy was in the QSS component, but this component had low intelligibility. The transient component had much lower energy but was almost as intelligible as the original speech, suggesting that the transient component included speech elements important to speech perception. A modified version of speech was produced by amplifying the transient component and recombining it with the original speech. The intelligibility of the modified speech in background noise was compared to that of the original speech, using a psychoacoustic procedure based on the modified rhyme protocol. Word recognition rates for the modified speech were significantly higher at low signal-to-noise ratios (SNRs), with minimal effect on intelligibility at higher SNRs. These results suggest that amplification of transient information may improve the intelligibility of speech in noise and that this improvement is more effective in severe noise conditions.

Multizone Speech Reinforcement

“…Also, many speech reinforcement schemes are designed taking empirical and/or heuristic considerations into account without a formal quantitative methodology for pre-processing (e.g., [2], [3], [7]- [9]). In contrast, in our approach, we work by quantifying the degradation between source and received speech at the listener by a general functional measure , dubbed as the distortion measure.…”

“…Another observation about current speech reinforcement algorithms is that they either disregard explicit channel information [17], [8], [7], [9], or the dependence is only on noise, not taking any convolutive (reverberant) effects into account [18], [2]- [6]. Although some work has also been done on reverberant channels [19]- [21], the so-called "reverberation pre-processing" algorithms, that work does in turn not take the effect of noise into account.…”

Effectiveness of a loudness model for time-varying sounds in equating the loudness of sentences subjected to different forms of signal processing

“…There has been considerable interest in recent years in the development of methods of processing speech so as to enhance its intelligibility when background noise and/or reverberation are added after the processing has been applied (Yoo et al, 2007;Zorila et al, 2012;Cooke et al, 2013). Such methods have potential applications in public address systems and in classrooms for use with special populations, such as children with "auditory processing disorder" (Moore et al, 2013).…”

The potential of onset enhancement for increased speech intelligibility in auditory prostheses

“…It is even more crucial to focus on these speech parts because the dynamic changes can easily be affected in adverse listening conditions. A number of studies (Hazan and Simpson, 1998;Lorenzi et al, 1999;Apoux et al, 2004;Skowronski and Harris, 2006;Yoo et al, 2007;Rasetshwane et al, 2009) investigated the effect of increasing the CV ratio or amplifying the transient parts of the target speech signal with NH listeners. Amplifying the transient parts lead to an improvement in speech intelligibility in various noisy conditions.…”