A Comprehensive Vowel Space for Whispered Speech

Abstract

Whispered speech is a relatively common form of communications, used primarily to selectively exclude or include potential listeners from hearing a spoken message. Despite the everyday nature of whispering, and its undoubted usefulness in vocal communications, whispers have received relatively little research effort to date, apart from some studies analysing the main whispered vowels and some quite general estimations of whispered speech characteristics. In particular, a classic vowel space determination has been lacking for whispers. For voiced speech, this type of information has played an important role in the development and testing of recognition and processing theories over the past few decades, and can be expected to be equally useful for whisper-mode communications and recognition systems.This paper aims to redress the shortfall by presenting a vowel formant space for whispered speech, and comparing the results with corresponding phonated samples. In addition, since the study was conducted using speakers from Birmingham, the analysis extends to discuss the effect of the common British West Midlands (WM) accent in comparison with Standard English (RP). Thus, the paper presents the analysis of formant data showing differences between normal and whispered speech while also considering an accentual effect on whispered speech.

Finding the Most Uniform Changes in Vowel Polygon Caused by Psychological Stress

“…Recently, vowel polygons, mainly called as vowel spaces, were used in other fields of speech processing, e.g. achieving children age differences [25], whisper analysis [26] and observation of the Parkinson disease [27], but not applied on stressed speech. The presented method can be also possibly useful to active hypoxia level detection [28].…”

Comparative whisper vowel space for Singapore English and British English accents

“…The acoustic analysis including details of the recording, speakers, equipment and measurement methods, are described in Section 2, while Section 3 outlines the results separately for men and women. Section 3 also provides a discussion on findings including the comparison of the shift amounts obtained from the current study in SgE with the corresponding results in British WM (presented in our previous paper [17]) for each vowel in whispers and normal speech; finally, Section 4 concludes the paper. …”

“…Our previously published work [17] tries to present an acoustic vowel space determination (a classic F 2xF 1 plane) for this purpose along with the shift amounts between normal and whispers for each vowel through the experiments conducted on British West Midlands (WM) accent speakers; however, the extent of generalisation of vowel shift amounts between two spaces is not yet generalised across other English accents. The current paper analyses Singapore English (SgE) to firstly present the whisper vowel diagrams for SgE, and secondly, extends the discussion to assess the generalisation of the shift amounts between the two vowel spaces for dissimilar English accents.…”

Variants of mel-frequency cepstral coefficients for improved whispered speech speaker verification in mismatched conditions

“…i) whispered speech has a lower and flatter power spectral density [3] and ii) the formants shift towards higher frequencies [2], [16]. This last observation is more noticeable for the first three formants (F1, F2 and F3), where, F1 shifts can be up to 71% for men and 52% for women; F2 shifts can be up to 24% for men and 20% for women; and F3 shifts can be of 10% and 4.8%, respectively [16]. To illustrate this, Figure 1 depicts the average power spectrum of amplitude-normalized and preemphasized recordings from 36 speakers (male and female).…”