Vowel formant dispersion as a measure of articulation proficiency

Abstract

The articulatory range of a speaker has previously been estimated by the shape formed by first and second formant measurements of produced vowels. In a majority of the currently available metrics, formant frequency measurements are reduced to a single estimate for a condition, which has adverse consequences for subsequent statistical testing. Other metrics provide estimates of size of vowel articulation changes only, and do not provide a method for studying the direction of the change. This paper proposes an alternative approach. Vowel formant frequencies are redefined as vectors originating from a defined center point of the vowel space fixed to a basic three-vowel frame. The Euclidean length of the vectors, the vowel formant dispersion (VFD), can be compared across conditions for evidence of articulatory expansions or reductions across conditions or speaker groups. Further, the angle component of the vowel vectors allows for analyses of direction of the reduction or expansion. Based on the range of investigations afforded by the VFD metric, and simulation experiments that compare its statistical properties with those of other proposed metrics, it is argued that the VFD procedure offers an enhanced view of vowel articulation change over rival metrics.

Vowel Acoustics in Parkinson's Disease and Multiple Sclerosis: Comparison of Clear, Loud, and Slow Speaking Conditions

“…By itself, however, the measure does not capture the impact of rate reduction, clear speech, and increased loudness on the production of individual vowels in F1 × F2 space. Thus, vowel space area measures were supplemented with distance measures for individual vowels computed from a speaker-specific habitual centroid (see also Chung, Edwards, Weismer, Fourakis, & Hwang, 2012; Karlsson & van Doorn, 2012; Turner et al, 1995), as well as a measure reflecting angle components of these distances or vectors (for a similar approach, see Chung et al, 2012; Karlsson & van Doorn, 2012). Each of these measures is considered in the following paragraph.…”

“…A similar measure was reported by Chung et al (2012) to characterize language-related differences in the location of vowels in formant space. Karlsson and van Doorn (2012) also suggested using angle components of Euclidean distance vectors to capture the direction of vowel changes in F1 and F2 space. A schematic illustrating the absolute angle difference measure is provided in Figure 2.…”

Articulation Features of Parkinson’s Disease Patients with Subthalamic Nucleus Deep Brain Stimulation

“…Weismer et al [44] reported that the VSA does not provide straightforward information on movement range, speed, or any other derivative of motion. In contrast, other studies in PD patients reported that the VSA reflected the working space of mouth/jaw and tongue movements [45], [46]. Skodda et al suggested that reduced VSA may be a marker for disease progression in PD [28].…”

Deep Brain Stimulation of the Subthalamic Nucleus Parameter Optimization for Vowel Acoustics and Speech Intelligibility in Parkinson's Disease

“…The degree of vowel centralization was measured using the four-vowel articulation index (VAI; Roy, Nissen, Dromey, & Sapir, 2009;Sapir, Ramig, Spielman, & Fox, 2011), a metric that has successfully distinguished acoustic vowel production in individuals with PD (Sapir et al, 2011). Measures such as the VAI and its inverse, the formant centralization ratio, were developed in order to capture acoustic vowel differences in speakers with dysarthria for whom conventional vowel metrics (such as vowel space area) were not sufficiently sensitive (Karlsson & van Doorn, 2012;Martel-Sauvageau et al, 2014;Martel-Sauvageau, Roy, Cantin, et al, 2015;Roy et al, 2009;Rusz et al, 2013;Sapir, Ramig, Spielman, & Fox, 2010;Skodda et al, 2011). The VAI measures the coefficient of vowel centralization in such a way that minimizes the effects of interspeaker variability (Roy et al, 2009).…”