The Effect of Speaking Rate onConsonant Vowel Coarticulation

Abstract

In 2007 Lindblom et al. introduced a methodological tool to disentangle consonant-vowel (CV) coarticulation attributable to emphatic stress apart from the vowel expansion effects known to accompany the prosodic overlay. After empirically accounting for the altered vowel positions, they reported small but consistent increases in F2 transition onsets in emphatically produced CVs that could not be attributed to vowel context influences, and that differed across stop place. At issue is whether the findings of these authors can be replicated, but in the opposite direction, for CVs produced at fast speaking rates. A modified locus equation regression metric was similarly used to account for rate-induced vowel reduction effects in predicting frequencies of F2 transition onsets in rapid speech. Six American-English speakers produced [V1.CV2] sequences embedded in a carrier sentence, at three speaking tempos: normal, fast, and fastest. Significant differences were found between ‘predicted’ and ‘observed’ F2 onsets across stops, with alveolars and velars showing greater decreases in F2 onsets during more rapid speech than labials. The complementary findings are discussed relative to a unified view of anticipatory coarticulation in CV production across a continuum of hyperarticulated spectral expansion to hypoarticulated spectral reduction.

Effects of the Rate of Formant-Frequency Variation on the Grouping of Formants in Speech Perception

“…Increasing or decreasing the rate at which speech is produced-which talkers do frequently in normal discourse for a variety of linguistic and paralinguistic reasons-has complex articulatory (and therefore acoustic) consequences in addition to the manifest effect of decreasing or increasing overall utterance duration. Increases in speech rate typically involve, among other things: (1) increased co-articulation, reduction, and assimilation (Gay 1981;Agwuele et al 2008); (2) reductions in segment duration that are dependent on stress placement and phoneme class (Lehiste 1970;Byrd and Tan 1996); and (3) increases in articulator 2 There is no distinction between reversed and constant amplitude when the competitor rate is set to zero. Hence, the data for C2 served as the zero-rate case for both amplitude-contour conditions in the two-factor ANOVA.…”

Speaking Rate Effects on Normal Aspects of Articulation: Outcomes and Issues

“…Despite the preceding discussion, which suggests that acoustic effects of rate change may be small (relative to kinematic variability), studies of the acoustic effects of speaking rate change have identified a variety of speaker and context-related effects on formants (Agwuele, Sussman, & Lindblom, 2008;Fourakis, 1991;Gay, 1968Gay, , 1978Hertrich & Ackermann, 1995;Lindblom, 1963;Pitermann, 2000;Rosen et al, 2011;Tjaden & Weismer, 1998;van Son & Pols, 1992;Weismer & Berry, 2003). Taken together, these studies indicate a wide range of possible speaker-specific and context-related effects of speaking rate change.…”

A Duration-Dependent Account of Coarticulation for Hyper- and Hypoarticulation

“…In total there were 270 tokens analyzed per speaker in the rate study. Further details about the specific stimuli and conditions can be found in Lindblom et al [2007] and Agwuele et al [2008].…”

“…Specifics of this methodology have been provided in detail in both Lindblom et al [2007] and Agwuele et al [2008], but will be briefly summarized below. Following this, a new figure, displaying both emphasis and rate data in the same plot, will be presented.…”

“…The purpose of this research was to revisit two databases that investigated anticipatory coarticulation from opposite ends of the hyperarticulation-to-hypoarticulation continuum: emphatic stress and reduced speech brought about via fast speaking rates [Lindblom et al, 2007;Agwuele et al, 2008]. The departure point in both studies was the realization that coarticulation has several causal factors that interact to determine its overall extent.…”