This study investigates the relationship between articulatory and acoustic variability in Latin American Spanish vowel production. While Whalen et al. (2018) found that American English vowels show similar degrees of variability across domains and a correlation between them, it remains unclear whether this pattern generalizes across languages. We analyzed audio, tongue ultrasound, and lip video data from 11 native speakers reading a word list. Tongue and lip contours were semi-automatically tracked using DeepLabCut Mathis, et al. 2018 in AAA Articulate Instruments Ltd. 2012, and vowel formants were extracted at midpoint. For each word, mean formant values and tongue and lip contours were computed, and variability across repetitions was quantified using the coefficient of variation (CV). Preliminary results from eight speakers showed no statistically significant correlation between tongue and formant variability (Pearson’s r = −0.64, p = 0.087). A linear mixed-effects model revealed marginally higher CV in articulation than in acoustics (p = 0.071), with no significant vowel-specific effects. The weaker correlation compared to English suggests that the correspondence between articulatory and acoustic variability may differ cross-linguistically, highlighting the need to expand research to languages with diverse sound systems. This work contributes to a broader understanding of cross-linguistic variability patterns in speech production.
Kandler et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: