Abstract :
[en] The phonetic implementation of nasal vowels involve much more than velo-pharyngeal port opening. For example, Delvaux (2003) found that the difference between French nasal vowels and their oral counterparts lies not only in the lowering of the velum, but also in a set of articulatory changes on e.g., tongue and lip shape to change the relation between the oral and nasal tracts. These articulatory changes lead to additional differences in the air flow and air pressure between the different nasal vowels, with /??/ having the largest nasal airflow and /??/ the smallest. Delvaux also found midsagittal articulatory differences in the tongue and velum positions between the four subjects investigated. While some of the differences may be explained by intersubject variation in speaker style, it may be hypothesized that others are due to anatomy, i.e., that different speakers need to use different articulations in order to achieve similar acoustic targets. Since speech motor control mechanisms are highly adaptive and compensatory (Lindblom, 1996), it is plausible that subjects make articulatory adjustments to compensate for anatomical differences. This study is a first attempt to test this hypothesis. The articulatory analysis made by Delvaux (2003) is extended in several aspects, since the same four speakers are analyzed, but investigating the 3D vocal tract shape rather than the midsagittal; the cross-sectional areas of the velum port opening rather than the midsagittal contour; and by adding the analysis of the subjects'3D nasal tract shape and volume.Results show (i) that the four subjects differ anatomically in the size of their oral and nasal tracts, both in the absolute volume and in the relation between the two; (ii) that the subjects use different articulatory strategies to contrast the oral and nasal vowels.Results also suggest that (i) and (ii) are related, i.e. that the subjects' anatomy in the vocal and nasal tracts influences the articulation. The subject with the smallest nasal cavity, F1, makes larger changes than the other subjects in both the oral tract articulation and the velum port opening. The subjects with relatively larger nasal tracts achieve the contrast either by large differences in the velum port opening, while keeping the oral articulation similar (subject F2), or with small changes at the velum, but compensating for this by changes in the oral articulation to alter the oral tract volume and area function (subject M2). An intermediate option is to combine smaller changes in velum port opening and oral articulation (subject M1). As this study investigates four subjects only, it will have to be extended in order to show that the differences in articulation are a direct consequence of the differences in anatomy. But this study nevertheless indicates that such a relation may exist and that it could be an important factor for intersubject variations
