Using Phonetic Constraints to improve inversion
Abstract
The goal of this work is to recover articulatory information from the speech signal by acoustic-to-articulatory inversion. One of the main difficulties with inversion is that the problem is under-constrained and inversion methods generally offer no guarantee on the phonetical realism of the inverse solutions. A way to adress this issue is to use additional constraints: in this work, contraints on phonetically relevant caracteristics of the vocal tracts. Inversion experiments were conducted on two different speakers (male and female), using different implementations of phonetical constraints. The articulatory trajectories were compared with those obtained without using these constraints and to articulatory trajectories measured from X-ray data.