Predicting Articulatory Movement from Text Using Deep Architecture with Stacked Bottleneck Features

Using speech or text to predict articulatory movements can have potential benefits for speech related applications. Many approaches have been proposed to solve the acoustic-to-articulatory inversion problem, which is much more than the exploration for predicting articulatory movements from text. In this paper, we investigate the feasibility of using deep neural network (DNN) for articulartory movement prediction from text. We also combine full-context features, state and phone information with stacked bottleneck features which provide wide linguistic context as network input, to improve the performance of articulatory movements' prediction. We show on the MNGU0 data set that our DNN approach achieves a root mean-squared error (RMSE) of 0.7370 mm, the lowest RMSE reported in the literature. We also confirmed the effectiveness of stacked bottleneck features, which could include important contextual information.