Text this: Articulatory-to-Acoustic Conversion Using BiLSTM-CNN Word-Attention-Based Method