Gravar-mail: Modeling Consonant-Vowel Coarticulation for Articulatory Speech Synthesis