Speech Recognition Using Articulatory and Excitation Source Features