Deriving articulatory representations from speech with various excitation modes

A new approach is described which estimates vocal tract shape sequences for speech consisting of voiceless speech and periods of silence as well as voiced speech. This method, based on the use of articulatory codebooks, has proved successful in identifying the place position of stops and fricatives. Secondly, the authors focus on voiced speech in particular. A fast analysis-by-synthesis scheme, which gives continuously-valued area estimates, has been developed. Savings in computation of 50:1 have been achieved by using an MLP to perform the synthesis in this method. The technique also allows a more complex dynamic model to be used.