for modelling vocal tract parameters