Synthesis of voiced sounds using low-dimensional models of the vocal cords and time-varying subglottal pressure