A synthesizer of diplophonic voices

Voice disorders affect various social categories. Diplophonia is a common phonetic symptom of disordered voices for which a little knowledge is available resulting in its misinterpretation. Diplophonia refers to a voice containing two simultaneous pitches and it is characterized by a periodic signal that has at least two harmonic series having their fundamental frequencies in a rational ratio. In this paper we propose a synthesizer that simulates diplophnic voices. It is a combination of a physical-based synthesizer and a formant synthesizer. The synthesizer includes three principal models; the first model is the smooth phase delayed overlapping sinusoid (S-PDOS) model that simulates the glottal area, the second model is the iterative numerical solution of the Rothenberg equation which simulates the airflow rate and the third model is the vocal tract model which is the cascade configuration of the formant synthesizer of Klatt. The primary tests show the ability of the synthesizer to generate diplophonic voices.