Synthesis and Analysis-By-Synthesis of Modulated Diplophonic Glottal Area Waveforms

Diplophonia is a type of disordered voice in which two simultaneous pitches are perceived. Most commonly diplophonic voices, the vocal folds divided into parts that vibrate at different frequencies. The glottal area projected space between folds. time referred to as waveform (GAW). GAW modeled for by superimposing partial GAWs (pGAWs) trains single-peak pulses with pulse frequencies, i.e., fundamental frequencies ($f_o$s). In current kinematic models fold vibration, pGAWs assumed be quasiperiodic. This assumption mitigated here modulating pulse-to-pulse cycle length and amplitude. Both random deterministic modulations considered. Deterministic depend on difference pGAWs' instantaneous phases. Model fitted input using an analysis-by-synthesis approach we refer `modulated decomposition' (MPD). MPD shown applicable well nondiplophonic types dysphonia, include multi-pulse patterns, timing behaviours, chaos. It mostly robust against but degraded large modulations. compared deep autoencoder neural network, WaveGlow network. terms time-domain fitting errors, outperforms other approaches unless large. best up approximately 5 dB. For modulations, network achieves smallest errors. magnitude spectrum superior except natural containing only dysphonia. Also errors advantageous MPD.