Formant Tracking Using Quasi-Closed Phase Forward-Backward Linear Prediction Analysis and Deep Neural Networks

Formant tracking is investigated in this study by using trackers based on dynamic programming (DP) and deep neural nets (DNNs). Using the DP approach, six formant estimation methods were first compared. The include linear prediction (LP) algorithms, weighted LP algorithms recently developed quasi-closed phase forward-backward (QCP-FB) method. QCP-FB gave best performance comparison. Therefore, a novel which combines benefits of learning signal processing QCP-FB, was proposed. In formants predicted DNN-based tracker from speech frame are refined peaks all-pole spectrum computed same frame. Results show that proposed performed better both detection rate error for lowest three compared to reference trackers. Compared popular Wavesurfer, example, reduction 29%, 48% 35% formants, respectively.