Formant Tracking

Formant tracking aims to accurately identify and follow the changing frequencies of vocal tract resonances (formants) in speech signals, crucial for speech analysis and synthesis. Current research focuses on improving accuracy and robustness through hybrid approaches combining model-driven techniques like linear prediction (including variations such as quasi-closed-phase analysis) with data-driven methods such as deep neural networks. These advancements lead to more reliable formant estimations across diverse speech conditions, impacting applications in speech recognition, synthesis, and clinical diagnostics of speech disorders.