Using estimated formants tracks for formants smoothing in text to speech (TTS) synthesis | IEEE Conference Publication | IEEE Xplore