A frequency-weighted post-filtering transform for compensation of the over-smoothing effect in HMM-based speech synthesis.

Florian Eyben, Yannis Agiomyrgiannakis

Browse the full ICASSP paper archive.