Directly modeling voiced and unvoiced components in speech waveforms by neural networks.

Keiichi Tokuda, Heiga Zen

Browse the full ICASSP paper archive.