Speech enhancement for reverberation-robust TTS

Experiments with reverberant data

Enhancement methods:
OMLSA: baseline speech enhancement method
Postfish: baseline speech enhancement method
RNN-DFT-N: RNN enhancement of DFT based parameters trained with noisy data
RNN-DFT-R: RNN enhancement of DFT based parameters trained with reverb data

Natural and vocoded speech samples:

	FEMALE				MALE
	Small room		Medium size room		Small room		Medium size room
	Natural	Vocoded	Natural	Vocoded	Natural	Vocoded	Natural	Vocoded
CLEAN
REVERB
OMLSA
POSTFISH
RNN-DFT-N
RNN-DFT-R

HMM-based synthetic speech samples:

	FEMALE				MALE
	Sample 1	Sample 2	Sample 3	Sample 4	Sample 1	Sample 2	Sample 3	Sample 4
CLEAN
REVERB
OMLSA
POSTFISH
RNN-DFT-N
RNN-DFT-R