Examples of HMM-based speech synthesis using a DSM of the residual signal
For a detailed explanation of the technique used for these examples, please refer to our Interspeech09 paper.
The use of headphones is recommended for a better appreciation.
Speaker Bruno (French Male)
File 1: Pulse
File 1: Proposed
File 2: Pulse
File 2: Proposed
Speaker Julie (French Female)
File 1: Pulse
File 1: Proposed
File 2: Pulse
File 2: Proposed
Speaker SLT (US Female)
File 1: Pulse
File 1: Proposed
File 2: Pulse
File 2: Proposed
Speaker AWB (Scottish Male)
File 1: Pulse
File 1: Proposed
File 2: Pulse
File 2: Proposed