I am using Marytts software (open source Text to Speech software) with voice model "cmu-rms-hsmm en_US male hmm" to generate audio from text. However, the sound feels robotic. Is it possible to remove that robotic effect by changing some aspect of sound and what are the general directions to do it?
There is an online demo portal for Marytts: Marytts link. Please use this portal incase the link for buggy sound expires. You can type any text to test it.
Example buggy sound: sound link