Synthesized speech quality of Indonesian natural text-to-speech by using HTS and CLUSTERGEN
Elok Cahyaningtyas, Dhany Arifianto · 2016
In this paper, we present the comparison of the synthesized speech of natural Indonesian text-to-speech by using CLUSTERGEN and HTS. The experiment conducted by using 1529 Indonesian speech databases for male (mmht) and female (fena) voice. The process is both contain of training and synthesis part, but in CLUSTERGEN there give a difference method that is trajectory model. After the synthetic voice were synthesized, we evaluate using objective and subjective test to measure the synthesized speech quality. The analysis is done using fundamental frequency track and plot of mel-cepstrum. The result shown that there are not big different in the speech quality of synthesized speech by using CLUSTERGEN and HTS.