Prediction of Perceived Sound Quality of Synthetic Speech

This paper investigates the performance of objective speech and audio quality measures for the prediction of perceived sound quality of synthetic speech. A number of existing quality measures have been applied to synthetic speech generated by different speech synthesizers such like LP synthesizer, HSM synthesizer, STRAIGHT synthesizer and several HMM based text-to-speech synthesis systems. The subjective quality rating were obtained using the ITU-T P.85 methodology designed to evaluate the quality of synthetic speech along three dimension: speech naturalness, speech similarity, and overall quality. The correlation of several quality measures with these three subjective rating scales were evaluated among normal subjects. This paper reports the correlations of five objective measures with these three subjective measures and point out the research direction in the future.