Pruning of redundant synthesis instances based on weighted vector quantization

A new method of pruning redundant synthesis unit instances in a large-scale synthesis database was proposed based on weighted vector quantization (WVQ). WVQ takes relative importance of each instance into account when clustering the similar instances using vector quantization (VQ) technique. The proposed method was compared with two conventional pruning methods through objective and subjective evaluations of synthetic speech quality: one to simply limit maximum number of instance, and the other based on normal VQ-based clustering. For the same reduction rate of instance number, the proposed method showed the best performance. The synthetic speech with reduction rate 50% sounded with no perceptible degradation as compared to that without instance reduction.