Interspeech'2005 - Eurospeech
A new method for generating sentence F0 contours of Mandarin speech is proposed. The method assumes the F0 contour generation process model, but generates the tone and phrase components in different ways and sums them to produce a sentence F0 contour. The tone component is generated concatenating F0 patterns of tone nuclei, which are predicted by a corpus-based scheme (binary decision trees). Experiments of F0 contour generation were conducted by using 100 news utterances by a female speaker. The results showed that the method could generate F0 contours close to those of target speech. A perceptual evaluation was also conducted on the synthetic speech using the F0 contours generated by the method. An average score of 4.5 in a 5-point scale indicates the high naturalness, verifying the validity of the method.
Bibliographic reference. Sun, Qinghua / Hirose, Keikichi / Gu, Wentao / Minematsu, Nobuaki (2005): "Generation of fundamental frequency contours for Mandarin speech synthesis based on tone nucleus model", In INTERSPEECH-2005, 3265-3268.