Multi-level Prosody And Short-term Spectrum Transform For Emotional Speech Synthesis

In daily life, the speech not only contains the meaning of the text content, but also delivers some emotional information. For the same sentences, if the style that speaker expresses is different, the information listener get will also be different. Emotional speech conversion, that is, under the condition of the same text, realizing voice conversion between different emotions. Therefore, the research of emotion transformation plays a significant role in expressive speech synthesis.In order to synthesize high quality of emotional speech, this paper use a multi-level emotional prosody and short-time spectrum transform for emotional speech synthesis method. By using the method of multi-level, we build a corresponding prosodic model for happiness, anger, sadness and neutral speech. Based on it, we realize the prosody conversion after training the mapping relationship between neutral and emotional speech, and then complete emotional transformation. In the end, combined with the short-term spectrum transformation, we use STRAIGHT to synthesize obvious emotional speech.In this paper, the subjective evaluation method of MOS and ABX is used to test the converted emotional speech and the results show that the proposed method can improve the transformation result. At the same time, compared with traditional method, the result of spectral distortion test for synthesized emotional speech, shows that the, in this paper, the conversion performances for happiness, anger, and sadness increased by 2%, 4% and 6%,respectively.