Voice Conversion

Voice Conversion is a speech processing technology designed to transform the vocal characteristics of one speaker into those of another while maintaining the original content of the speech. This technique achieves cross-individual voice transformation by learning the acoustic feature mappings between different speakers, and it has a wide range of application values, such as personalized voice assistants, speech synthesis, and privacy protection. Under the joint training framework of multi-source Tacotron and WaveNet, Voice Conversion can more efficiently achieve high-quality voice conversion effects.