why we need both voice design and voice cloning module?

[JiachuanDENG]

Looks like voice design can already got we expect where prompt description control how the voice sound and target text control the content. Looks like it already can give us all we want. Why in the diagram, we still need a voice cloning module?

[hujingbin1]

Hello, because the timbre we generate is not fixed, we need a voice clone model for speech synthesis tasks.

[cacard]

most audio pipline is, 1) generate voice style , 2) use this style as tts ...