You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Parameters for mixed tuning, selecting the sources you want to mix. Must be on the singer list and feature description
speaker_id
string
No
Effective when mix_info is not set. Single source of synthesis, refer to the list of singers and feature description. Default is "1".
extra
string
否
Users can input any string, which will be filled into the returned data structure
Note: Each request file must have a synthesis duration shorter than 18 seconds, and the number of uploaded files should not exceed 4.
Voice Blending Feature Explanation
Through the interface, you can achieve the desired voice by more nuanced tone blending. You can control the different blending ratios based on the following 7 dimensions:
duration mainly controls the articulation
pitch mainly controls the singing style, and it also affects some aspects of articulation
air controls the amount of breathiness in singing
falsetto controls the amount of falsetto in singing
tension controls the tension/relaxation of the vocal cords during singing
energy controls the intensity of singing
mel primarily controls the basic timbre
To understand the tone blending more intuitively, here's an example: currently, there are two singers with IDs 1 and 82. If we want to sound 30% like singer 1 and 70% like singer 82 in the above dimensions, then we can create a mix_info with the following JSON string to achieve this function: