Timing and copying

Still working through stuff …

Two things,

1 Solaria and Kevin start singing at slightly different times. Why??

2 When I try to copy, instead of a straight copy, all the lines go haywire. Surely a copy should be a copy, not to have to start editing again? Why??


1 Like

Each AI voice database is based on machine learning analysis of a real human singer, so they inherit articulation and pitch behaviors similar to the original voice provider.

In this case the slight timing differences are likely a result of Kevin and Solaria’s voice providers articulating the lyrics differently when they sing, and the synthesis engine attempting to replicate that.

Pitch differences are a result of Sing mode re-generating the curves based on the original voice provider’s singing style, such as their tendency to under/overshoot notes, where they use vibrato, etc.

Timing adjustments have to be done on a case-by-case basis, but you can avoid the software re-generating the pitch curve by switching the notes to Manual mode before copying. Make sure to also right click on the selection and use “Select Parameters for Notes” (ctrl+alt+A) to keep all the parameter data when copying.


Thanks …l again.

Didn’t mean to put the ‘I’
Can’t see how to edit previous posts.