does dsa correspond to dsb

[tuanad121]

Thanks for your project.
I realize that we form pair of elements from dsa and dsb for training models.
Because we have parallel data in VC. Is it necessary that the pairs of spectrum from source and target speakers from same sentence?
I think you don't need that in your training. Is it correct?

[marcoppasini]

That's correct! You don't need parallel data to train the system.