i-vector system training data configuration

[s...@oben.com]

Hello,

I am trying speaker recognition using i-vector/PLDA framework,

specifically, now I am making a protocol referring to the NIST SRE2012 database configuration.

(https://github.com/bioidiap/bob.db.nist_sre12)

What I am wondering is how each list file in the "norm" folder is used for training.

Is it something like train_world.lst for UBM and train_world_optional_1.list for total variability space?

Thank you!

Best,

Soo.