search

ALIZE-Speaker-Recognition / LIA_RAL

A high-level toolkit for speaker recognition, build on top of ALIZE-Core.
http://alize.univ-avignon.fr
GNU Lesser General Public License v3.0
125 stars 27 forks source link

What the normalization did? #32

Closed 9crk closed 5 years ago

9crk commented 5 years ago
  1. do feature extraction. 2.do normFeat 3.do EnergyDetection 4.do normFeat 5.do TrainWorld

that's the question: what's the step 2&4 did? it does normalization in a single speaker feature data or on all speaker's feature data ?

9crk commented 5 years ago

tried put in one speaker vs. two speaker. the result was the same. so it's doing normalization in a single feature file.