This example demonstrates a traditional iVector system evaluated on NIST
 SRE 2016.  It is based on the recipe in ../../sre10/v1/.  In addition to the
 standard features of the SRE10 recipe, it also demonstrates the use of data
 augmentation for PLDA training.

 The recipe uses the following data for system development.  This is in
 addition to the NIST SRE 2016 dataset used for evaluation (see ../README.txt).
 
     Corpus              LDC Catalog No.
     SWBD2 Phase 1       LDC98S75
     SWBD2 Phase 2       LDC99S79
     SWBD2 Phase 3       LDC2002S06
     SWBD Cellular 1     LDC2001S13
     SWBD Cellular 2     LDC2004S07
     SRE2004             LDC2006S44
     SRE2005 Train       LDC2011S01
     SRE2005 Test        LDC2011S04
     SRE2006 Train       LDC2011S09
     SRE2006 Test 1      LDC2011S10
     SRE2006 Test 2      LDC2012S01
     SRE2008 Train       LDC2011S05
     SRE2008 Test        LDC2011S08
     SRE2010 Eval        LDC2017S06
     Mixer 6             LDC2013S03

 The following datasets are used in data augmentation.

     MUSAN               http://www.openslr.org/17
     RIR_NOISES          http://www.openslr.org/28