[Subject Prev][Subject Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:02148] HInit in Trainning.pl


Hello,

What is an influence of HInit in HTS-demo_CMU-ARCTIC-SLT/scripts/Training.pl  (**) script on the final quality
of models, and as a result, on the quality of the generated speech?
 
(**) I mean this piece of code in Training.pl
#######################################
# HInit & HRest (initialization & reestimation)
if ($IN_RE) {
...
}
########################################

In other words, can I use row data described by corresponding labels without start/end times of phonemes, but rather using  just
phonetic transcriptions of data?
  
------instead of this-------       --just this-
start_time end_time ah   --->  ah


Thank You and Best Regards,

Stas

Follow-Ups
[hts-users:02149] Re: HInit in Trainning.pl, Sébastien Le Maguer