Skip to content

how to generate the full-contextual labels from un-seen text for the application of HMM-based speech synthesis (HTS)

Notifications You must be signed in to change notification settings

bajibabu/make_full_labels

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 

Repository files navigation

follow the steps to generate the full contextual labels

  1. convert the normal text into festival prompts format python ./src/text2prompt.py example/example.txt example/example.data

  2. generate the festival prompts into utts file $FESTVOXDIR/src/promptselect/text2utts -all -level Text -odir example/utts -otype utts -itype data example/example.data

  3. generate the utts into full-contextual and mono label files sh ./src/generate_labels.sh example/labels example/utts $FESTDIR/festival/examples/dumpfeats src/

Where $FESTVOXDIR and $FESTDIR are enviornment variables. They provides the path for festvox and festival tools.

About

how to generate the full-contextual labels from un-seen text for the application of HMM-based speech synthesis (HTS)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published