- import model.voca from my backup file as Julius dictionary (as active dictionary);
- import Ralf's German dictionary as PLS dictionary (as shadow dictionary), or alternatively import shadow.voca as Julius dictionary (as shadow dictionary).

This makes sense. Then I would have to import my prompts file and the corresponding training.data folder via the import training samples wizard.

I hope that I will learn how to import the information that is stored in my backup folder. I think that I will try it again.

I am not to sure if “Julius dictionary” is the correct description, either.

You might know that simon uses Julius for the recognition. The Julius vocabulary used for this system is now generated from the active scenarios but was – up to 0.2 – stored in the original format that Julius uses (have a look at older model.voca files). The format is quite simple:

% Terminal
Word\tPronunciation
Word2\tPronunciation2

% Terminal2
Word\tPronunciation

I implemented the import feature to make it possible to “restore” the word list of an older simon version so that users upgrading from 0.2 don’t have to re-enter the words manually. They can instead import the old file ~/.kde/share/apps/simon/model/model.voca.

Of course importing the lexicon (~/.kde/share/apps/simon/model/lexicon) as HTK lexicon would work too but terminal information would have be lost as an HTK dictionary does not contain that information.

Greetings,
Peter