Posts Tagged ‘model.voca’

Importing my German backup lexicon

Tuesday, January 26th, 2010

Now, I am importing my German backup lexicon into simon. I import it as Julius vocabulary (into the active vocabulary) from the location /home/am3msi/Documents/201001/model/lexicon. Obviously something went wrong, this is the result:

active-vocabulary

1. The pronunciation is not displayed.
2. The words in the vocabulary are all upper-case.

My guess is that I imported the wrong file as Julius vocabulary. I should have imported a different file. Which one is it? Probably I should have imported model.voca. I will try that now. I am now importing /home/am3msi/Documents/201001/model/model.voca as Julius vocabulary. model.voca is the right choice.

Now, I will have to delete the existing simon active vocabulary because it contains a lot of garbage entries (because I imported lexicon instead of model.voca as Julius vocabulary). I will take a look into the folder with the shared files.

I just opened the file /home/am3msi/.kde/share/apps/simon/scenarios/general. Obviously, simon now stores the lexicon (with terminal information) in an XML file. This approach is obviously new. Is it sufficient to delete this file? I will try it.

After deleting the file /home/am3msi/.kde/share/apps/simon/scenarios/general, I restart simon. The active vocabulary is gone (this is what I intended). And the shadow dictionary is still available. Fine. It worked as intended.

You can see that it is necessary to know where the specific file with the active vocabulary is located to fix my error that I made when importing the wrong file as Julius vocabulary.

Tags:
Posted in Ubuntu, dictionary | 1 Comment »

Fresh start with the German speech model

Monday, August 24th, 2009

I am making a fresh start with the German speech model. I renamed the folder file:///home/liberty/.kde/share/apps/simon/model into file:///home/liberty/.kde/share/apps/simon/model-backup-20090824. So I can access my already recorded wav files (and the other files). The result of this operation is that the dictionary is now empty.

I am now importing the PLS dictionary /home/liberty/200908/voxDE20090209-modified.xml to the active lexicon. During the import process, the following three files were created:

1. file:///home/liberty/.kde/share/apps/simon/model/modelsrcrc

WordListDate=2009,8,24,3,59,11

2. file:///home/liberty/.kde/share/apps/simon/model/model.voca

% NS_B
<s>    sil
% NS_E
</s>    sil
% Unknown
übten    y: p t @ n
übte    y: p t @
übt    y: p t
übst    y: p s t
üblichsten    y: p l I C s t n=
[...]

3. file:///home/liberty/.kde/share/apps/simon/model/lexicon

A [A] a:
AACHEN [Aachen] a: x @ n
AACHEN [Aachen] a: x N
AACHEN [Aachen] a: x n=
AACHENS [Aachens] a: x @ n s
AACHENS [Aachens] a: x n= s
AB [ab] a p
ABBAU [Abbau] a p b aU
[...]

I can say that sam helped me to understand what is happening internally in simon. So it seems that I am on the right path. What is to do next?

My goal is to import wav files with the corresponding prompts.

I just defined a grammar. The sentence structure is just “Unknown”. I have only one kind of word: Unknown. Very simple. All words are of the category Unknown. I want to keep it as simple as possible.

OK, there seems to be something wrong. I just synchronized the speech model. And the wav files reappeared. This wasn’t intended. I think that maybe I should rename the folder /usr/share/kde4/apps/simon/model. But the owner is root. Maybe I should use the command gksudo nautilus. I renamed the folder into /usr/share/kde4/apps/simon/model-backup-20090824.

I don’t know how to reset the German speech model. simon just recognized correctly the words “aufnehmen”, and “sehr”. That is not too bad. But other words weren’t recognized.

How can I delete the German speech model? I just want to keep a backup of the wav files, and of the prompts.

Tags: ,
Posted in Ubuntu, dictionary | 2 Comments »

model.voca: changing verb to noun

Wednesday, July 22nd, 2009

I just found out that I can modify the file /home/liberty/.kde/share/apps/simon/model/model.voca. The entry unterhielt U n t @ r h i: l t was classified as % noun. I cut the entry with gedit, and pasted it under % verb. simon displays the changed category.

The next step would be to modify the file /home/liberty/.kde/share/apps/simon/model/lexicon directly with gedit. Take a look at the following two lines:

GÄNGE [Gänge] g & N @
GÄRTNER [Gärtner] g E r t n @ r

I want to see whether I can change the & to an E. After that, I will re-synchronize the speech model. Another pronunciation that I would like to change:

EREIGNIS [Ereignis] E @ r aI g n I s

I would like to remove the @. There is another thing that could be done:

FEUERWEHR [Feuerwehr] f O I @ r v e: @ r

Changing the O I to OI.

Tags: , , , ,
Posted in Ubuntu, dictionary | 3 Comments »