How to use a custom acoustic model

[levin]

I have read these posts:
https://www.politepix.com/forums/topic/creating-an-acoustic-model-bundle-for-openears-1-5-and-up/
https://www.politepix.com/forums/topic/creating-custom-acousticmodel-bundle-2015/

I have successfully created a custom acoustic bundle for the Dutch language, using the acoustic model from voxforge.org.

Halle says “you can’t use LanguageModelGenerator” for a custom acoustic model. I have created a dictionary and named it “LanguageModelGeneratorLookupList.txt”.

But, how can I use this method:

OEPocketsphinxController.sharedInstance().startListeningWithLanguageModelAtPath(lmPath, dictionaryAtPath: dicPath, acousticModelAtPath: acousticModel, languageModelIsJSGF: false)

I have the dicPath, which is the dictionary, I do have the acousticModel. But what is my LanguageModelPath (first parameter)?

Thanks in Advance!

[Halle Winkler]

Welcome,

Sorry for the delay, but this was not previous possible with OpenEars. However, as of today it is possible to use Dutch (as well as several other languages) with OpenEars 2.5: https://www.politepix.com/openears

The compatible language models can be downloaded from here: https://www.politepix.com/languages

[Author]

Posts