Issue when adding new language model

[xwang]

Hi Halle,

We are using Openears + rapidears + Rejecto.
Now our English and Chinese Version work well, and we want to add a
Russian model to our App.

Since Openears doesn’t support Russian yet, i am trying to make a Runssian.bundle myself.

I downloaded the Russian model from Sphinx, https://sourceforge.net/projects/cmusphinx/files/Acoustic%20and%20Language%20Models/Russian/.

and copied following file to my Russian bundle,
feat.params, feature_transform, mdef, means, mixture_weights, noisedict,
transition_matrices, variances,

also i put a LanguageModelGeneratorLookupList.text which contain the pronunciation of a bunch of words into the bundle.( i can find this file named ru.dic in that model i downloaded )

After all these done, i put this bundle into my App, open this openear logging,

and generateLanguageModelFromArray which contain one Russian word “метеор”.

then i got these errors from log
ERROR: “dict.c”, line 185: Line 1: No pronunciation for word ‘метеор’; ignored
ERROR: “ngram_search_fwdtree.c”, line 336: No word from the language model has pronunciation in the dictionary

Openears start listening normally though, but the hypothesis always be “”.
“The received hypothesis is with a score of -5551 and an ID of 0″

dicPath = [lmGenerator pathToSuccessfullyGeneratedDictionaryWithRequestedName:@”NameIWantForMyLanguageModelFiles”];

according to this path, i found the file Openears generated in my App name NameIWantForMyLanguageModelFiles.dic, there is one word “метеор” in it without pronunciation.( if i use default English Model there are words with their pronunciation in this file, lie “METEOR M IY T IY ER”)

So do you have idea about this? or is it even impossible to use Russian model by Openears?

I read
“https://stackoverflow.com/questions/14254266/add-a-new-language-to-openears”,
and your full discussion in
https://www.politepix.com/forums/topic/using-russian-acoustic-model
page can’t access now.

But it seems like someone have successfully added Russian model, so i am wondering which step i am doing wrong? or would you support Russian language later? Like giving the official version of AcousticModelRussian.bundle

Thanks,
Xi.

[Halle Winkler]

Hi Xi,

With OpenEars 2.5 and the addition of so many new languages, OpenEars moved to an OpenEars-specific bundle package including a new file format, which means that only acoustic model bundles issued by Politepix are compatible with OpenEars (previously, other language acoustic models were explicitly not supported by this project, but with the older and simpler acoustic model structure I was willing to give some hints for creating an unsupported model as long as it was understood that I didn’t support it). There is no 2.5-compatible Russian model, sorry. I unfortunately don’t discuss future plans (or lack of plans) here, apologies for the inconvenience.

[xwang]

So there is no 2.5-compatible Russian model, how about the older version of Openears?

Can i use Russian model in an older version of Openears like 1.5 or 2.0?

[Halle Winkler]

Hi Xi,

Sorry, older versions of OpenEars are not supported by me (or the plugins), so I wouldn’t be able to help with the project of using an obsolete version or with the project of configuring an unsupported acoustic model for use with it. I will close this up since I’m not able to assist with it.

[Author]

Posts