Hello! I'm a bit familiar with signal processing. But I am quite the newb when it comes to these types of APIs like CMUSphinx. The problem I'm trying to solve is I have about 500+ 3 second recordings of words pronounced in a dead language (Biblical Hebrew). Would it be possible to train Sphinx model to recognize these 500+ words.? I'm not sure the contemporary Hebrew model is similar enough to leverage against. The mobile app is for language memorization and pronunciation. I'm looking for a way for the user to speak and to have the app check whether that is a correct pronunciation or not.
If something like CMU is not the way to go, any suggestions on where to research?
Thanks!
Matt
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The problem I'm trying to solve is I have about 500+ 3 second recordings of words pronounced in a dead language (Biblical Hebrew). Would it be possible to train Sphinx model to recognize these 500+ words.? I'm not sure the contemporary Hebrew model is similar enough to leverage against.
Speech recognition training requires much more data.
The mobile app is for language memorization and pronunciation. I'm looking for a way for the user to speak and to have the app check whether that is a correct pronunciation or not.
Pronunciation scoring is a hard problem ortogonal to speech recognition problem.
If something like CMU is not the way to go, any suggestions on where to research?
You'd better look for an expert who can implement you such thing, you won't be able to do it yourself.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello! I'm a bit familiar with signal processing. But I am quite the newb when it comes to these types of APIs like CMUSphinx. The problem I'm trying to solve is I have about 500+ 3 second recordings of words pronounced in a dead language (Biblical Hebrew). Would it be possible to train Sphinx model to recognize these 500+ words.? I'm not sure the contemporary Hebrew model is similar enough to leverage against. The mobile app is for language memorization and pronunciation. I'm looking for a way for the user to speak and to have the app check whether that is a correct pronunciation or not.
If something like CMU is not the way to go, any suggestions on where to research?
Thanks!
Matt
Speech recognition training requires much more data.
Pronunciation scoring is a hard problem ortogonal to speech recognition problem.
You'd better look for an expert who can implement you such thing, you won't be able to do it yourself.