Speech Recognition Engines

Flat
Commercially-friendly toolkits for building a phoneme aligner
User: Olumide
Date: 12/22/2012 8:27 pm
Views: 5053
Rating: 8

Can anyone kindly suggest a commercially-friendly toolkit for building a phoneme aligner (I'm looking to align phonemes with an audio file). I know that the HTK3 allows models trained with it to be used commercially but the HVite itself cannot.

Regards,

Olumide

--- (Edited on 12/22/2012 8:27 pm [GMT-0600] by Olumide) ---

Re: Commercially-friendly toolkits for building a phoneme aligner
User: kmaclean
Date: 12/22/2012 9:57 pm
Views: 173
Rating: 8

>looking to align phonemes with an audio file

cmusphinx phonemerecognition

julius4-segmentation-kit-v1.0.tar.gz

 

--- (Edited on 12/22/2012 10:57 pm [GMT-0500] by kmaclean) ---

Re: Commercially-friendly toolkits for building a phoneme aligner
User: Olumide
Date: 12/25/2012 8:42 pm
Views: 168
Rating: 6

Thanks for your reply. I have been taking a close look at CMU Sphinx and Julius. While I'm sure they are very both very powerful, neither of them seems to have an extensive tutorial for beginners such as myself. (Although I have a post grad degree in computer science and am a professional software developer my training was not in speech processing.) Would you recommend that I start with HTK in order to learn some of the basics, considering that it has a more detailed tutorial-style documentation?

--- (Edited on 12/25/2012 8:42 pm [GMT-0600] by Olumide) ---

Re: Commercially-friendly toolkits for building a phoneme aligner
User: kmaclean
Date: 1/2/2013 12:56 pm
Views: 1895
Rating: 8

>neither of them seems to have an extensive tutorial for beginners such as myself.

Go with CMU Sphinx - though it does not have a comprehensive manual like HTK, it is fully open (recognizer and acoustic model trainer) and has an excellent wiki and forum.

--- (Edited on 1/2/2013 1:56 pm [GMT-0500] by kmaclean) ---

PreviousNext