Hi
I want to build a simple voice recognition system(at first I think it's better to call it command-controll-like system), and want to use pocketsphinx. So I like to know, are there any sources like dictionary, lanuage model, acostic model, train data set, test data set,etc for adaption?
or Should I do everything from scratch with myself?
what's the road map and a comprehensive guide to do that?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
How can I use espeak to build a Farsi dictionary that I can use with pocketsphinx?
I saw their site but couldn't find Farsi in their language list!
Is there any tutorial about this?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Persian language uses Arabic alphabet(فارسی) and our text resources are in that alphabet, so, is it posible to use them in those toolkits introduced in tutorial to build a language model?!
Persian language uses Arabic alphabet(فارسی) and our text resources are in that alphabet, so, is it posible to use them in those toolkits introduced in tutorial to build a language model?!
Yes, you can use ti. We expect UTF-8 encoding of your input data.
you say we can build a model with using of many of these pages in Persian alphabet?!
Yes.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi
I want to build a simple voice recognition system(at first I think it's better to call it command-controll-like system), and want to use pocketsphinx. So I like to know, are there any sources like dictionary, lanuage model, acostic model, train data set, test data set,etc for adaption?
or Should I do everything from scratch with myself?
what's the road map and a comprehensive guide to do that?
Most of the methods are listed in our http://cmusphinx.sourceforge.net/wiki/tutorial, you just need to read it.
You can use espeak to create a phonetic dictionary.
You can train language model from wikipedia dump and from movie subtitles
You can find transcribed podcasts, audiobooks, radio shows to get acoustic model training data. You can run crowdsourcing data collection too.
How can I use espeak to build a Farsi dictionary that I can use with pocketsphinx?
I saw their site but couldn't find Farsi in their language list!
Is there any tutorial about this?
Persian language uses Arabic alphabet(فارسی) and our text resources are in that alphabet, so, is it posible to use them in those toolkits introduced in tutorial to build a language model?!
this is a Persian page in wikipedia : https://fa.wikipedia.org/wiki/%D8%B2%D8%A8%D8%A7%D9%86_%D9%81%D8%A7%D8%B1%D8%B3%DB%8C
you say we can build a model with using of many of these pages in Persian alphabet?!
How?
Yes, you can use ti. We expect UTF-8 encoding of your input data.
Yes.