GitHub - iBehruz/speach_recognition: The Speach recognition app for android, desktop. Avalibale languages are Russion, English, Uzbek.

iBehruz / speach_recognition Public

Notifications You must be signed in to change notification settings
Fork 0
Star 1

The Speach recognition app for android, desktop. Avalibale languages are Russion, English, Uzbek.

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
pocketsphinx		pocketsphinx
sphinxtrain		sphinxtrain
zero_ru.cd_cont_4000		zero_ru.cd_cont_4000
zero_ru.cd_ptm_4000		zero_ru.cd_ptm_4000
zero_ru.cd_semi_4000		zero_ru.cd_semi_4000
.gitignore		.gitignore
COPYING		COPYING
README		README
decoder-test.sh		decoder-test.sh
decoder-test.wav		decoder-test.wav
ru.dic		ru.dic

Repository files navigation

Official site: https://cmusphinx.github.io/

This package contains Russian acoustic model and Russian langauge
model based on morphological composition

Acoustic model:

 - sample rate 8000Hz
 - 4000 senones
 - context depend
 - continuous
 - all noise marked as SIL
 - created used the data from http://www.voxforge.org and https://github.com/zamiron/ru4sphinx

Language model:

 - 3gram
 - 500000 base word in vocabulary
 - created using the data from http://www.lib.ru

Example how to use it:

    pocketsphinx_continuous -infile test.wav \
	-hmm zero_ru.cd_cont_4000 -dict ru.dic -lm ru.lm -remove_noise no