Skip to content

ArafatTheGoldenBoy/Speech-Recognizer-for-other-language

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Speech-Recognizer-for-other-language

Building a speech recognizer by using cmu-sphinx for all language

Step1:

Install python and perl.

Step2:

After the install need to check by using the command in command prompt, the command is: for python = python -V for perl = perl -v

Step3:

For latest sphinxbase,pocketsphinx and sphinxtrain, need visual studio 2012 or newer version of visual studio. Install vs studio.

Step4:

We need to download sphinx from their site. https://github.com/cmusphinx/sphinxbase https://github.com/cmusphinx/pocketsphinx https://github.com/cmusphinx/sphinxtrain

Step5:

Now we need to extract them and rebuild them and build batch file.

Step6:

we need to follow the documentation from cmu-sphinx. https://cmusphinx.github.io/wiki/tutorialam/

Step7:

we need notepad++ and cmder for showing other language.cmder actually optional.If your language supported normal command prompt then don't need it.

Step8:

we need to change some settings in notepad++ and install TextFX plugin. For this part follow the video to understand how to use them.

Step9:

Create the necessary file for building an acoustic model and language model or Grammer.

Step10:

For language model, http://www.speech.cs.cmu.edu/tools/lmtool-new.html For grammer, #JSGF V1.0 ISO8859-5; grammar sample;

public = (শূন্য | এক | দুই | তিন | চার | পাঁচ | ছয় | সাত | আট | নয় );

Step11:

Now recording the voice and convert them wav file.you can use this site for convert audio file: https://audio.online-convert.com/convert-to-wav

Step12:

For generating training scripts, python ../sphinxtrain/scripts/sphinxtrain -t an4 setup and change an4 to your database name. Now configure some line in sphinx_train.cfg. for a small amount of data, change the number densities from $CFG_CD_TRAIN = 'no'; [165] $DEC_CFG_MODEL_NAME = "$CFG_EXPTNAME.ci_cont"; [242]

Step13:

Now train the data, python ../sphinxtrain/scripts/sphinxtrain run

Step14: (for showing UTF-8 on console)

To change the codepage for the console only, do the following: 1. Start -> Run -> Regedit 2. Go to [HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor\Autorun] 3. Change the value to chcp 65001

Step15:

if there are any error you didnt know how to solve , just asked in cmu-sphinx comminuty.There are always helpful. https://sourceforge.net/p/cmusphinx/discussion/

About

Building a speech recognizer by using cmu-sphinx for all language

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published