Code Monkey home page Code Monkey logo

Comments (8)

GitPiyush avatar GitPiyush commented on September 13, 2024

i tried by changing samplerate also. using both 8000 or 16000, result is same. Please help.

from pocketsphinx.js.

syl22-00 avatar syl22-00 commented on September 13, 2024

@GitPiyush

  • Yes, you should use 8kHz for hub4. you can set outputSampleRate: 8000 in your config, when initializing the recorder.
  • I looked a little bit and can't tell what's wrong yet. How did you train your language model? I tried with hub4 language model, and the output made more sense to me.
  • I don't think hub4 is a good acoustic model, I would recommend to build one on RM1 as explained on sphinx tutorial (but with more senones, like 1000), or the voxforge models.
  • I recently documented how to split your JavaScript file into several smaller files containing the acoustic and language models. It should be useful to you as your pocketspinx.js is becoming very large.

from pocketsphinx.js.

GitPiyush avatar GitPiyush commented on September 13, 2024

Thanks for your response @syl22-00 .
i have created my language model and dictionary using weblink http://www.speech.cs.cmu.edu/tools/lmtool-new.html.
i used rm1_200(default) AM and included my LM and dictionary also but getting same issue. Problem is even if you not speak a single word, hyp is continuously printing words like - ICE NO ALL etc.
Please let me know if you can provide a good AM location.
the main reason why i am in doubt, the same shared AM and LM are working perfect with pocketsphinx desktop application. Is there any problem with generated javascript file.

from pocketsphinx.js.

GitPiyush avatar GitPiyush commented on September 13, 2024

Any inputs?

from pocketsphinx.js.

syl22-00 avatar syl22-00 commented on September 13, 2024

@GitPiyush Your language model is not in the right format. It is in ARPA format and must be converted to CMU's binary format: sphinx_lm_convert -i ECL_Vocab.lm -o ECL_Vocab.DMP. Please refer to the language model documentation of CMUSphinx: http://cmusphinx.sourceforge.net/wiki/tutoriallm

After I converted your file, recognition was actually pretty good with the default RM1 model. I haven't tried with hub4.

from pocketsphinx.js.

GitPiyush avatar GitPiyush commented on September 13, 2024

@syl22-00 it is working fine now. Thanks a lot for all your support.

from pocketsphinx.js.

raziyev avatar raziyev commented on September 13, 2024

Hi every body!
I installed pocketsphinx in windows10 and I run it with the programm in Github.
but it doe'nt recognise any word correctly??
what should I do?
and I don't know how to use "sphinx_lm_convert -i ECL_Vocab.lm -o ECL_Vocab.DMP"
if my problem is it. please help me.
Thank you.

from pocketsphinx.js.

syl22-00 avatar syl22-00 commented on September 13, 2024

@VBR1016 This is the pocketsphinx.js repository, you should probably ask your question directly in https://github.com/cmusphinx

from pocketsphinx.js.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.