Menu

CMU Sphinx 4 for Spanish recognition

2015-10-13
2015-10-13
  • Henrik A. D.

    Henrik A. D. - 2015-10-13

    Hi,
    I'm trying to obtain the "AcousticModel", "Dictionary" and "LanguageModel" for spanish. Is any site with theese files for download?

    I would like download theese files and configure Sphinx-4 as its documentation at http://cmusphinx.sourceforge.net/wiki/tutorialsphinx4 said:

    // Set path to acoustic model.
    configuration.setAcousticModelPath("resource:/edu/cmu/sphinx/models/es-es/es-es");
    // Set path to dictionary.
    configuration.setDictionaryPath("resource:/edu/cmu/sphinx/models/es-es/cmudict-es-es.dict");
    // Set language model.
    configuration.setLanguageModelPath("resource:/edu/cmu/sphinx/models/es-es/es-es.lm.bin");

    But at models download section I only find the package: "Spanish Voxforge" (voxforge-es-0.2.tar.gz) and this package don't contains the files that the configuration java object requires.
    Perhaps this package contais the files to compilate the model? I'm already try to compile the spanish model from this package but I got some errors... (Another history... :S)

    Thanks for your help!! And sorry for my bad english :)

     
  • Nickolay V. Shmyrev

    But at models download section I only find the package: "Spanish Voxforge" (voxforge-es-0.2.tar.gz) and this package don't contains the files that the configuration java object requires.

    It does contain everything required. Dictionary is in etc/voxforge_es_sphinx.dic. Language model is etc/voxforge_es_sphinx.transcription.test.lm. Acoustic model is model_parameters/voxforge_es_sphinx.cd_ptm_3000.

    Take into acount that langauge model is only provided for testing and is quite simplistic. It is recommended to train your own language model for basic tasks.

     
  • Henrik A. D.

    Henrik A. D. - 2015-10-13

    Hi Nickolay!, thanks for your response!

    Is that files trained with the source repository of audios from http://www.repository.voxforge1.org/downloads/es/Trunk/Audio/Main/16kHz_16bit?

    Or do you recomend to me that I should build the new files with that repository of samples?

    Thanks for your help and your patience, I am very new to cmu sphinx!

     
    • Nickolay V. Shmyrev

      I recommend you to read our tutorial and learn the difference between acoustic model and language model.

       
  • Henrik A. D.

    Henrik A. D. - 2015-10-13

    Ok, Thanks for the recomendation Nickolay. I will do :)

     

Log in to post a comment.