[Sphinx] [Speech recognition] Learning notes

　　

"Http://www.voxforge.org/home/dev"

1. Acoustic Model Training

The acoustics model's training process is to use your voice data to generate speaker-independent acoustic models, or to adapt the acoustic model of Voxforge online to your sound characteristics.

Linux

• Train your own acoustic model
• How-to-using a script
• Tutorial-step-by-step (with explanations)

• To improve the recognition rate of the acoustic model you create yourself, you need to use more hours of recording to train an acoustic model, a quicker way to adapt the Voxforge's online model to your sound features.

  • Adapt pre-existing acoustic Model to Your Voice
• Want to know how much data is needed to improve your acoustics model? You can use the following tests to learn:
  • Testing Your Acoustic Model with HTK & Julius

Windows

Create Acoustic model & Run Simple Dialog manager:

• How-to-using a script
• Tutorial-step-by-step (with explanations)

"Http://cmusphinx.sourceforge.net/wiki/sphinx4:an4"

[Sphinx] [Speech recognition] Learning notes