Acoustic model
From A Cat's Wiki
An acoustic model is a necessary component of a speech recognition software. An acoustic model can be created by analyzing a speech corpus in a specific kind of way. Statistical representations of the sounds of the words have to be created during the compiling process. Every single distinct sound is related to a phoneme.
Another necessary component is the language model.
[edit] Links
- acoustic model in the German edition of the Parteibuch Wiki
- download acoustic models at VoxForge.org - available in English, Dutch, German, Russian.
