I've just created my first HTK acoustic model and would like to test it. Unfortunately the only tutorials that I've come across assume the digits example (which I am not working on) or require a word-net (which I do not have).

Can anyone recommend a corpus neutral guide to using Julius or HTK for large vocabulary CSR.

Thanks,

- Olumide

PS: I do not seek to run the recognizer live either.

--- (Edited on 4/13/2013 3:26 pm [GMT-0500] by Olumide) ---