All,

I have few pre-recorded audio files (X.wav) in wav format.

I am planning to use HTK (AM) for building my ACOUSTIC MODEL and Julius (http://julius.sourceforge.jp/en_index.php) for building the LANGUAGE MODEL (.voca and .grammar). I am using Voxforge (http://www.voxforge.org/home/dev/acousticmodels/linux/create/htkjulius/tutorial) based HTK.

If carefully designed LANGUAGE MODEL files using Julius and ACOUSTIC MODEL files using HTK, can my approach generate the final TEXT output with command -

$ julian -input rawfile -filelist ../train/wav/X.wav -C julian.jconf

P.S: The example as given in Voxforge usually demonstrate to speak and the output would be in TEXT but not about a pre-recorded audio wav files whose output can be obtained in TEXT mode.

I am using Julius becoz it has Large Vocabulary Compressed Speech Recognition (LVCSR) than any other tools as available in open source.

Please help & correct my initiative before I KICK START..!!

-----

--- (Edited on 12/8/2014 8:16 am [GMT-0600] by srivmuk) ---