Download of Julius_AcousticModels_8kHz16-bit_MFCC_0_D_build396.tgz (Julius_AcousticModels_8kHz16-bit_MFCC_0_D_build396.tgz ( external link: SF.net): 702,602 bytes) will begin shortly. If not so, click link on the left.
VoxForge collects user-submitted speech audio files for the creation of Acoustic Models for Free and Open Source Speech Recognition Engines such as HTK, Julius, ISIP and Sphinx.