Description: RASTA is a program for performing feature extraction in speech
recognition systems. It takes files or streams of audio data and
produces files or streams of feature vectors. Three
feature extraction approaches are supported- PLP, log-rasta and
J-rasta. For each case, cepstral features are typically produced,
although there are some options for spectrum-like features.
RASTA handles various file formats and can be used in either
batch mode or an interactive demonstration system.
File list (Check if you may need any files):
Download