readme.upto031

ears-0.32, linux下有用的语音信号处理工具包

==========================================================================
Users of version 0.31 and below, please note:  As of 0.32, the
default feature extractor really is Rasta-PLP, it was vanilla PLP
before.  You *have* to "rm ~/.ears/MRASTA-V/*" and train_ears again!
The old behaviour can be gotten by configuring FEATURE PLP and retraining.
Sorry for the inconvenience.   ralf
==========================================================================

Later, if you have a set of patterns for each PLP and MRASTA, you can
switch between them by going through the following sequence:
- run ./train_ears
- change FEATURE in config
- save config
- hit enter
- save the recognizer
- ./listen

This changes the recognizer file accordingly.  In a later version, this
won't be necessary anymore because of different recognizer files for
different feature patterns.  Also, the upcoming 'ears' program will
have both training and listening functionality, so there is all in
one place.