Closed hackerekcah closed 6 years ago
The system does not support this type of feature handling directly. To implement this, you would need to bypass feature stacking in application_core.py. Depending a bit on what you want to do with these multichannel features, you could collect features from multiple channels together in system_training-method before feeding it to the Learner class or implement custom Learner class able to handle multiple feature streams.
How to keep two channel features? if I setup mfcc: mono:false Stacking_recipe: mfcc it wont go with two channel, and if i setup like Stacking_recipe: mfcc=0;mfcc=1 it will go two channel, but stacked instead. so How can i get two channel without stack?