Thank you very much for your work, noticed that the default shortTermWindow and shortTermStep in the audioTrainTest.py code are both 50 milliseconds, my task here is to classify a segmented audio that is roughly 10 milliseconds. Is it possible to set shortTermWindow to 2ms during training?