file structure

|- data
|     |- clean_speech   <- speech wav files
|     |- noise          <- noise wav files
|- dnn
|- feat
|- list                 <- contains lists of speech file that will be used in training
|- config.m             <- configuration w.r.t. training, network structure, etc.
|- main.m               <- run this code for training & testing
|- proc_noisy.m         <- uses pre-trained network to process wav files

how to run

1. prepare the training data
2. tune the parameters in config.m
3. run main.m