DDAE speech enhancement

Hyper-parameters are not tuned to optimal

Prerequisites:

Extract spectrogram features:

python spectrum.py data.h5 list_noisy list_clean

Train DDAE using Keras:

python train_DNN.py data.h5

Enhance test wave files using trained model:

python test_gen_spec.py model.hdf5 list_noisy

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
enhanced		enhanced
wav		wav
README.md		README.md
list_clean		list_clean
list_noisy		list_noisy
spectrum.py		spectrum.py
test_gen_spec.py		test_gen_spec.py
train_DNN.py		train_DNN.py