Automatic Speech recognition based on TIMIT, HTK, Theano and using Deep Learning and Feature Extraction.