Wideband speech recognition using feature extraction techniques and supervised learning for real time applications