Text this: Feature extraction based on mel-scaled wavelet packet transform for the diagnosis of voice disorders