Researchers in Tunisia have published details of a speech enhancement system that uses two distinct tools to reduce the noise from a recorded or sampled voice signal. In a new article, the researchers explain how a bionic wavelet transform and a recurrent neural network can be used for speech enhancement.