Speech Signal Enhancement Using CNN-Driven Encoder-Decoder Architectures for Reverberation Mitigation

There can be lots of challenges in the signal quality due to reverberation noise. This noise badly effects the signal to noise ratio and sound quality. This research article proposes a convolutional neural network (CNN) based encoder decoder structure to address the problem of speech signal with reverberations. This model uses the hierarchical feature extraction ability of CNNs to learn and localise the reverberant contents from speech signals. On the basis of a number of experiments, this method is proved to be more effective than traditional ones. It is very useful in upgrading the speech signals, and also provides a good performance under different conditions which makes it useful for practical uses in communication systems and hearing aids.