Learning algorithms for audio signal enhancement. Part 1: Neural network implementation for the removal of impulse distortions
Andrzej Czyżewski · Journal of the Audio Engineering Society · 1997
Learning algorithms were implemented for the elimination of impulse distortions found in old records and in transmitted audio signals. Neural network simulations were tested with regard to the detection of scratches and clicks. Some special algorithms for the interpolation of distorted signal fragments were studied. Applied methods, results of musical signal processing, and conclusions are presented.