Methods of reducing background noise, with digital filtering techniques in old recordings are presented. Different high pass filters have been used to detect impulsive noise, including the -compactly supported wavelets- of I.Daubechies. Two techniques of reconstructing the musical signal during the impulse are described, the linear prediction, and a method using the wavelets of I.Daubechies respectively. Finally, an algorithm worked out by Ephraim and Malah is used to suppress the permanent surface noise.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.