Digital Sound Restoration Laboratory


Digital Sound Restoration Laboratory (DSR-Lab) was a NeXTStep Release 3.2 application designed for digital restoration of sound from old records or magnetic capes. An artificial intelligence learning algorithm, based on neural networks, was used for the detection and elimination of scratches and similar types of impulse noise. Another artificial intelligence learning algorithm, implementing the Wiener adaptive filtering procedures, was used for noise suppression.

The neural network approach, used in the DSR-Lab, brought a "human touch" to the recognition of clicks and noise, because the training process was based on a human perception. This in turn leads to an obvious supremacy of the DSR-Lab algorithms over other methods that exist today for the detection of impulse sounds and noise in the old recordings. Also, the processing speed achieved with the DSR-Lab algorithms was much higher than with other algorithms.

The DSR-Lab application was written in collaboration with the Sound Engineering Department of the Technical University of Gdansk, Gdansk, Poland.

Digital Recordings Canada
$995
Academic discounts were available.
Qty in the collection: 0