This is a link to the slides and source code.

Abstract:

This is a presentation on how I've developed audio preprocessing for noise reduction in archival recordings to improve their inteligibility for human annotators, as well as for automated processors.