Transient and steady-state component separation for audio signals

Abstract

In this work the problem of transient and steady-state component separation of an audio signal was addressed. In particular, a recently proposed method for separation of transient and steady-state components based on the median filter was investigated. For a better understanding of the processes involved, a modification of the filtering stage of the algorithm was proposed. This modification was evaluated subjectively by listening tests and objectively by an application-based comparison. Also some extensions to the model were presented in conjunction with different possible applications for the transient and steady-state decomposition in the area of audio editing and processing.