A Review on Signal Decomposition Techniques

Analysis of audio and musical information signals deals with the decomposition into atoms. This subject is most interesting and useful for the researchers who want to invent the inherent properties of the signal under decomposition and to construct a new version of it. Many algorithms have been proposed for the decomposition of audio and musical content and methodologies have been demonstrated to visualizing the behavioral characteristics of such signals.This article outlines an overview of various decomposition algorithms such as high resolution matching pursuit, molecular matching pursuit, cyclic matching pursuit etc. and states the basis for that structural decomposition. Atomic decomposition of signal leads to represent the structure of audio or musical signal, allows to morph on sound into another by varying the parameter of an atom. This paper provides an overview of sound decomposition by Dictionary Based Methods (DBM). Audio signal can be decomposed into atoms by means of user-defined dictionary. This decomposition of signal using DBM can be achieved in different methods based on different parameters. DBM is useful to analyze the signal in multiple time scales such as music, environmental sound and biomedical signals.