Abstract: Sound is a primary form of communication between human beings. Sound coming to our ears is a mixer of sounds from various sources. It is ability of human hearing capability to extract a sound of particular source while rejecting the sounds from other sources. Human ear can do this effortlessly but how machine can do this?. This problem is noted as cocktail party problem. In this paper it is study of various literature and methods used to solve cocktail party problem.

Keywords: CASA (Computational Auditory Sense Analysis), MVDR (Minimum Variance Distortion less Response), Ideal Binary Mask, cocktail party.