Signal Enhancement of Source Separation Techniques

Blind Source Separation (BSS) refers to a problem where both the sources and the mixing techniques are unknown, only mixture signals are available for advanced separation process. In several circumstances it is beneficial to recuperate all individual sources from the mixed signal, or partially to segregate a particular source. In laboratory conditions, most of the algorithms work acceptably where input signals, number of source present in the mixture, mixing methodology etc. are well known to the separation process. But in real-life scenarios the problem is much more tangled and it begins with the input signal, a mixture where most of the parameters are unknown. Several approaches have been proposed for the solution of this problem but development is currently still very much in progress. Some of the more fortunate approaches works well when there are no delays or echoes present. The paper summarizes these approaches taken previously to solve this problem and an experiment of source separation which will mix and then de-mix those source signals using various methods like Conv-TasNet, Demucs, IVA, etc. The novel method implemented in this paper involves using LMS filters to enhance the weaker signal strength, thereby improving the quality of the estimated signal in separation through upgradation in signal to noise ratio.