AudioSep : Separate Anything You Describe
LASS or Language-queried Audio Source Separation is the new paradigm for CASA or Computational Auditory Scene Analysis that aims to separate a target sound from a given mixture of audio using a natural language query that provides the natural yet …