This thesis addresses different aspects of machine listening using two different approaches, namely (1) A supervised and adaptive sparse representation based approach for identifying the type of background noise and the ...