Abstrato

Audio Signal Separation and Classification: A Review Paper

Silk Smita, Sharmila Biswas, Sandeep Singh Solanki

Music signals are not solely characterized because of other mixed audio signals. Mixed audio signals contain music signals mixed with speech signals, voice and even background noise.Thus, mixed signals need to classify separately. Researchers have developed many algorithms to solve this problem keeping in mind with their characteristic features of music signals: by timbre, harmony, pitch, loudness etc. The algorithm ICA (Independent component analysis) uses basis of Blind source separation, HSS(Harmonic structure stability), "SOSM" APPROACH (Second Order Statistical Measures Approach), Sinusoidal Parameters based audio classification using FDMSM etc. are some of the mixed signal classification algorithms. This paper highlights all these existing methods and their experimental results.