Performance Analysis of a Source Separation Algorithm
Source separation is an attractive preprocessing step for applying digital audio effects to a single source inside a signal mix. We present a performance analysis of a source separation algorithm based on time-frequency processing and its application to digital audio effects. The performance analysis gives insight to the main analysis parameters for the detection of the number of source signals inside the signal mix. We also analyze the main design parameters for the demixing operation which extracts a single source out of the signal mix.