Realtime Multiple-pitch and Multiple-instrument Recognition For Music Signals using Sparse Non-negative Constraints

摘要: In this paper we introduce a simple and fast method for realtime recognition of multiple-pitches produced by multiple musical instruments. Our proposed is based on two important facts: one that timbral information any instrument pitch-dependant two, the modulation spectrum same pitch seems to result into persistent representation characteristics instrumental family, as discussed in paper. Using these basic facts, construct learning algorithm obtain templates all possible notes various instruments then devise an online decompose audio buffer using learned templates. The decomposition here are inspired non-negative matrix factorization methods but differ introduction explicit sparsity control. test results show significant rate system real music recordings. We discuss further improvements can be made over system.