Automatic sample recognition in hip-hop music based on non-negative matrix factorization

We present a method for automatic recognition of samples in hip-hop music. A sample is defined as a short extraction from a source audio corpus that may have been embedded into another audio mixture. A series of non-negative matrix factorizations are applied to log-frequency spectrograms of hip-hop music and the source material from a master corpus. The factorizations result in matrices of base spectra and amplitude envelopes for the original and mixed audio. Each window of the mixed audio is compared to the original audio by examining the extracted amplitude envelopes. Several image-similarity metrics are employed to determine how closely the sampled and mixed amplitude envelopes match. Preliminary testing indicates that, as distinct from existing audio fingerprinting algorithms, the algorithm we describe is able to confirm instances of sampling in a hip-hop music mixture that the untrained listener is frequently unable to detect.