Reverse Engineering the Amen Break - Score-informed Separation and Restoration applied to Drum Recordings

[Back to Main Page]

This page provides audio examples that illustrate the results of the different levels of NMFD side information as well as our proposed post-processing methods. The intention is to point out differences that are not well captured by the PEASS and BSSEval [1] metrics. The audio item on this page has also been used as an illustrative example for the cross-talk reduction in the paper.

Test set item: WaveDrum02_56

Mixture

Kick drum

Snare drum

Hi-hat

License and Acknowledgements

We would like to thank the colleagues from Fraunhofer IDMT for making the "IDMT-SMT-Drums" (http://www.idmt.fraunhofer.de/en/business_units/smt/drums.html) data set publicly available under the Creative Commons Attribution-ShareAlike 4.0 International License ("by-sa"). Our reconstructed signals are derived from the original dataset and consequently fall under the same license.

References

  1. Valentin Emiya, Emmanuel Vincent, Niklas Harlander, and Volker Hohmann
    Subjective and Objective Quality Assessment of Audio Source Separation
    IEEE Transactions on Audio, Speech, and Language Processing, 19(7): 2046–2057, 2011.
    @article{EmiyaVHH11_QualityEvalSourceSep_IEEE-TASLP,
    author    = {Valentin Emiya and Emmanuel Vincent and Niklas Harlander and Volker Hohmann},
    title     = {{S}ubjective and {O}bjective {Q}uality {A}ssessment of {A}udio {S}ource {S}eparation},
    journal   = {{IEEE} Transactions on Audio, Speech, and Language Processing},
    volume    = {19},
    number    = {7},
    year      = {2011},
    pages     = {2046--2057},
    }