Harmonic-Percussive-Residual Sound Separation using the Structure Tensor on Spectrograms
This is the accompanying website for the paper "Harmonic-Percussive-Residual Sound Separation using the Structure Tensor on Spectrograms" by Richard Füg, Andreas Niedermeier, Jonathan Driedger, Sascha Disch, and Meinard Müller.
Additional Material: Audio Examples for Table 1
Item 1
IBM |
HPR-M |
HPR-ST |
IBM |
HPR-M |
HPR-ST |
IBM |
HPR-M |
HPR-ST |
||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Vibrato | 29.43 | 11.51 | 21.25 | 34.26 | 27.94 | 30.01 | 31.16 | 11.61 | 21.88 | |||
| Impulse | 8.56 | -10.33 | -1.47 | 20.31 | -7.96 | 12.03 | 8.90 | 2.02 | -1.00 | |||
| Noise | 8.49 | -13.53 | 2.58 | 24.70 | -11.99 | 14.12 | 8.61 | 3.97 | 3.06 | |||
Item 2
IBM |
HPR-M |
HPR-ST |
IBM |
HPR-M |
HPR-ST |
IBM |
HPR-M |
HPR-ST |
||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Vocals | 14.82 | 6.48 | 9.18 | 22.75 | 20.83 | 15.61 | 15.60 | 6.68 | 10.42 | |||
| Castanets | 8.48 | 3.79 | 2.37 | 21.59 | 16.09 | 17.96 | 8.73 | 4.16 | 2.56 | |||
| Applause | 7.39 | -2.03 | -0.37 | 20.31 | 1.11 | 6.34 | 7.66 | 3.33 | 1.58 | |||
Table 1. Objective evaluation measures. All values are given in dB. Click on the values to listen to the respective components.
