2015:Singing Voice Separation Results

From MIREX Wiki

Introduction

Description

These are the results for the 2015 running of the Singing Voice Separation task set. The evaluation set is kindly provided by iKala. If you need to cite this page, please also cite T.-S. Chan, T.-C. Yeh, Z.-C. Fan, H.-W. Chen, L. Su, Y.-H. Yang, and R. Jang, "Vocal activity informed singing voice separation with the iKala dataset," in Proc. IEEE Int. Conf. Acoust., Speech and Signal Process., 2015, pp. 718-722. For more information about this task set please refer to the 2015:Singing Voice Separation page.

Legend

Submission code Submission name Abstract PDF Contributors
FJ1 MIREX 2015 Submission for Singing Voice Separation - Zhe-Cheng Fan, Jyh-Shing Roger Jang
FJ2 MIREX 2015 Submission for Singing Voice Separation - Zhe-Cheng Fan, Jyh-Shing Roger Jang
IIY3 MIREX2015: Singing Voice Separation PDF Yukara Ikemiya, Katsutoshi Itoyama, Kazuyoshi Yoshii
IIY4 MIREX2015: Singing Voice Separation PDF Yukara Ikemiya, Katsutoshi Itoyama, Kazuyoshi Yoshii
MD3 An Ensemble Method for Learning to Extract Vocals from Polyphonic Musical Audio PDF Matt McVicar, Tijl De Bie
MD4 An Ensemble Method for Learning to Extract Vocals from Polyphonic Musical Audio PDF Matt McVicar, Tijl De Bie

Evaluation Criteria

GNSDR = Global Normalized Signal-to-Distortion Ratio
NSDR = Normalized Signal-to-Distortion Ratio
SIR = Signal-to-Interference Ratio
SAR = Signal-to-Artifacts Ratio

Summary

Summary Results

Algorithm Voice GNSDR (dB) Music GNSDR (dB) Runtime (m)
FJ1 6.8236 10.135 3.4014
FJ2 6.3487 9.8678 2.9135
IIY3 4.9862 8.2138 99.1737
IIY4 5.3953 8.77 46.6621
MD3 2.9831 6.3671 121.6655
MD4 3.1022 7.4657 121.1348

NSDR

For the Singing Voice (dB)

Algorithm Mean SD Min Max Median
FJ1 6.8236 3.3945 -1.0987 17.087 6.693
FJ2 6.3487 3.4062 -1.5522 16.139 6.3066
IIY3 4.9862 3.0974 -1.4384 15.753 4.5152
IIY4 5.3953 3.1454 -1.5708 16.296 5.321
MD3 2.9831 3.359 -3.6909 14.99 2.991
MD4 3.1022 2.3219 -1.5255 10.167 3.082

download these results as csv

For the Music Accompaniment (dB)

Algorithm Mean SD Min Max Median
FJ1 10.135 2.9105 2.2446 17.147 10.139
FJ2 9.8678 2.7308 3.379 15.962 9.6999
IIY3 8.2138 3.5171 -11.987 14.139 8.8945
IIY4 8.77 3.8443 -14.661 16.429 9.2513
MD3 6.3671 3.0928 -1.2317 14.238 6.3533
MD4 7.4657 3.8335 -7.6267 17.156 7.6152

download these results as csv

Boxplots

2015-svs-nsdr.png

SIR

For the Singing Voice (dB)

Algorithm Mean SD Min Max Median
FJ1 13.346 10.351 -23.786 37.811 15.755
FJ2 13.746 10.18 -23.665 33.926 16.683
IIY3 15.717 11.713 -27.859 32.436 18.944
IIY4 14.626 11.323 -28.429 31.364 16.947
MD3 11.361 10.162 -24.183 25.137 14.065
MD4 6.6966 9.8508 -30.777 19.816 9.1638

download these results as csv

For the Music Accompaniment (dB)

Algorithm Mean SD Min Max Median
FJ1 11.203 8.9932 -2.3519 43.1 9.5306
FJ2 11.653 9.2526 -4.0102 44.696 9.6655
IIY3 13.288 8.3022 1.3372 40.385 12.376
IIY4 15.057 8.4537 2.1673 43.782 14.952
MD3 9.7413 9.8699 -5.4239 46.37 8.2292
MD4 16.616 8.4294 1.1789 41.702 15.186

download these results as csv

Boxplots

2015-svs-sir.png

SAR

For the Singing Voice (dB)

Algorithm Mean SD Min Max Median
FJ1 11.517 8.3821 -22.529 20.922 13.88
FJ2 10.586 8.1887 -23.897 17.987 13.214
IIY3 8.5224 7.4896 -25.194 17.422 10.527
IIY4 9.214 7.6286 -24.313 17.835 11.522
MD3 7.1543 6.743 -19.385 19.311 8.2378
MD4 11.651 7.329 -13.815 27.014 12.629

download these results as csv

For the Music Accompaniment (dB)

Algorithm Mean SD Min Max Median
FJ1 9.9624 3.2432 -0.75164 18.359 10.182
FJ2 9.082 3.2594 0.040332 17.077 9.1409
IIY3 5.754 3.3033 -4.0821 16.133 5.4503
IIY4 5.9472 3.4381 -6.7539 16.601 5.9557
MD3 5.3175 3.8621 -6.6715 15.684 5.324
MD4 4.2005 3.5672 -5.4033 11.8 4.6933

download these results as csv

Boxplots

2015-svs-sar.png

Runtime Data

Submission Code Runtime (m)
FJ1 3.4014
FJ2 2.9135
IIY3 99.1737
IIY4 46.6621
MD3 121.6655
MD4 121.1348

download these results as csv