2017:Multiple Fundamental Frequency Estimation & Tracking Results - MIREX Dataset

From MIREX Wiki
Revision as of 23:39, 22 October 2017 by Yun Hao (talk | contribs) (General Legend)

Introduction

These are the results for the 2017 running of the Multiple Fundamental Frequency Estimation and Tracking task on MIREX dataset. For background information about this task set please refer to the 2017:Multiple Fundamental Frequency Estimation & Tracking page.

General Legend

Sub code Submission name Abstract Contributors
CB1 Silvet PDF Chris Cannam, Emmanouil Benetos
CB2 Silvet Live PDF Chris Cannam, Emmanouil Benetos
KD1 multiF0_sampled PDF Karin Dressler
KD2 multiF0_midi PDF Karin Dressler
MHMTM1 End-to-End Multi-instrumental ConvNet PDF Gaku Hatanaka, Shinjiro Mita, Alexis Meneses, Daiki Miura, Nattapong Thammasan
MHMTM2 Ensemble category ConvNet to F0 ConvNet PDF Gaku Hatanaka, Shinjiro Mita, Alexis Meneses, Daiki Miura, Nattapong Thammasan
PR1 LPCR PDF Leonid Pogorelyuk, Clarence Rowley
PRGR1 SOT MFFE&T 901 PDF Katarzyna Rokicka, Adam Pluta, Rafal Rokicki, Marcin Gawrysz
PRGR2 SOT MFFE&T 902 PDF Katarzyna Rokicka, Adam Pluta, Rafal Rokicki, Marcin Gawrysz
THK1 Spectral Convolutions PDF John Thickstun, Zaid Harchaoui, Dean Foster, Sham Kakade
WCS1 Piano_Transcription PDF Li Su, Derek Wu, Berlin Chen
ZCY2 Multiple pitch estimation PDF Fuliang Yin, Weiwei Zhang, Zhe Chen
CT1 convlstm PDF Carl Thomé
SL1 samuel-li-onset-detector PDF Samuel Li

Task 1: Multiple Fundamental Frequency Estimation (MF0E)

MF0E Overall Summary Results

Below are the average scores across 40 test files. These files come from 3 different sources: woodwind quintet recording of bassoon, clarinet, horn,flute and oboe (UIUC); Rendered MIDI using RWC database donated by IRCAM and a quartet recording of bassoon, clarinet, violin and sax donated by Dr. Bryan Pardo`s Interactive Audio Lab (IAL). 20 files coming from 5 sections of the woodwind recording where each section has 4 files ranging from 2 polyphony to 5 polyphony. 12 files from IAL, coming from 4 different songs ranging from 2 polyphony to 4 polyphony and 8 files from RWC synthesized midi ranging from 2 different songs ranging from 2 polphony to 5 polyphony.

Detailed Results

Precision Recall Accuracy Etot Esubs Emiss Efa
CB1.results.task1 0.804 0.519 0.498 0.529 0.093 0.389 0.047
CB2.results.task1 0.656 0.460 0.420 0.636 0.174 0.367 0.095
KD1.results.task1 0.724 0.811 0.669 0.419 0.104 0.085 0.231
KD2.results.task1 0.724 0.811 0.669 0.419 0.104 0.085 0.231
MHMTM1.results.task1 0.727 0.782 0.655 0.441 0.119 0.099 0.223
MHMTM2.results.task1 0.391 0.200 0.190 0.860 0.288 0.512 0.060
PR1.results.task1 0.487 0.584 0.418 0.926 0.237 0.178 0.511
PRGR1.results.task1 0.603 0.443 0.408 0.646 0.225 0.332 0.089
PRGR2.results.task1 0.514 0.681 0.476 0.781 0.216 0.103 0.462
THK1.results.task1 0.822 0.789 0.720 0.316 0.076 0.136 0.104
WCS1.results.task1 0.640 0.806 0.593 0.569 0.101 0.094 0.375
ZCY2.results.task1 0.627 0.562 0.506 0.601 0.236 0.203 0.163

download these results as csv

Detailed Chroma Results

Here, accuracy is assessed on chroma results (i.e. all F0's are mapped to a single octave before evaluating)

Precision Recall Accuracy Etot Esubs Emiss Efa
CB1.results.task1 0.851 0.551 0.527 0.497 0.062 0.389 0.047
CB2.results.task1 0.747 0.527 0.479 0.568 0.106 0.367 0.095
KD1.results.task1 0.751 0.842 0.694 0.389 0.074 0.085 0.231
KD2.results.task1 0.751 0.842 0.694 0.389 0.074 0.085 0.231
MHMTM1.results.task1 0.758 0.817 0.682 0.406 0.084 0.099 0.223
MHMTM2.results.task1 0.602 0.314 0.295 0.746 0.174 0.512 0.060
PR1.results.task1 0.569 0.689 0.489 0.822 0.132 0.178 0.511
PRGR1.results.task1 0.712 0.522 0.479 0.567 0.146 0.332 0.089
PRGR2.results.task1 0.595 0.783 0.547 0.680 0.115 0.103 0.462
THK1.results.task1 0.833 0.800 0.730 0.304 0.064 0.136 0.104
WCS1.results.task1 0.660 0.831 0.612 0.544 0.075 0.094 0.375
ZCY2.results.task1 0.740 0.670 0.600 0.493 0.128 0.203 0.163

download these results as csv

Individual Results Files for Task 1

CB1= Chris Cannam, Emmanouil Benetos
CB2= Chris Cannam, Emmanouil Benetos
KD1= Karin Dressler
KD2= Karin Dressler
MHMTM1= Gaku Hatanaka, Shinjiro Mita, Alexis Meneses, Daiki Miura, Nattapong Thammasan
MHMTM2= Gaku Hatanaka, Shinjiro Mita, Alexis Meneses, Daiki Miura, Nattapong Thammasan
PR1= Leonid Pogorelyuk, Clarence Rowley
PRGR1= Katarzyna Rokicka, Adam Pluta, Rafal Rokicki, Marcin Gawrysz
PRGR2= Katarzyna Rokicka, Adam Pluta, Rafal Rokicki, Marcin Gawrysz
THK1= John Thickstun, Sham Kakade, Zaid Harchaoui
WCS1= Li Su, Derek Wu, Berlin Chen
ZCY2= Fuliang Yin, Weiwei Zhang, Zhe Chen


Info about the filenames

The filenames starting with part* comes from acoustic woodwind recording, the ones starting with RWC are synthesized. The legend about the instruments are:

bs = bassoon, cl = clarinet, fl = flute, hn = horn, ob = oboe, vl = violin, cel = cello, gtr = guitar, sax = saxophone, bass = electric bass guitar

Friedman tests for Multiple Fundamental Frequency Estimation (MF0E)

The Friedman test was run in MATLAB to test significant differences amongst systems with regard to the performance (accuracy) on individual files.

Tukey-Kramer HSD Multi-Comparison

TeamID TeamID Lowerbound Mean Upperbound Significance
THK1.results.task1 KD2.results.task1 -0.9423 1.6875 4.3173 FALSE
THK1.results.task1 KD1.results.task1 -0.9423 1.6875 4.3173 FALSE
THK1.results.task1 MHMTM1.results.task1 -0.5298 2.1000 4.7298 FALSE
THK1.results.task1 WCS1.results.task1 1.3452 3.9750 6.6048 TRUE
THK1.results.task1 ZCY2.results.task1 3.1702 5.8000 8.4298 TRUE
THK1.results.task1 CB1.results.task1 3.0202 5.6500 8.2798 TRUE
THK1.results.task1 PRGR2.results.task1 3.7702 6.4000 9.0298 TRUE
THK1.results.task1 CB2.results.task1 5.1202 7.7500 10.3798 TRUE
THK1.results.task1 PR1.results.task1 4.8952 7.5250 10.1548 TRUE
THK1.results.task1 PRGR1.results.task1 5.2702 7.9000 10.5298 TRUE
THK1.results.task1 MHMTM2.results.task1 7.7952 10.4250 13.0548 TRUE
KD2.results.task1 KD1.results.task1 -2.6298 0.0000 2.6298 FALSE
KD2.results.task1 MHMTM1.results.task1 -2.2173 0.4125 3.0423 FALSE
KD2.results.task1 WCS1.results.task1 -0.3423 2.2875 4.9173 FALSE
KD2.results.task1 ZCY2.results.task1 1.4827 4.1125 6.7423 TRUE
KD2.results.task1 CB1.results.task1 1.3327 3.9625 6.5923 TRUE
KD2.results.task1 PRGR2.results.task1 2.0827 4.7125 7.3423 TRUE
KD2.results.task1 CB2.results.task1 3.4327 6.0625 8.6923 TRUE
KD2.results.task1 PR1.results.task1 3.2077 5.8375 8.4673 TRUE
KD2.results.task1 PRGR1.results.task1 3.5827 6.2125 8.8423 TRUE
KD2.results.task1 MHMTM2.results.task1 6.1077 8.7375 11.3673 TRUE
KD1.results.task1 MHMTM1.results.task1 -2.2173 0.4125 3.0423 FALSE
KD1.results.task1 WCS1.results.task1 -0.3423 2.2875 4.9173 FALSE
KD1.results.task1 ZCY2.results.task1 1.4827 4.1125 6.7423 TRUE
KD1.results.task1 CB1.results.task1 1.3327 3.9625 6.5923 TRUE
KD1.results.task1 PRGR2.results.task1 2.0827 4.7125 7.3423 TRUE
KD1.results.task1 CB2.results.task1 3.4327 6.0625 8.6923 TRUE
KD1.results.task1 PR1.results.task1 3.2077 5.8375 8.4673 TRUE
KD1.results.task1 PRGR1.results.task1 3.5827 6.2125 8.8423 TRUE
KD1.results.task1 MHMTM2.results.task1 6.1077 8.7375 11.3673 TRUE
MHMTM1.results.task1 WCS1.results.task1 -0.7548 1.8750 4.5048 FALSE
MHMTM1.results.task1 ZCY2.results.task1 1.0702 3.7000 6.3298 TRUE
MHMTM1.results.task1 CB1.results.task1 0.9202 3.5500 6.1798 TRUE
MHMTM1.results.task1 PRGR2.results.task1 1.6702 4.3000 6.9298 TRUE
MHMTM1.results.task1 CB2.results.task1 3.0202 5.6500 8.2798 TRUE
MHMTM1.results.task1 PR1.results.task1 2.7952 5.4250 8.0548 TRUE
MHMTM1.results.task1 PRGR1.results.task1 3.1702 5.8000 8.4298 TRUE
MHMTM1.results.task1 MHMTM2.results.task1 5.6952 8.3250 10.9548 TRUE
WCS1.results.task1 ZCY2.results.task1 -0.8048 1.8250 4.4548 FALSE
WCS1.results.task1 CB1.results.task1 -0.9548 1.6750 4.3048 FALSE
WCS1.results.task1 PRGR2.results.task1 -0.2048 2.4250 5.0548 FALSE
WCS1.results.task1 CB2.results.task1 1.1452 3.7750 6.4048 TRUE
WCS1.results.task1 PR1.results.task1 0.9202 3.5500 6.1798 TRUE
WCS1.results.task1 PRGR1.results.task1 1.2952 3.9250 6.5548 TRUE
WCS1.results.task1 MHMTM2.results.task1 3.8202 6.4500 9.0798 TRUE
ZCY2.results.task1 CB1.results.task1 -2.7798 -0.1500 2.4798 FALSE
ZCY2.results.task1 PRGR2.results.task1 -2.0298 0.6000 3.2298 FALSE
ZCY2.results.task1 CB2.results.task1 -0.6798 1.9500 4.5798 FALSE
ZCY2.results.task1 PR1.results.task1 -0.9048 1.7250 4.3548 FALSE
ZCY2.results.task1 PRGR1.results.task1 -0.5298 2.1000 4.7298 FALSE
ZCY2.results.task1 MHMTM2.results.task1 1.9952 4.6250 7.2548 TRUE
CB1.results.task1 PRGR2.results.task1 -1.8798 0.7500 3.3798 FALSE
CB1.results.task1 CB2.results.task1 -0.5298 2.1000 4.7298 FALSE
CB1.results.task1 PR1.results.task1 -0.7548 1.8750 4.5048 FALSE
CB1.results.task1 PRGR1.results.task1 -0.3798 2.2500 4.8798 FALSE
CB1.results.task1 MHMTM2.results.task1 2.1452 4.7750 7.4048 TRUE
PRGR2.results.task1 CB2.results.task1 -1.2798 1.3500 3.9798 FALSE
PRGR2.results.task1 PR1.results.task1 -1.5048 1.1250 3.7548 FALSE
PRGR2.results.task1 PRGR1.results.task1 -1.1298 1.5000 4.1298 FALSE
PRGR2.results.task1 MHMTM2.results.task1 1.3952 4.0250 6.6548 TRUE
CB2.results.task1 PR1.results.task1 -2.8548 -0.2250 2.4048 FALSE
CB2.results.task1 PRGR1.results.task1 -2.4798 0.1500 2.7798 FALSE
CB2.results.task1 MHMTM2.results.task1 0.0452 2.6750 5.3048 TRUE
PR1.results.task1 PRGR1.results.task1 -2.2548 0.3750 3.0048 FALSE
PR1.results.task1 MHMTM2.results.task1 0.2702 2.9000 5.5298 TRUE
PRGR1.results.task1 MHMTM2.results.task1 -0.1048 2.5250 5.1548 FALSE

download these results as csv

2017 Accuracy Per Song Friedman Mean Rankstask1.friedman.Friedman Mean Ranks.png

Task 2:Note Tracking (NT)

NT Mixed Set Overall Summary Results

This subtask is evaluated in two different ways. In the first setup , a returned note is assumed correct if its onset is within +-50ms of a ref note and its F0 is within +- quarter tone of the corresponding reference note, ignoring the returned offset values. In the second setup, on top of the above requirements, a correct returned note is required to have an offset value within 20% of the ref notes duration around the ref note`s offset, or within 50ms whichever is larger.

A total of 34 files were used in this subtask: 16 from woodwind recording, 8 from IAL quintet recording and 6 piano.

CB1 CB2 CT1 KD2 PR1 PRGR1 PRGR2 SL1 ZCY2
Ave. F-Measure Onset-Offset 0.3047 0.2064 0.5292 0.4357 0.2489 0.1955 0.2826 0.0011 0.0854
Ave. F-Measure Onset Only 0.5029 0.3742 0.7675 0.6969 0.4497 0.4982 0.4908 0.3593 0.2256
Ave. F-Measure Chroma 0.3207 0.2365 0.5381 0.4465 0.3036 0.2219 0.3319 0.0014 0.0996
Ave. F-Measure Onset Only Chroma 0.5343 0.4276 0.7785 0.7020 0.5099 0.5595 0.5489 0.4080 0.2717

download these results as csv

Detailed Results

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.312 0.304 0.305 0.865
CB2 0.201 0.230 0.206 0.862
CT1 0.513 0.550 0.529 0.863
KD2 0.412 0.468 0.436 0.886
PR1 0.202 0.340 0.249 0.863
PRGR1 0.229 0.176 0.195 0.882
PRGR2 0.264 0.346 0.283 0.903
SL1 0.001 0.001 0.001 -0.039
ZCY2 0.080 0.095 0.085 0.780

download these results as csv

Detailed Chroma Results

Here, accuracy is assessed on chroma results (i.e. all F0's are mapped to a single octave before evaluating)

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.329 0.320 0.321 0.860
CB2 0.229 0.265 0.237 0.858
CT1 0.522 0.560 0.538 0.861
KD2 0.423 0.479 0.447 0.883
PR1 0.245 0.415 0.304 0.862
PRGR1 0.261 0.199 0.222 0.883
PRGR2 0.308 0.411 0.332 0.904
SL1 0.001 0.002 0.001 -0.064
ZCY2 0.093 0.111 0.100 0.826

download these results as csv


Results Based on Onset Only

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.525 0.493 0.503 0.720
CB2 0.375 0.403 0.374 0.677
CT1 0.741 0.802 0.767 0.728
KD2 0.666 0.741 0.697 0.767
PR1 0.363 0.619 0.450 0.696
PRGR1 0.604 0.444 0.498 -2.108
PRGR2 0.488 0.582 0.491 0.731
SL1 0.301 0.477 0.359 -0.023
ZCY2 0.214 0.246 0.226 0.656

download these results as csv

Chroma Results Based on Onset Only

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.558 0.524 0.534 0.700
CB2 0.426 0.465 0.428 0.652
CT1 0.751 0.814 0.778 0.700
KD2 0.671 0.746 0.702 0.760
PR1 0.410 0.704 0.510 0.689
PRGR1 0.678 0.498 0.559 -2.246
PRGR2 0.540 0.659 0.549 0.728
SL1 0.342 0.544 0.408 -0.025
ZCY2 0.257 0.297 0.272 0.634

download these results as csv


Friedman Tests for Note Tracking

The Friedman test was run in MATLAB to test significant differences amongst systems with regard to the F-measure on individual files.

Tukey-Kramer HSD Multi-Comparison for Task2
TeamID TeamID Lowerbound Mean Upperbound Significance
CT1 KD2 -1.2803 0.7794 2.8391 FALSE
CT1 CB1 1.4109 3.4706 5.5303 TRUE
CT1 PRGR1 0.9844 3.0441 5.1038 TRUE
CT1 PRGR2 0.9403 3.0000 5.0597 TRUE
CT1 PR1 2.0285 4.0882 6.1479 TRUE
CT1 CB2 3.6168 5.6765 7.7362 TRUE
CT1 SL1 3.2197 5.2794 7.3391 TRUE
CT1 ZCY2 4.8962 6.9559 9.0156 TRUE
KD2 CB1 0.6315 2.6912 4.7509 TRUE
KD2 PRGR1 0.2050 2.2647 4.3244 TRUE
KD2 PRGR2 0.1609 2.2206 4.2803 TRUE
KD2 PR1 1.2491 3.3088 5.3685 TRUE
KD2 CB2 2.8374 4.8971 6.9568 TRUE
KD2 SL1 2.4403 4.5000 6.5597 TRUE
KD2 ZCY2 4.1168 6.1765 8.2362 TRUE
CB1 PRGR1 -2.4862 -0.4265 1.6332 FALSE
CB1 PRGR2 -2.5303 -0.4706 1.5891 FALSE
CB1 PR1 -1.4421 0.6176 2.6773 FALSE
CB1 CB2 0.1462 2.2059 4.2656 TRUE
CB1 SL1 -0.2509 1.8088 3.8685 FALSE
CB1 ZCY2 1.4256 3.4853 5.5450 TRUE
PRGR1 PRGR2 -2.1038 -0.0441 2.0156 FALSE
PRGR1 PR1 -1.0156 1.0441 3.1038 FALSE
PRGR1 CB2 0.5727 2.6324 4.6921 TRUE
PRGR1 SL1 0.1756 2.2353 4.2950 TRUE
PRGR1 ZCY2 1.8521 3.9118 5.9715 TRUE
PRGR2 PR1 -0.9715 1.0882 3.1479 FALSE
PRGR2 CB2 0.6168 2.6765 4.7362 TRUE
PRGR2 SL1 0.2197 2.2794 4.3391 TRUE
PRGR2 ZCY2 1.8962 3.9559 6.0156 TRUE
PR1 CB2 -0.4715 1.5882 3.6479 FALSE
PR1 SL1 -0.8685 1.1912 3.2509 FALSE
PR1 ZCY2 0.8079 2.8676 4.9273 TRUE
CB2 SL1 -2.4568 -0.3971 1.6626 FALSE
CB2 ZCY2 -0.7803 1.2794 3.3391 FALSE
SL1 ZCY2 -0.3832 1.6765 3.7362 FALSE

download these results as csv

2017 Accuracy Per Song Friedman Mean Rankstask2.onsetOnly.friedman.Friedman Mean Ranks.png

NT Piano-Only Overall Summary Results

This subtask is evaluated in two different ways. In the first setup , a returned note is assumed correct if its onset is within +-50ms of a ref note and its F0 is within +- quarter tone of the corresponding reference note, ignoring the returned offset values. In the second setup, on top of the above requirements, a correct returned note is required to have an offset value within 20% of the ref notes duration around the ref note`s offset, or within 50ms whichever is larger. 6 piano recordings are evaluated separately for this subtask.

CB1 CB2 CT1 KD2 PR1 PRGR1 PRGR2 SL1 ZCY2
Ave. F-Measure Onset-Offset 0.2380 0.1750 0.6201 0.1549 0.2104 0.0462 0.1250 0.0052 0.0131
Ave. F-Measure Onset Only 0.6681 0.4970 0.8139 0.7196 0.5568 0.4242 0.5529 0.3122 0.1482
Ave. F-Measure Chroma 0.2538 0.1863 0.6260 0.1370 0.2204 0.0462 0.1205 0.0061 0.0227
Ave. F-Measure Onset Only Chroma 0.6787 0.5145 0.8150 0.6614 0.5276 0.4311 0.5333 0.3485 0.1846

download these results as csv


Detailed Results

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.275 0.211 0.238 0.813
CB2 0.209 0.153 0.175 0.797
CT1 0.622 0.621 0.620 0.786
KD2 0.150 0.161 0.155 0.844
PR1 0.192 0.249 0.210 0.818
PRGR1 0.070 0.036 0.046 0.844
PRGR2 0.163 0.107 0.125 0.860
SL1 0.005 0.005 0.005 -0.056
ZCY2 0.012 0.016 0.013 0.440

download these results as csv


Detailed Chroma Results

Here, accuracy is assessed on chroma results (i.e. all F0's are mapped to a single octave before evaluating)

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.292 0.226 0.254 0.801
CB2 0.221 0.164 0.186 0.796
CT1 0.627 0.627 0.626 0.782
KD2 0.132 0.142 0.137 0.835
PR1 0.199 0.263 0.220 0.817
PRGR1 0.070 0.036 0.046 0.837
PRGR2 0.157 0.103 0.121 0.859
SL1 0.006 0.006 0.006 -0.195
ZCY2 0.020 0.028 0.023 0.697

download these results as csv


Results Based on Onset Only

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.744 0.613 0.668 0.585
CB2 0.561 0.454 0.497 0.576
CT1 0.814 0.817 0.814 0.694
KD2 0.690 0.756 0.720 0.559
PR1 0.488 0.687 0.557 0.606
PRGR1 0.639 0.334 0.424 -0.438
PRGR2 0.689 0.493 0.553 0.544
SL1 0.325 0.324 0.312 -0.053
ZCY2 0.142 0.161 0.148 0.453

download these results as csv


Chroma Results Based on Onset Only

Precision Recall Ave. F-measure Ave. Overlap
CB1 0.755 0.623 0.679 0.585
CB2 0.580 0.470 0.514 0.577
CT1 0.815 0.819 0.815 0.676
KD2 0.634 0.696 0.661 0.557
PR1 0.461 0.652 0.528 0.606
PRGR1 0.643 0.341 0.431 -0.557
PRGR2 0.665 0.477 0.533 0.541
SL1 0.368 0.361 0.348 -0.060
ZCY2 0.176 0.203 0.185 0.449

download these results as csv

Individual Results Files for Task 2

CB1= Chris Cannam, Emmanouil Benetos
CB2= Chris Cannam, Emmanouil Benetos
CT1= Carl Thomé
KD2= Karin Dressler
PR1= Leonid Pogorelyuk, Clarence Rowley
PRGR1= Katarzyna Rokicka, Adam Pluta, Rafal Rokicki, Marcin Gawrysz
PRGR2= Katarzyna Rokicka, Adam Pluta, Rafal Rokicki, Marcin Gawrysz
SL1= Samuel Li
ZCY2= Fuliang Yin, Weiwei Zhang, Zhe Chen