2014:Audio Cover Song Identification Results

From MIREX Wiki

Jump to: navigation, search

Contents

Introduction

Audio Cover Song (ACS) Identification was evaluated against two datasets: Mixed Collection and Sapp's Mazurka Collection.

Mixed Collection Information

This is the "original" ACS collection. Within the 1000 pieces in the Audio Cover Song database, there are embedded 30 different "cover songs" each represented by 11 different "versions" for a total of 330 audio files (16bit, monophonic, 22.05khz, wav). The "cover songs" represent a variety of genres (e.g., classical, jazz, gospel, rock, folk-rock, etc.) and the variations span a variety of styles and orchestrations.

Using each of these cover song files in turn as the "seed/query" file, we will examine the returned lists of items for the presence of the other 10 versions of the "seed/query" file.

Sapp's Mazurka Collection Information

In addition to our original ACS dataset, we used the Mazurka.org dataset put together by Craig Sapp. We randomly chose 11 versions from 49 mazurkas and ran it as a separate ACS subtask. Systems returned a distance matrix of 539x539 from which we located the ranks of each of the associated cover versions.


General Legend

Sub code Submission name Abstract Contributors
NC2 NC2 PDF Ning Chen
NC3 NC3 PDF Ning Chen

Results

Mixed Collection

Summary Results

NC2 NC3
Total number of covers identified in top 10 33.00 29.00
Mean number of covers identified in top 10 (average performance) 0 0
Mean (arithmetic) of Avg. Precisions 0.02 0.02
Mean rank of first correctly identified cover 77.96 76.26

download these results as csv

Number of Correct Covers at Rank X Returned in Top Ten

Rank NC2 NC3
1 9 9
2 2
3 3 2
4
5 5 5
6 3 1
7 5 1
8 2 3
9 3 4
10 3 2
Total 33 29

download these results as csv

Average Performance per Query Group

Group NC2 NC3
1 0.01 0.01
2 0.02 0.01
3 0.12 0.12
4 0.01 0.01
5 0.01 0.01
6 0.04 0.03
7 0.02 0.01
8 0.01 0.01
9 0.01 0.01
10 0.04 0.04
11 0.02 0.02
12 0.03 0.04
13 0.01 0.01
14 0.01 0.01
15 0.01 0.01
16 0.02 0.01
17 0.01 0.01
18 0.01 0.01
19 0.01 0.01
20 0.01 0.01
21 0.02 0.02
22 0.02 0.01
23 0.01 0.01
24 0.01 0.03
25 0.01 0.01
26 0.12 0.12
27 0.02 0.01
28 0.01 0.01
29 0.01 0.01
30 0.01 0.02

download these results as csv

Sapp's Mazuraka Collection

Summary results

NC2 NC3
Total number of covers identified in top 10 592.00 674.00
Mean number of covers identified in top 10 (average performance) 1.09 1.25
Mean (arithmetic) of Avg. Precisions 0.11 0.13
Mean rank of first correctly identified cover 19.51 18.09

download these results as csv

Number of Correct Covers at Rank X Returned in Top Ten

Rank NC2 NC3
1 67 73
2 57 70
3 51 64
4 61 66
5 57 71
6 53 65
7 58 66
8 59 59
9 70 64
10 59 76
Total 592 674

download these results as csv

Average Performance per Query Group

Group NC2 NC3
1 0.49 0.72
2 0.17 0.27
3 0.16 0.14
4 0.03 0.02
5 0.12 0.07
6 0.36 0.71
7 0.24 0.22
8 0.02 0.02
9 0.04 0.05
10 0.14 0.13
11 0.12 0.12
12 0.27 0.11
13 0.37 0.52
14 0.08 0.07
15 0.22 0.16
16 0.04 0.05
17 0.11 0.07
18 0.19 0.20
19 0.03 0.04
20 0.13 0.10
21 0.16 0.34
22 0.08 0.11
23 0.19 0.19
24 0.59 0.61
25 0.01 0.02
26 0.05 0.06
27 0.03 0.03
28 0.16 0.10
29 0.02 0.03
30 0.07 0.10
31 0.11 0.12
32 0.35 0.46
33 0.46 0.49
34 0.16 0.53
35 0.52 0.52
36 0.18 0.29
37 0.12 0.19
38 0.02 0.12
39 0.03 0.05
40 0.56 0.32
41 0.06 0.11
42 0.02 0.03
43 0.10 0.09
44 0.04 0.05
45 0.02 0.09
46 0.04 0.06
47 0.05 0.04
48 0.04 0.03
49 0.06 0.05

download these results as csv

Individual Results Files

Mixed Collection

Average Precision by Query

NC2 : Ning Chen

NC3 : Ning Chen

Rank Lists

NC2 : Ning Chen

NC3 : Ning Chen

Mixed Collection

Average Precision by Query

NC2 : Ning Chen

NC3 : Ning Chen

Rank Lists

NC2 : Ning Chen

NC3 : Ning Chen

Personal tools
Account Request