Junyan: MIREX 2026 Audio Downbeat Estimation results

2026-10-14T18:56:10Z

MIREX 2026 Audio Downbeat Estimation results

New page

== Submissions ==

{| class="wikitable"
|-
! Submission
! Title
! PDF
! Authors
|-
| style="vertical-align:bottom; background-color:#F8F9FA; color:#222;" | Baseline: CD4
| style="vertical-align:bottom; background-color:#F8F9FA; color:#222;" | QM Bar & Beat Tracker (beats per bar = 4)
| [https://vamp-plugins.org/plugin-doc/qm-vamp-plugins.html Link]
|
|-
| style="vertical-align:bottom; background-color:#F8F9FA; color:#222;" | Baseline: BeatThis
| style="vertical-align:bottom; background-color:#F8F9FA; color:#222;" | Beat This! Accurate Beat Tracking Without DBN Postprocessing
| [https://github.com/CPJKU/beat_this Link]
|
|-
| CASM
| Context-Aware Semi-Markov Post-Processor for Beat and Downbeat Estimation
| [http://futuremirex.com/portal/wp-content/uploads/2026/audio-downbeat-estimation/CASM.pdf PDF]
| Zhanhong He, Hanyu Meng, Jiahao Zhao, Ganghui Ru, Yaolong Ju
|-
| infgon
| Multi-Scale Expert Reverification for Joint Beat and Downbeat Tracking
| [http://futuremirex.com/portal/wp-content/uploads/2026/audio-downbeat-estimation/infgon.pdf PDF]
| Tianle Yang
|}

== Test Sets ==

* '''GTZAN''': 993 excerpts of the GTZAN dataset with bar annotations (gtzan_tempo_beat, completed with GTZAN-Rhythm; 6 jazz excerpts without bar annotations are excluded).
* '''osu2beat2025''' (new in 2026): 703 full-length songs from ranked osu! beatmaps (anime, Vocaloid, game music and more) with metered beat annotations ([https://github.com/ziyunliu4444/osu2mir Osu2MIR], ISMIR 2025 LBD); audio taken from public osu! mirrors and verified against the MD5 given with each annotation.
* '''Yamaha_JPop''': A private dataset annotated by Yamaha Corporation (250 JPOP songs).
* '''Yamaha_Balanced''': A private dataset annotated by Yamaha Corporation (239 songs). While it is still biased towards JPOP songs, the dataset covers a wider range of genres: J.Pop (10.37%), Rock (10.37%), J.Enka (10.37%), J.Kayoukyoku (10.37%), Soundtrack (10.37%), Western Pop (10.37%), Children's Song (10.37%), R&B (6.22%), Hiphop (4.56%), Jazz (2.49%), Dance (2.49%), World (2.07%), Techno (1.24%), Easy listening (1.24%), J.Minyou (1.24%), Others (5.81%).
* '''JSIR''': to be announced.

For each metric, the best result is shown in bold.

== GTZAN ==

{| class="wikitable" style="text-align:right;"
|- style="font-weight:bold; text-align:left;"
! style="vertical-align:bottom;" | Group
! style="vertical-align:bottom;" | F-measure
! style="vertical-align:bottom;" | Precision
! style="vertical-align:bottom;" | Recall
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-casm (default)
| 77.24
| 77.43
| 79.51
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-direct
| 77.33
| 75.28
| 82.39
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn55_215
| 76.56
| 78.35
| 76.66
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn30_300
| 76.88
| '''78.70'''
| 77.08
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-casm
| 70.95
| 69.42
| 75.27
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-direct
| 68.67
| 63.43
| 77.91
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn55_215
| 71.56
| 73.42
| 71.46
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn30_300
| 70.89
| 73.18
| 70.67
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-casm
| 65.38
| 61.91
| 76.31
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-direct
| 60.75
| 49.62
| '''85.32'''
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn55_215
| 66.41
| 67.51
| 67.89
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn30_300
| 65.81
| 67.05
| 67.18
|- style="vertical-align:bottom;"
| style="text-align:left;" | infgon
| 13.95
| 14.19
| 14.31
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: CD4
| 46.36
| 46.43
| 47.95
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: BeatThis
| '''78.12'''
| 75.67
| 83.93
|}

== osu2beat2025 ==

{| class="wikitable" style="text-align:right;"
|- style="font-weight:bold; text-align:left;"
! style="vertical-align:bottom;" | Group
! style="vertical-align:bottom;" | F-measure
! style="vertical-align:bottom;" | Precision
! style="vertical-align:bottom;" | Recall
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-casm (default)
| 81.99
| 87.64
| 78.48
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-direct
| '''83.78'''
| 86.91
| 82.33
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn55_215
| 82.19
| '''89.97'''
| 77.56
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn30_300
| 82.38
| '''89.97'''
| 77.82
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-casm
| 71.86
| 73.57
| 71.81
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-direct
| 72.46
| 71.43
| 75.07
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn55_215
| 71.01
| 81.18
| 65.74
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn30_300
| 70.54
| 81.57
| 65.16
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-casm
| 63.92
| 56.82
| 80.80
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-direct
| 60.89
| 48.89
| '''85.16'''
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn55_215
| 67.08
| 75.88
| 62.71
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn30_300
| 66.48
| 75.44
| 62.23
|- style="vertical-align:bottom;"
| style="text-align:left;" | infgon
| 11.15
| 14.18
| 9.45
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: CD4
| 44.01
| 48.77
| 41.75
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: BeatThis
| 83.22
| 86.69
| 81.77
|}


== YAMAHA_Balanced ==

{| class="wikitable" style="text-align:right;"
|- style="font-weight:bold; text-align:left;"
! style="vertical-align:bottom;" | Group
! style="vertical-align:bottom;" | F-measure
! style="vertical-align:bottom;" | Precision
! style="vertical-align:bottom;" | Recall
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-casm (default)
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-casm
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-casm
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | infgon
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: CD4
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: BeatThis
|
|
|
|}

== YAMAHA_JPop ==

{| class="wikitable" style="text-align:right;"
|- style="font-weight:bold; text-align:left;"
! style="vertical-align:bottom;" | Group
! style="vertical-align:bottom;" | F-measure
! style="vertical-align:bottom;" | Precision
! style="vertical-align:bottom;" | Recall
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-casm (default)
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-casm
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-casm
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | infgon
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: CD4
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: BeatThis
|
|
|
|}


== JSIR ==

{| class="wikitable" style="text-align:right;"
|- style="font-weight:bold; text-align:left;"
! style="vertical-align:bottom;" | Group
! style="vertical-align:bottom;" | F-measure
! style="vertical-align:bottom;" | Precision
! style="vertical-align:bottom;" | Recall
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-casm (default)
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM beatthis-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-casm
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM mscnn-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-casm
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-direct
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn55_215
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | CASM tcn-dbn30_300
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | infgon
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: CD4
|
|
|
|- style="vertical-align:bottom;"
| style="text-align:left;" | Baseline: BeatThis
|
|
|
|}

2026:Audio Downbeat Estimation Results - Revision history

Junyan: MIREX 2026 Audio Downbeat Estimation results