2025:Audio Chord Estimation Results - Revision history

Junyan at 14:10, 19 September 2025

2025-09-19T14:10:42Z

Junyan at 12:10, 9 September 2025

2025-09-09T12:10:37Z

Show changes

Junyan: /* Submissions */

2025-09-09T03:58:31Z

‎Submissions

Junyan: /* Test Sets */

2025-09-08T09:03:09Z

‎Test Sets

Junyan at 08:37, 8 September 2025

2025-09-08T08:37:08Z

Junyan at 06:07, 8 September 2025

2025-09-08T06:07:03Z

Junyan: /* Test Sets */

2025-09-08T04:58:32Z

‎Test Sets

Junyan: /* Test Sets */

2025-09-08T04:57:58Z

‎Test Sets

Junyan: /* Test Sets */

2025-09-08T04:57:10Z

‎Test Sets

Junyan at 04:53, 8 September 2025

2025-09-08T04:53:59Z

@@ Line 21: / Line 21: @@
 | MD1
 | Degree-Based Automatic Chord Recognition with Enharmonic Distinction
-| TBA
+| [http://futuremirex.com/portal/wp-content/uploads/2025/audio-chord-estimation/MD1.pdf PDF]
 | Masayuki Doai
 |-
 | wu-ensemble
 | wu-ensemble
-| TBA
+| [http://futuremirex.com/portal/wp-content/uploads/2025/audio-chord-estimation/wu-ensemble.pdf PDF]
 | Yiwei Ding, Christof Weiß
 |-
 | wu-single
 | wu-single
-| TBA
+| [http://futuremirex.com/portal/wp-content/uploads/2025/audio-chord-estimation/wu-single.pdf PDF]
 | Yiwei Ding, Christof Weiß
 |-
 | YK1
 | Semi-Supervised Audio Chord Estimator Based on Disentangled Generative Modeling
-| TBA
+| [http://futuremirex.com/portal/wp-content/uploads/2025/audio-chord-estimation/YK1.pdf PDF]
 | Yiming Wu, Kento Yoshida
 |-
 | BMACE
 | A Mamba-Based Model for Automatic Chord Recognition
-| TBA
+| [http://futuremirex.com/portal/wp-content/uploads/2025/audio-chord-estimation/BMACE.pdf PDF]
 | Chunyu Yuan, Jiyeoung Sim, Johanna Devaney
 |}

@@ Line 23: / Line 23: @@
 | Degree-Based Automatic Chord Recognition with Enharmonic Distinction
 | TBA
-| Muhammad Waseem Akram et al. [*]
+| Masayuki Doai
 |-
 | wu-ensemble
@@ Line 45: / Line 45: @@
 | Chunyu Yuan, Jiyeoung Sim, Johanna Devaney
 |}
 = Test Sets =

@@ Line 51: / Line 51: @@
 ====Main Test Sets====
 * '''Billboard 2013''': The held-out portion of the McGill Billboard dataset, containing mainly western pop songs from the Billboard chart.
 * '''Yamaha_JPOP''': A private dataset annotated by Yamaha Corporation. The dataset contains 200 JPOP songs.
@@ Line 58: / Line 61: @@
 These are datasets that may not be strictly held-out test sets. Some models might have been trained on these datasets; for specific details, please refer to the extended abstracts of each model.
 * '''Billboard 2012''': The public portion of the McGill Billboard dataset.
 * '''RWC Popular''': 100 pop songs from the RWC (Real World Computing) Music Database. 20% songs with English lyrics and 80% songs with Japanese lyrics.

← Older revision		Revision as of 08:37, 8 September 2025
Line 424:		Line 424:
	\| 0.92		\| 0.92
	\|}		\|}
		+
		+
		+	= Task Captain's Note =
		+
		+	* Results on Billboard & RWC Popular are competible with previous years.
		+	* Evaluation tools: https://github.com/ismir-mirex/ace-task-captain-note
		+	* Model Raw outputs: https://github.com/ismir-mirex/ace-output
		+	* Detailed evaluation results: https://github.com/ismir-mirex/ace-results

@@ Line 8: / Line 8: @@
 ! Title
 ! PDF
 |-
 | style="vertical-align:bottom; background-color:#F8F9FA; color:#222;" | Baseline: Chordino
 | style="vertical-align:bottom; background-color:#F8F9FA; color:#222;" | NNLS Chroma v1.1
 | [https://github.com/ismir-mirex/ace-task-captain-notes?tab=readme-ov-file#baselines Link]
 |-
 | Baseline: ISMIR2019
 | Large-Vocabulary Chord Transcription via Chord Structure Decomposition
 | [https://github.com/ismir-mirex/ace-task-captain-notes?tab=readme-ov-file#baselines Link]
 |-
 | MD1
 | Degree-Based Automatic Chord Recognition with Enharmonic Distinction
 | TBA
 |-
 | wu-ensemble
 | wu-ensemble
 | TBA
 |-
 | wu-single
 | wu-single
 | TBA
 |-
 | YK1
 | Semi-Supervised Audio Chord Estimator Based on Disentangled Generative Modeling
 | TBA
 |-
 | BMACE
 | A Mamba-Based Model for Automatic Chord Recognition
 | TBA
 |}
 = Test Sets =

@@ Line 41: / Line 41: @@
 Main Test Sets
-* Billboard 2013: The held-out portion of the McGill Billboard dataset, containing mainly western pop songs from the Billboard chart.
+* '''Billboard 2013''': The held-out portion of the McGill Billboard dataset, containing mainly western pop songs from the Billboard chart.
-* Yamaha_JPOP: A private dataset annotated by Yamaha Corporation. The dataset contains 200 JPOP songs.
+* '''Yamaha_JPOP''': A private dataset annotated by Yamaha Corporation. The dataset contains 200 JPOP songs.
-* Yamaha_Balanced: A private dataset annotated by Yamaha Corporation. The dataset contains 241 songs. While still biased towards JPOP songs, the dataset contains a wider genre distribution: 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%).
+* '''Yamaha_Balanced''': A private dataset annotated by Yamaha Corporation. The dataset contains 241 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%).
 Additional Test Sets
-* Billboard 2012: The public portion of the McGill Billboard dataset.
+* '''Billboard 2012''': The public portion of the McGill Billboard dataset.
-* RWC Popular: 100 pop songs from the RWC (Real World Computing) Music Database. 20% songs with English lyrics and 80% songs with Japanese lyrics.
+* '''RWC Popular''': 100 pop songs from the RWC (Real World Computing) Music Database. 20% songs with English lyrics and 80% songs with Japanese lyrics.
 = Main Results =

@@ Line 38: / Line 38: @@
 |}
-= Main Evaluation Results =
+= Test Sets =
 The following datasets are served as pure test sets. No system is allowed to train on them.
@@ Line 267: / Line 278: @@
 |}
-= Additional Test Sets =
+= Additional Results =
-Below are datasets that are not pure test sets. Some models might have trained on them; please see their extended abstracts for details.
+Below are results on datasets that may not be strictly held-out test sets. Some models might have been trained on these datasets; for specific details, please refer to the extended abstracts of each model.
 == Billboard2012 ==