The cadenza woodwind dataset - Synthesised quartets for music information retrieval and machine learning - Data Article

Aug 4, 2026

Upload by Noah

This paper presents the Cadenza Woodwind Dataset: publicly available synthesised audio for woodwind quartets, including isolated renderings of each instrument. The dataset was created for training in Cadenza’s second open machine learning challenge (CAD2) focused on rebalancing classical music ensembles, and it supports broader music information retrieval (MIR) algorithm development with machine learning. Scores were selected from the OpenScore String Quartet corpus and rendered by a professional producer using industry-standard software, with convolution reverb to simulate performance spaces and mixed ensembles. Audio and accompanying metadata are provided for research use.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

The Cadenza Woodwind Dataset

Article overview

Dataset creation and rendering

Dataset specifications and accessibility

Value of the data

Background

Data in Brief 57 (2024) 111199 Contents lists available at ScienceDirect Data in Brief journal [homepage: www.elsevier.com/locate/dib](homepage: www.elsevier.com/locate/dib) Data Article The cadenza woodwind dataset: Synthesised quartets for music information retrieval and machine learning Gerardo Roa Dabikea, Trevor J. Cox a,∗, Alex J. Miller a, Bruno M. Fazenda a, Simone Graetzera, Rebecca R. Vosa, Michael A. Akeroydb, Jennifer Firth b, William M. Whitmer b, Scott Bannister c, Alinka Greasley c, Jon P. Barker d a Acoustics Research Centre, University of Salford, UK b Hearing Sciences, Mental Health and Clinical Neurosciences, School of Medicine, University of Nottingham, UK c School of Music, University of Leeds, UK d Department of Computer Science, University of Sheﬃeld, UK a r t i c l e i n f o Article history: Received 12 September 2024 Revised 26 November 2024 Accepted 28 November 2024 Available online 4 December 2024 Dataset link: Cadenza Challenge (CAD2): databases for rebalancing classical music task (Original data). Keywords: MIR Audio Ensemble Deep learning ∗ Corresponding author. E-mail address: [t.j.cox@salford.ac.uk](t.j.cox@salford.ac.uk) (T.J. Cox). Social media: @trevor_cox (T.J. Cox) a b s t r a c t This paper presents the Cadenza Woodwind Dataset. This publicly available data is synthesised audio for woodwind quartets including renderings of each instrument in isolation. The data was created to be used as training data within Cadenza’s second open machine learning challenge (CAD2) for the task on rebalancing classical music ensembles. The dataset is also intended for developing other music information retrieval (MIR) algorithms using machine learning. It was created because of the lack of large-scale datasets of classical woodwind music with separate audio for each instrument and permissive license for reuse. Music scores were selected from the OpenScore String Quartet corpus. These were rendered for two woodwind ensembles of (i) ﬂute, oboe, clarinet and bassoon; and (ii) ﬂute, oboe, alto saxophone and bassoon. This was done by a professional music producer using industry-standard software. Virtual instruments were used to create the audio for each instrument using software that interpreted expression markings in the [https://doi.org/10.1016/j.dib.2024.111199](https://doi.org/10.1016/j.dib.2024.111199) 2352-3409/© 2024 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)) 2 G. Roa Dabike, T.J. Cox and A.J. Miller et al. / Data in Brief 57 (2024) 111199 score. Convolution reverberation was used to simulate a performance space and the ensembles mixed. The dataset consists of the audio and associated metadata. © 2024 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)) Speciﬁcations Table | Subject | Data Science: Applied Machine Learning | | --- | --- | | Speciﬁc subject area | Developing algorithms for Music Information Retrieval (MIR), music signal processing and deep learning. | | Type of data | Digital audio ﬁles
Metadata in ∗ .json format | | Data collection | Nineteen scores were randomly selected from the OpenScore String Quartet corpus. The synthesis of these as woodwind ensembles was performed by a sound engineering professional using professional software. The scores were loaded into the music notation software Steinberg’s Dorico. The string parts were allocated to ﬂute, oboe, clarinet (or alto saxophone) and bassoon. Virtual instruments were used to create the audio for each instrument: Miroslav | | Data source location | Philharmonik 2 by IK Multimedia for the saxophone and Intimate Studio Winds by 8Dio for the other parts. The quartets were then imported into Avid’s

FAQs

What is included in the Cadenza Woodwind Dataset?

The dataset provides synthesised audio for woodwind quartets, including separate isolated renderings for each instrument, along with associated metadata.

Why was the dataset created?

It was created to support training for Cadenza’s second open machine learning challenge (CAD2) on rebalancing classical music ensembles and to help develop other MIR algorithms using machine learning.

How were the music audio files generated?

Music scores were chosen from the OpenScore String Quartet corpus, rendered using professional software with virtual instruments, then mixed and processed with convolution reverberation to simulate performance spaces.

Aug 4, 2026

Upload by Noah