Cadenza Project

There is a newer version of the record available.

Published August 8, 2024 | Version V1.0.0
Dataset Open

Cadenza Challenge (CAD1): Submission audio samples for listening to music over the headphones challenge

Authors/Creators

  • 1. ROR icon University of Salford
  • 2. ROR icon University of Sheffield

Contributors

  • 1. ROR icon University of Nottingham
  • 2. ROR icon University of Salford
  • 3. ROR icon University of Leeds
  • 4. ROR icon University of Sheffield

Description

Cadenza

This is the submission data for the listening over headphones task from the First Cadenza Machine Learning Challenge (CAD1).

The Cadenza Challenges are improving music production and processing for people with a hearing loss. According to The World Health Organization, 430 million people worldwide have a disabling hearing loss. Studies show that not being able to understand lyrics is an important problem to tackle for those with hearing loss. Consequently, this task is about improving the intelligibility of lyrics when listening to pop/rock over headphones. But this needs to be done without losing too much audio quality - you can't improve intelligibility just by turning off the rest of the band! We will be using one metric for intelligibility and another metric for audio quality, and giving you different targets to explore the balance between these metrics.

Please see the Cadenza website for a full description of the data

Technical info (English)

This dataset contains the submission audio signals for the CAD1 task1. The signals correspond to 30-second segments of 49 tracks of the MUSDB18-HQ test split.  The signals were processed according the CAD1 requirements. Please refer to the Cadenza challenge website and to the paper for details.

Total number of audio samples: 25,971.

Description of files:

  1. CAD1_data.zip: package containing the audio signals
  2. listeners.json: JSON file with the annonimized listeners' audiograms.
  3. musdb18.test.json: JSON file with the 49 MUSDB18-HQ tracks included.
  4. musdb18.segments.json: JSON file with details of the 30-second segments used.
  5. HAAQI_scores.csv: CSV file with HAAQI scores

The audio signals are organised as:

<TEAM_ID>/<Listener_ID>/<Listener_ID>_<Track_ID>_remix.flac

where:

  • TEAM_ID:  9 unique ids to identify each Team.
  • Listener_ID:  53 unique ids to identify each listener.

Other

Cite as:

G. Roa Dabike, M. A. Akeroyd, S. Bannister, J. P. Barker, T. J. Cox, B. Fazenda, J. Firth, S. Graetzer, A. Greasley, R. R. Vos and W. M. Whitmer, "The First Cadenza Challenges: Using Machine Learning Competitions to Improve Music for Listeners With a Hearing Loss," in IEEE Open Journal of Signal Processing, under review.

Files

CAD1_data.zip

Files (31.8 GB)

Name Size Download all
md5:637cd24b9f382b2d4679076d69c5c3dd
31.8 GB Preview Download
md5:0e0e0cf3a94e12decf09608e8b84c683
2.4 MB Preview Download
md5:8bd264e7fcf0c44028ad05ab28d8e11c
20.6 kB Preview Download
md5:f70095902b5a3e27fc9da1c22bc96dff
5.7 kB Preview Download
md5:9eaa89be1980de0842fe4f431d494e68
7.3 kB Preview Download

Additional details

Funding

UK Research and Innovation
EnhanceMusic: Machine Learning Challenges to Revolutionise Music Listening for People with Hearing Loss EP/W019434/1

Software

Repository URL
https://github.com/claritychallenge/clarity
Programming language
Python