UK Sort Codes - ASR Evaluation

License icon

License:

CC-BY-4.0

Shield icon

Steward:

MDC Curators

Task: ASR

Release Date: 4/15/2026

Format: WEBM, TSV

Size: 23.76 MB

Share

Description

This dataset consists of 1,000 UK bank sort codes read out aloud by a single male speaker of UK English from the Midlands. Sort codes are used to identify bank branches for making money transfers in the United Kingdom. They consist of six digits, split into groups of two, for example `40-30-24`.

Specifics

Licensing

Creative Commons Attribution 4.0 International (CC-BY-4.0)

https://spdx.org/licenses/CC-BY-4.0.html

Considerations

Restrictions/Special Constraints

n/a

Forbidden Usage

Any attempt to clone the voice or train models that imitate the speakers in this dataset is forbidden

Processes

Intended Use

Testing number recognition systems.

Metadata

This dataset consists of 1,000 UK bank sort codes read out aloud by a single speaker of UK English from the Midlands.

Sort codes are used to identify bank branches for making money transfers in the United Kingdom. They consist of six digits, split into groups of two, for example 40-30-24.

Acoustic and recording environment

All samples were recorded on a Xiaomi mobile device in different acoustic environments.

  • Indoors (different room sizes) and outdoors

  • Background noise (road noise, background music, children's noise)

Examples

The numbers are read out naturally including both zero and oh for the number 0, and one one and eleven for the number 11.

  • 40-30-24: Forty thirty twenty four

  • 09-00-44: Oh nine oh oh forty four

  • 01-02-14: Zero one zero two one four

  • 09-00-96: Zero nine zero zero ninety six

  • 60-06-39: Sixty zero six three nine

Banks

The sort codes were selected randomly from a list of the top five UK banks: HSBC, Lloyds, Barclays, National Westminster (Natwest) and Santander. The following number of codes are represented per bank:

 332 HSBC BANK PLC
 284 NATIONAL WESTMINSTER BANK PLC
 231 LLOYDS BANK PLC
 104 BARCLAYS BANK PLC
  32 SANTANDER
   5 NATWEST OFFSHORE LTD
   3 LLOYDS BANK (JERSEY) LTD
   3 BARCLAYS PRIVATE CLIENTS INTERNATIONAL
   2 LLOYDS OFFSHORE
   2 C&G (LLOYDS BANK)
   1 LLOYDS PRIVATE BANKING LTD
   1 HSBC PRIVATE BANK (UK) LTD

Format

There is one directory audio/ with the audio files, in .WEBM format and two TSV files:

  • mapping.tsv: Contains a mapping between the sort code and the audio recording.

  • sortcodes.tsv: Contains a mapping between sort code and bank name.

mapping.tsv

audio_filename	key	sentence
f702763094e195baffbe3388f880ffbf.webm	f702763094e195baffbe3388f880ffbf	77-16-07
cbf80ac96db87e6d0452c852c2912ffb.webm	cbf80ac96db87e6d0452c852c2912ffb	30-11-12
1ee0a79a7449da9ce4675ba278f22904.webm	1ee0a79a7449da9ce4675ba278f22904	60-40-09

sortcodes.tsv

id	sort_code	bank_name
1	77-16-07	LLOYDS BANK PLC
2	30-11-12	LLOYDS BANK PLC
3	60-40-09	NATIONAL WESTMINSTER BANK PLC

Intended usage

Testing number recognition systems.