This page contains information about and links for downloading speech databases distributed by Språkbanken.

The content of the individual databases is described in the pdf-files linked to below. Thus far, these descriptions are only available in Norwegian.
Because of the large volume, the material is split into smaller files. To a certain extent, the individual files reflect natural units.

The databases may be downloaded and used without restrictions for language technology research and development.
Questions and feedback in connection with these resources may be sent to sprakbanken@nb.no .

Speech databases for Norwegian, originally produced by NST

• Descriptions of the databases (pdf – Norwegian)
• Recordings for telephony, 8kHz
  • This database is currently not available
• Recordings for speech recognition (ASR/dictation), 16kHz
  • Download Part 1 (20,2 GB)
  • Download Part 2  (20,7 GB)
  • Download Part 3  (22,2 GB)
  • Download Part 4  (9,2 GB)
  • Download Part 5  (19,5 GB)
  • Download Part 6  (20,2 GB)
  • Download Part 7  (16,8 GB)
  • The validation files are currently not available
• Recordings for dictation, 22kHz
  • Download Part 1  (12,2 GB)
  • Download Part 2  (2,0 GB)
  • Download Part 3  (15 MB)
  • Download Part 4  (6,0 GB)
• Database with recorded hesitation sounds, 16kHz
  • Download the database  (145 MB)
• Database for speech synthesis
  • Download the database  (3,6 GB)

Speech databases for Swedish, originally produced by NST

• Descriptions of the databases (pdf – Norwegian)
• Recordings for speech recognition (ASR/dictation), 16kHz
  • Download Part 1 (25,6 GB)
  • Download Part 2 (22,8 GB)
  • Download Part 3 (23,7 GB)
  • Download Part 4 (17,4 GB)
  • Download Part 5 (19,2 GB)
  • Download Part 6 (16,3 GB)
  • The validation files are currently not available
• Recordings for dictation, 22kHz
  • Download Part 1 (3,9 GB)
  • Download Part 2 (2,7 GB)
  • Download Part 3 (7,5 GB)
  • Download Part 4 (17,5 GB)
  • Download Part 5 (620 KB)
• Database with recorded hesitation sounds, 16kHz
  • Download the database (128 MB)
• Database for speech synthesis
  • Download the database (5,2 GB)

Speech databases for Danish, originally produced by NST

• Descriptions of the databases (pdf – Norwegian)
• Recordings for telephony, 8kHz
  • This database is currently not available
• Recordings for speech recognition (ASR/dictation), 16kHz
  • Download Part 1 (24,0 GB)
  • Download Part 2 (21,4 GB)
  • Download Part 3 (13,1 GB)
  • Download Part 4 (8,0 GB)
  • The validation files are currently not available
• Recordings for dictation, 22kHz
  • Download the database (6,7 GB)
• Database for speech synthesis
  • Download the database (3,9 GB)