NST: Swedish Automatic Speech Recognition Database (16 kHz)

A speech database created by Nordic Language Technology for developing automatic speech recognition and dictation systems in Swedish. The dataset has been reorganized from its original version to improve its usefulness, with changes to the file and folder structure. It is hosted by KTH and was last updated on March 26, 2026.

Use Cases

• Training Swedish speech recognition models based on the described speech corpus.
• Fine-tuning dictation systems for Swedish using the reorganized audio data.
• Benchmarking ASR model performance on Swedish speech data.
• Studying the impact of dataset reorganization on model training utility.

Strengths

• Created specifically for Swedish automatic speech recognition and dictation development.
• Reorganized from an original version to improve the usefulness of the database.
• Hosted by KTH, a recognized institution, and last updated on March 26, 2026.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license information are unknown, which may limit suitability assessment.

Provenance

Source

Nordic Language Technology, hosted by KTH.

Collection Method

Created for ASR development; specific collection method not detailed.

Freshness

Last updated 2026-03-26 15:33:28; freshness should be verified.