TAU Urban Acoustic Scenes 2019 Openset: Audio Segments for Open Set Classification

10-second audio segments from 10 urban acoustic scenes, including airport, metro station, and park, designed for open set classification research. The dataset includes additional 'unknown' scene material from 4 other environments, such as beach and library, extracted from the TUT Acoustic scenes 2017 dataset. It was created by Toni Heittola of Tampere University.

Use Cases

• Train open set acoustic scene classifiers based on the 10 target urban environments.
• Benchmark model performance on unknown acoustic scenes based on the provided 'unknown' class material.
• Research domain shift in audio classification based on the mismatch between training and test environments.

Strengths

• Designed for a specific research problem: open set classification in acoustic scenes.
• Includes material for an 'unknown' class from 4 distinct scenes to test model generalization.
• Audio segments have a standardized 10-second duration.

Limitations

• Row count and total dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Tampere University

Collection Method

Likely contains recorded or synthesized audio segments from urban environments.