TAU Urban Acoustic Scenes 2019, Development dataset

40 hours of audio from 10 distinct urban acoustic scenes, each represented by 1440 ten-second segments. The development dataset, created by Toni Heittola of Tampere University, supports research in environmental sound classification. It includes scenes such as airports, metro stations, pedestrian streets, and urban parks.

Use Cases

• Train audio scene classifiers based on the 10 labeled acoustic scenes.
• Benchmark urban sound recognition algorithms using the 10-second audio segments.
• Study acoustic characteristics of specific environments like metro stations or public squares.
• Develop models for urban noise monitoring based on traffic and pedestrian street audio.

Strengths

• Contains 40 hours of total audio data.
• Each of the 10 acoustic scenes has 1440 segments, providing 240 minutes of balanced audio per class.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Toni Heittola, Tampere University