Benchmark RadarAI BENCHMARK PROFILE
RealDESED
RealDESED is a real-world domestic sound event detection benchmark with 5,710 recordings from 652 participants, 15 classes, and temporally precise annotations. Includes multi-annotator labeling and rich metadata.
- Released
- 2026-07-18
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
Provides a realistic alternative to synthetic or web-crawled SED datasets, supporting evaluation of systems under natural domestic conditions for deployment.
Motivation
This paper presents RealDESED, a real-world domestic sound event detection (SED) benchmark comprising 5,710 audio recordings collected by 652 participants in their homes.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.