CUSTOM GRADENON-SPEECH AUDIO
Acoustic & Environmental Audio
Urban, industrial, and natural soundscapes with annotated events and scene classification for audio AI and context-aware systems.
- Scenes
- Urban · Industrial · Natural
- Quality
- Multi-layer QC
- Availability
- Sample clips on request
[ OVERVIEW ]
Environmental and non-speech audio recordings captured across urban streets, industrial sites, natural environments, and indoor acoustic spaces. Every recording carries event-level annotations (door slam, siren, machinery hum, birdsong, footsteps, and more), scene-level classification, and ambient metadata like time of day and proximity. Built for teams training acoustic scene classification, event detection, audio-forensics, and context-aware assistants. Scoped per scene type or custom class list.
[ KEY HIGHLIGHTS ]
- Event-level labels across a taxonomy of 200+ acoustic events
- Scene classification: urban street, industrial floor, natural outdoor, indoor residential, office
- Ambient metadata per recording: time of day, weather, proximity, crowd size
- Polyphonic events captured (multiple overlapping sounds per clip)
- Balanced class distributions scoped to your model architecture
- Field recordings with calibrated capture devices and known SPL
- Licensed per scene type or custom event-class list
[ TECHNICAL SPECIFICATIONS ]
- Files
- WAV, 44.1 or 48 kHz, 16-24-bit, mono or stereo as the scene requires
- Annotations
- Event tags with onset/offset timestamps · polyphonic labels · scene classification
- Schema
- JSON with AudioSet-compatible taxonomy · custom class lists on request
- Licensing
- Commercial training rights · per-scene or full-corpus · location-release handled per file
More from the catalog.
Explore the full catalog, or scope a custom build matched to your brief.