Ambient audio collection format for acoustic scene classification, sound event detection, noise robustness, hearable AI, and multimodal context evaluation.
Acoustic Environment Sets are built for teams evaluating how audio models perform across real-world sound environments.
Each clip is structured around an environment type, scene sub-category, recording context, duration, file format, country or region where scoped, and review status. The dataset is designed to support acoustic model evaluation without relying on private conversations or unclear-rights recordings.
Structured collection
Acoustic Environment Sets use defined scene categories and recording rules.
Metadata depth
The format supports timestamped sound event alignment where scoped, connecting audio segments to scene labels, event labels, and…
Delivery-ready package
Includes audio quality, scene label, speech presence reviews, plus dataset manifest, CSV catalog, consent status metadata, QA…
Request sample access
Share your target scene types, recording conditions, speech-screening rules, annotation needs, and evaluation goals. We’ll reply with a technical scope for the collection.
Ambient audio files · Scene category labels · Environment sub-type metadata · Indoor/outdoor metadata
Time-of-day metadata where scoped · Country or region metadata where scoped · Recording device metadata · Duration metadata
Audio format metadata · Approximate noise level · Speech presence review status · Optional sound event labels
Acoustic Environment Sets use defined scene categories and recording rules. Contributors capture approved ambient environments under instructions for duration, device placement, privacy boundaries, and scene labeling.
Audio quality review · Scene label review · Speech presence review · Metadata review · Annotation consistency review · Consent status review · Delivery approval
Audio files · Scene metadata CSV · Sound event annotations where scoped · Speech screening fields · Dataset manifest JSON · CSV file catalog · QA summary · Consent status metadata · Delivery notes
Multi-speaker English dialogue sets with speaker-labeled transcripts, timestamped turns, accent/session metadata, optional separated-channel audio, and buyer-scoped annotations for ASR, diarization, and voice-agent evaluation.
Multilingual dialogue sets with speaker-labeled transcripts, locale metadata, language-span fields, accent or dialect notes, and aligned transcript structures for multilingual ASR and voice-agent evaluation.
Activity video sets captured across varied real-world environments, with task labels, environment metadata, camera position fields, lighting conditions, duration, and QA status for activity recognition and embodied AI evaluation.
Paired multimodal datasets linking video, audio, screen, text, or image-derived records through shared file IDs, cross-modal manifests, metadata schemas, and QA fields.
Fully scoped custom data collection with buyer-defined modality, schema, metadata fields, annotation layers, QA criteria, manifest structure, delivery package, and collection rules.
Scene-text image sets with language/script metadata, text-region boxes, ground-truth transcription, scene category labels, QA fields, and manifest records for OCR and visual translation.
