Audio · Text / AI & Machine Learning
10K+ Hours of Whispered Speech | AI Training Data | Annotated audio data for AI | Stems, Diarized Transcripts, Scene Descriptions | Global Cover
Advertised by providerLicensing rights not independently verifiedPrice on request
Overview
10K+ Hours of Whispered Speech | AI Training Data | Annotated audio data for AI | Stems, Diarized Transcripts, Scene Descriptions | Global Cover is an audio and text dataset. The provider states a volume of 10,000 hours. Typical applications include speech recognition, speaker recognition and conversational AI. Relevant sectors: AI & Machine Learning. The listing mentions transcription, speaker labels and polygons / segmentation. Pricing is not published and is provided on request.
Primary applications
What the listing offers
- Stated volume of 10k hours.
- Listing mentions transcription, speaker labels, polygons / segmentation.
This dataset contains over 10,000 hours of whispered speech recordings, collected from contributors around the world. Built specifically for AI and machine learning applications, it offers deeply annotated, context-rich audio designed to support next-generation conversational AI, speech recognition, and multimodal understanding. Key Features 1. Comprehensive Annotation Layers Each audio file is delivered with multiple synchronized metadata components: Speaker diarization (per-speaker segmentation and labeling). Individual stems for clean, isolated audio sources. High-accuracy transcripts, including disfluencies and natural speech patterns. Scene descriptions detailing environment, context, background sounds, and conversational intent. Acoustic metadata such as noise profile, channel setup, and recording conditions. This depth of annotation enables advanced tasks including ASR development, intent modeling, emotion recognition, and dialogue-state tracking. 2. Unique Sourcing Capabilities Data is sourced through structured, gamified contribution pipelines that encourage natural, unscripted dialogue across thousands of global participants. This ensures: Authentic, real-world speech. Wide demographic and linguistic variation. Consistent inflow of fresh recordings. Ability to produce custom datasets, including requests for specific accents, age groups, conversation types, emotional tones, or environments (e.g., indoors, outdoors, noisy, clean, telephone, VoIP). 3. Global Linguistic
Related datasets
More in Speech & Audio →Audio · Text · MultimodalAdvertised
Podcast Audio Data | Egyptian Arabic Multi-Speaker Dialogue Audio Corpus | 10,000+ Hours with Full Transcription | .WAV .JSON
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording ‣ Recording content:…
- Language
- Arabic
- Volume
- 10k h
- Domain
- Education +1
- Use case
- Speech recognition
Price on request
More speech & audio datasets
All 1,727 speech & audio datasets →- 10K+ Local News Interviews | AI Training Data | Annotated audio data for AI | Stems, Diarized Transcripts, Scene Descriptions | Global Coverage
- 16kHz Conversational Speech Data | 35,000 Hours | Large Language Model(LLM) Data | Speech AI Datasets | Conversation Audio Data
- 1M+ Hours of Adult Conversations | AI Training Data | Annotated audio data for AI | Stems, Diarized Transcripts, Scene Descriptions | Global Coverage
- 20,000+ Hours of Mandarin Chinese Conversational Speech & Transcripts | Multi-Speaker, Diarized | ASR & Conversational AI Training | Voice Data
- 20K+ Hours of Audio Diaries (Adults) | AI Training Data | Annotated audio data for AI | Stems, Diarized Transcripts, Scene Descriptions | Global Cover
- 20K+ Hours of Debate & Discussion Audio | AI Training Data | Annotated audio data for AI | Stems, Diarized Transcripts, Scene Descriptions | Global