2,172 Hours British-English Speech – Let Me Bore You To Sleep
Audio, Speech & Acoustic Datasets
Tags and Keywords
$195,528
About
Let Me Bore You To Sleep – British English Speech Dataset
About
A substantial collection of 2,172 hours, 32 minutes and 1 second of original, naturally spoken British English audio by Jason Newland, creator of free sleep and relaxation recordings since 2006.
Let Me Bore You To Sleep is a long-running spoken-word podcast featuring extended solo recordings covering everyday subjects, personal observations, general knowledge, trivia, magazines, newspapers and intentionally unhurried conversations designed to help listeners relax and fall asleep.
The recordings provide authentic examples of human speech, including natural hesitation, repetition, spontaneous narration, conversational pacing and extended uninterrupted monologues.
Data Product Features
- 1,678 original MP3 recordings
- 2,172 hours, 32 minutes and 1 second of recorded audio
- Approximately 125.7 GB of audio data
- Single adult male British English speaker
- 100% natural human voice, not AI-generated
- Voice-only recordings without background music
- Long-form conversational monologues and narration
- Natural pauses, hesitation, repetition and varied speaking patterns
- Wide variety of everyday subjects and vocabulary
- Original recordings created by Jason Newland
Distribution
The dataset contains 1,678 MP3 audio recordings, individually numbered and organised for identification.
The complete collection occupies approximately 125.7 GB and is securely stored in private Backblaze B2 cloud storage.
Custom delivery is available following agreement on licensing conditions and buyer requirements.
Audio samples and additional technical information can be provided upon request.
Usage and Business Value
This extensive collection of long-form British English speech offers potential value for appropriately licensed applications including:
- Automatic speech recognition (ASR)
- Speech-to-text transcription
- Natural language processing (NLP)
- Spoken-language understanding
- Linguistic and acoustic research
- Evaluation of speech recognition systems
- Long-form conversational speech analysis
- Research into natural speech patterns, hesitation, repetition and pacing
The dataset provides more than 2,172 hours of natural speech from a consistent speaker, recorded across multiple years and covering a broad range of topics.
Its extended monologues and natural conversational style make it particularly relevant to research and development involving long-duration spoken language.
Coverage
- Language: English (British English)
- Accent: Native British English
- Speaker: One adult male — Jason Newland
- Geographic availability: Worldwide
- Recording style: Natural, conversational, calm and unhurried
- Content: Everyday conversation, trivia, personal observations, magazines, newspapers and general knowledge
- Recording period: 2018–2026
Ownership and Licensing
All original recording rights remain with Jason Newland.
Commercial AI training and research licensing may be available under a separately negotiated written agreement.
The following uses are prohibited unless expressly authorised in writing:
- Voice cloning or synthetic replication of Jason Newland's voice
- Speaker impersonation
- Creation of unauthorised synthetic voice replicas
- Resale or redistribution of the original audio recordings
- Transfer or sublicensing of the dataset outside the agreed licence
No transfer of copyright or ownership is included.
All permitted commercial uses, including any AI training rights, must be explicitly agreed before delivery.
Additional Information
These recordings were originally produced for free public listening, relaxation and sleep support.
Commercial dataset licensing is offered separately from their original podcast distribution.
The collection has been organised for licensing as a voice-only British English speech dataset, excluding recordings identified as containing additional speakers.
Further information, audio samples and licensing discussions are available upon request.
TRUSTLoading...
$195,528
Download Dataset in Unknown Format
Recommended Datasets
Loading recommendations...