Datasets:
You need to agree to share your Hugging Face information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
Welcome to the Free To Use Sounds Environmental Audio Dataset.
This evaluation collection features original field recordings
created by Free To Use Sounds LLC, accompanied by written
descriptions and metadata.
Access is provided for internal, non-commercial research and
evaluation of existing audio-processing and machine-learning systems.
By requesting access, you acknowledge that:
AI model training, fine-tuning, and model development require
separate written authorization, including for academic research.
Redistribution, resale, sublicensing, and commercial use of the
recordings or accompanying metadata are not permitted.
Recordings must not be used to identify or profile individuals.
All evaluation activities are subject to the FTUS Research &
Evaluation License included in this repository.
Interested in licensing our larger collection for commercial AI
development? Contact hello@freetousesounds.com.
Log in or Sign Up to review the conditions and access this dataset content.
Free To Use Sounds — Environmental Audio Dataset
Nine Years of Original Field Recordings. One Source. Direct Licensing.
Free To Use Sounds LLC is an independent professional field recording company specializing in high-quality environmental audio, sound effects, and immersive recordings from around the world.
Over the past nine years, every recording in our collection has been personally captured by us, Libby and Marcel.
No third-party recordings. No aggregators. No scraped audio. Every recording comes directly from its original creators.
Our extensive library contains thousands of hours of original audio recorded across more than 30 countries, documenting natural environments, urban soundscapes, transportation systems, industrial locations, wildlife, weather, and everyday human activity.
We offer professionally recorded audio datasets for artificial intelligence, machine learning, academic research, and commercial audio technology development through direct licensing agreements.
What's in this collection?
| Sound Recordings | 61 original field recordings |
| Total duration | ~41 minutes |
| Countries covered | 6 — Finland, Estonia, Thailand, Japan, Australia, Croatia |
| Categories | Environmental Ambiences, Nature, Urban Environments, Transportation, Everyday Sounds, Mechanical Sounds, Water, Weather, Human Activity |
| Format | WAV · 96 / 192 kHz · 24- and 32-bit · mono, stereo, binaural, and multichannel |
| Spatial audio | IRT Cross, Double Mid Side, ORTF 3D, Ambisonics |
| Metadata fields | Filename, Description, Keywords, Category, Duration, Channels, BitDepth, SampleRate, Recording Location, Microphone Information |
| Access | Gated · Manual approval required |
| License | FTUS Research & Evaluation License |
Each recording is paired with written metadata created by us. The accompanying metadata.csv connects the audio files to their descriptions, keywords, categories, and technical specifications.
This collection represents a small sample of the much larger Free To Use Sounds library, featuring over 1,000 hours of original recordings captured around the world.
Evaluation use only. AI model training and fine-tuning require a separate written license.
For commercial dataset licensing: hello@freetousesounds.com
What Makes Our Collection Different?
100% Original Recordings
Every sound in our collection was recorded by us.
We do not aggregate recordings from third-party contributors, scrape audio from online platforms, or redistribute externally sourced content.
This provides organizations with a direct relationship with the original creators and rights holders.
One-Stop Licensing
Free To Use Sounds LLC provides a single point of contact for dataset licensing.
Researchers and commercial developers can negotiate directly with us without coordinating with multiple sound library contributors, marketplaces, or aggregators.
Global Environmental Coverage
Our collection contains thousands of hours of authentic environmental recordings captured across more than 30 countries.
Recording categories include:
- Urban environments and city ambiences
- Transportation, vehicles, trains, and public transit
- Nature, forests, rivers, lakes, and oceans
- Wildlife, birds, insects, and animal vocalizations
- Weather, rain, wind, and thunderstorms
- Industrial environments, machinery, and construction
- Public spaces, markets, and everyday human activity
- Interior and exterior acoustic environments
- Mechanical sounds and specialized sound effects
Listen to Our Recordings
Experience the sounds of Free To Use Sounds before exploring our dataset.
Our playlist on our All In One Bundle main page features original field recordings from around the world, offering a preview of the environments, soundscapes, and acoustic details captured throughout our travels.
🎧 Listen to Free To Use Sounds on Spotify
Professional Audio Quality
Our recordings are captured using professional equipment from manufacturers including Schoeps, Neumann, Sanken, DPA, Sound Devices, and others.
Available recording formats include:
- Mono
- Stereo
- Binaural
- Multichannel surround
- Ambisonics
- Immersive and spatial audio
Our collection includes recordings captured at sample rates up to 192 kHz and 32 bit.
Metadata
Every recording is accompanied by original descriptive metadata created by our team.
Depending on the recording, metadata may include:
- Detailed sound descriptions
- Keywords and acoustic characteristics
- Recording locations and countries
- Environmental context
- Sound categories and classifications
- Technical recording specifications
Our written descriptions, annotations, keywords, and other protectable metadata are part of our intellectual property and are subject to the same licensing framework as our audio recordings.
AI & Machine Learning Applications
Our collection is relevant to research and development in:
- Environmental sound classification
- Sound event detection and recognition
- Audio source separation
- Speech enhancement and background noise analysis
- Audio captioning and audio-text alignment
- Acoustic scene classification
- Spatial and immersive audio research
- Generative audio and sound synthesis
- Multimodal audio-language systems
These are potential applications for separately licensed datasets and do not constitute permission to train AI models using publicly available evaluation materials.
Inside the Evaluation Collection
61 Original Recordings. Six Countries. Multiple Recording Formats.
This evaluation collection offers a carefully selected introduction to the diversity, recording quality, and detailed metadata available through Free To Use Sounds LLC.
The collection contains 61 original field recordings captured across Finland, Estonia, Thailand, Japan, Australia, and Croatia, featuring everything from quiet natural environments and everyday urban life to transportation systems, industrial activity, and individual sound effects.
Rather than focusing on a single type of sound, we've selected recordings that demonstrate the variety of acoustic environments, recording perspectives, and microphone configurations found throughout our larger library.
What Will You Hear?
Urban Environments & Everyday Life
Experience the acoustic character of different cities and public spaces.
Explore Bangkok's busy streets, residential neighborhoods, construction sites, local markets, and train stations. Hear people playing and laughing in Osaka, evening activities in Bangkok's Benchasiri Park, and the quieter urban atmosphere of Helsinki.
The recordings capture natural combinations of traffic, human activity, distant conversations, mechanical sounds, and environmental details.
Nature, Water & Weather
Discover contrasting natural soundscapes from several countries.
Listen to quiet Finnish forests, lakeside winds, coastal environments, and autumn rain. Experience ocean waves washing onto sandy beaches in Okinawa, gentle surf along Tasmania's rocky coastline, and powerful ocean surges rushing through a narrow coastal rock formation.
These recordings demonstrate a wide range of natural acoustic textures, from subtle background atmospheres to powerful, dynamic sound events.
Transportation & Mechanical Environments
Explore trains, buses, road traffic, harbor activity, and mechanical sounds.
The collection includes diesel trains in Bangkok, a quiet bus journey in Osaka, railway activity in Finland, aircraft approaching above a busy highway, and motocross motorcycles racing along a wet track.
You'll also find individual mechanical recordings, including doors, drawers, switches, locks, and motorized curtains.
Interior Spaces & Everyday Sound Effects
Not every recording needs to be spectacular to be useful.
Our collection also includes everyday sounds such as a refrigerator door, microwave door, bathroom door, kitchen drawers, light switches, packaging movements, and interior room tones.
These recordings provide examples of short, recognizable sound events captured in real environments.
Multiple Recording Techniques
One of the distinguishing features of this evaluation collection is the variety of recording configurations represented.
The collection includes:
- Mono recordings — Individual sound sources and environmental details.
- Stereo recordings — Natural left-to-right spatial perspectives.
- Binaural recordings — Immersive environmental perspectives designed for headphone listening.
- Ambisonic recordings — Spatial audio captured using a four-channel microphone configuration.
- Double Mid Side recordings — Flexible multichannel recordings for spatial audio applications.
- IRT Cross recordings — Four-channel environmental recordings captured using a surround microphone arrangement.
- Schoeps ORTF 3D recordings — Eight-channel immersive recordings capturing horizontal and vertical spatial information.
These different configurations allow researchers and developers to explore how recording techniques influence spatial representation, environmental detail, and the perception of sound.
Professional Recording Quality
The evaluation collection contains original WAV recordings captured at:
- 96 kHz and 192 kHz sample rates
- 24-bit and 32-bit recording depths
- Mono, stereo, and multichannel configurations
The recordings were captured using professional field recording equipment, including microphones and recorders from Schoeps, Sennheiser, Rode, Sony, Tascam, Zoom, and Sound Devices.
Metadata
Each recording is accompanied by descriptive metadata created by Free To Use Sounds.
Depending on the recording, this information includes:
- Original sound descriptions
- Descriptive keywords
- Sound categories and subcategories
- Recording country or location
- Recording duration
- Sample rate and bit depth
- Channel count and configuration
- Recording equipment and microphone information
The metadata provides valuable context for understanding the content, acoustic characteristics, and technical properties of each recording.
A Small Introduction to a Much Larger Collection
These 61 recordings represent only a small selection of the original audio available through Free To Use Sounds LLC.
Our complete collection contains over one thousand hours of recordings from more than 32 countries, covering an extensive range of environments, sound events, and recording techniques.
Organizations interested in larger datasets, specialized recording categories, or commercial AI and machine-learning applications can contact us directly to discuss licensing opportunities.
Original Recordings. Written Metadata. Direct Licensing.
Research & Evaluation Dataset
This repository is intended to introduce our professional audio collection to researchers, developers, and organizations interested in original, high-quality environmental recordings.
A limited selection of evaluation recordings and accompanying metadata may be made available under our custom Research & Evaluation License.
Evaluation access is intended for internal, non-commercial assessment of the recordings and testing of existing audio-processing systems.
Training, fine-tuning, or developing AI or machine-learning models using our recordings or metadata requires separate written authorization from Free To Use Sounds LLC, regardless of whether the intended use is commercial or non-commercial.
Access to evaluation materials does not grant commercial AI training rights.
Commercial AI & Machine Learning Licensing
Free To Use Sounds LLC offers individually negotiated licensing agreements for organizations seeking professionally recorded audio datasets.
Licensing opportunities include:
- Curated collections of original environmental recordings
- Large-scale audio datasets
- Commercial AI and machine-learning training
- Audio classification and recognition applications
- Source separation and audio enhancement
- Generative audio research and development
- Commercial model development and distribution
- Custom recording and dataset development
Licensing terms depend on the number of recordings, intended applications, training requirements, model distribution rights, and other project-specific considerations.
All commercial AI training and model distribution rights require a separate written agreement.
Copyright & Intellectual Property
All original audio recordings and protectable accompanying metadata are owned or controlled by Free To Use Sounds LLC.
This repository does not place any recordings or metadata into the public domain.
No rights are granted for unauthorized redistribution, commercial exploitation, dataset replication, scraping, or AI model training.
All rights not expressly granted under the applicable license are reserved.
Copyright © 2026 Free To Use Sounds LLC. All rights reserved.
Licensing Inquiries & Contact
We welcome inquiries from AI researchers, developers, academic institutions, and technology companies interested in licensing original environmental audio datasets.
Free To Use Sounds LLC
Email: hello@freetousesounds.com
Website: https://www.freetousesounds.com
Hugging Face: https://huggingface.co/FreeToUseSounds
Original Recordings. Human-Written Metadata. Direct Licensing.
- Downloads last month
- 97