About
SoundWatch is an accessibility-focused Android application developed by researchers at the University of Washington's Makeability Lab, led by hard-of-hearing innovator Dhruv Jain. The app runs on smartwatches and leverages machine learning to continuously monitor a user's environment for critical and everyday sounds — from fire alarms to speech — and delivers discreet, glanceable notifications directly on the wrist. Built with deep learning at its core, SoundWatch processes audio locally in real time and can be personalized to recognize the sounds most relevant to each individual user. The system was co-designed with the d/Deaf and hard-of-hearing community, ensuring the interface and alert mechanisms align with real-world accessibility needs. The project is completely free and open-source, welcoming contributions from researchers and developers who want to advance sound accessibility technology. It has been validated through multiple peer-reviewed studies published at top venues including CHI, ASSETS, and CACM, and has been recognized with a Best Artifact Award. SoundWatch is well-suited for accessibility researchers, assistive technology developers, and any d/Deaf or hard-of-hearing individual seeking greater environmental awareness through their smartwatch.
Key Features
- Real-Time Sound Recognition: Uses on-device deep learning to continuously detect and classify important environmental sounds such as fire alarms, doorbells, and speech as they happen.
- Glanceable Smartwatch Alerts: Delivers discreet, always-available notifications directly on the user's wrist, enabling sound awareness without requiring a phone or visible device.
- Personalized Sound Profiles: Allows users to customize and prioritize which sounds the app monitors, tailoring the experience to their specific environment and needs.
- Open-Source & Community-Driven: Fully open-source codebase available for researchers and developers to study, extend, and contribute to, advancing accessibility technology for all.
- Research-Backed Design: Co-designed with d/Deaf and hard-of-hearing communities and validated through multiple peer-reviewed studies at top HCI venues including CHI and ASSETS.
Use Cases
- Helping a d/Deaf user receive a discreet wrist alert when a fire alarm or smoke detector activates in their home or workplace.
- Providing hard-of-hearing individuals with real-time sound awareness in public environments such as offices, transit stations, or universities.
- Allowing accessibility researchers to study and extend sound recognition systems for wearable devices using the open-source codebase.
- Enabling developers to build upon or customize SoundWatch's deep learning models to detect domain-specific sounds for specialized accessibility needs.
- Supporting academic courses or workshops on human-computer interaction and AI accessibility as a hands-on research artifact.
Pros
- Completely Free and Open Source: SoundWatch is freely available at no cost and open-sourced, making it accessible to anyone and encouraging community contributions.
- Purpose-Built for Accessibility: Designed from the ground up with and for d/Deaf and hard-of-hearing users, ensuring relevant features and a thoughtful user experience.
- Peer-Reviewed and Award-Winning: Backed by rigorous academic research published at CHI, ASSETS, and CACM, including a Best Artifact Award, lending strong credibility to its approach.
Cons
- Android Only: Currently limited to Android-based smartwatches, excluding users with Apple Watch or other non-Android wearable platforms.
- Research Project Maturity: As an academic research project, ongoing maintenance and feature updates may be less consistent than a commercially supported product.
- Requires Technical Setup: Being open-source and research-oriented, setup may require more technical knowledge compared to consumer-grade accessibility apps.
Frequently Asked Questions
SoundWatch is designed specifically for d/Deaf and hard-of-hearing individuals who want real-time awareness of environmental sounds through their Android smartwatch.
Yes, SoundWatch is completely free and open-source. You can download the app and also explore or contribute to its source code.
SoundWatch can detect a wide range of environmental sounds including fire alarms, speech, doorbells, and other critical or everyday audio events. Users can also personalize which sounds to monitor.
SoundWatch is an Android app designed for Android-compatible smartwatches. It does not currently support iOS or Apple Watch.
The source code is publicly available. You can access it via the project page at the Makeability Lab or through soundwatch.dhruvjain.info, where contributions from the community are welcome.
