Developing Radio for the Deaf and the Deaf-Blind
To produce the first, live-captioned radio programming and Braille radio programming, NPR Labs software developers and engineers joined forces with a cognitive scientist/professor from Towson, Dr. Ellyn Sheffield, who researched, designed and developed a captioning center using a process known as voice writing. The center uses commercially available voice recognition software and proprietary editing tools to create virtual real-time captioning for radio broadcasts. The aim of the project was to use new digital technologies in an affordable way while producing timely, accurate captions.
How Live-Captioned Radio Works
To create live-captioned radio, a broadcast is fed from NPR headquarters to Towson University, where voice writers are listening on headphones and continually re-speaking everything that they hear. That audio output is fed to a caption editor who corrects and formats the information for readability before is the captions are transmitted to consumers. All this occurs within 20 seconds of the live broadcast. Captions are currently available on the program providers' websites, but in the future will be available through NPR Member Station websites as well.
The team also wanted to make the content available for the deaf-blind, so they converted the captioning feed for use with stand-alone Refreshable Braille Displays.
Interested Audiences and First Captioned Program
Throughout the process, the team had input from a range of national and regional organizations, including the Northern Virginia Resource Center for the Deaf and Hard of Hearing Persons, the Hearing Loss Association of America, the Helen Keller National Center, the National Association of the Deaf, the National Federation of the Blind, the American Association of the Deaf-Blind, Gallaudet University and more than 50 other interested organizations and their members. This initial work was supported by a grant from the Department of Education's National Institute on Disability and Rehabilitation Research.
As a result, the team recently introduced the first live-captioned radio show, Latino USA with Maria Hinojosa, distributed by NPR. Captions can be seen on the program's web site.
- Radio Text Display: Results from 3 Studies
June 2009 - In three studies, consumer preferences and memory performance for radio text displays were examined. In total over 140 consumers participated in this experimental program, including hearing, hard-of-hearing and deaf individuals.
- Results from NPR Election Night 2008 Captioned Radio Event
November 2008 - Approximately 150 guests visited 5 locations (Baltimore, Boston, Denver, Phoenix and Washington) on election night 2008. 75 deaf and hard-of-hearing participants responded to questions concerning 3 displays.
- Consolidated Requests for Information
- Fundamentals of Digital Radio Captioning
- Accuracy Analysis for Radio Broadcast Deployment of Captioned Radio Services
- NPR Labs Speech to Text RFI