Nieman Foundation at Harvard
HOME
          
LATEST STORY
A big week for tech blowback: Regulation, broken promises, and Facebook victimhood
ABOUT                    SUBSCRIBE
Jan. 31, 2011, 10 a.m.

Audio/visual: Adding captions to NPR to reach a text-based audience

Things to take into consideration when trying to caption a radio newscast: how to convey sarcasm, irony, or seriousness; how to represent sound or ambient noise that’s important to a story; how to differentiate the voices of multiple hosts and guests.

Oh, and how to enable captioning on a medium that typically comes with no visuals.

All of these are things NPR Labs has been working on for several years as they try to bring captioned radio into mainstream use. This fall, they’ll begin a pilot program to test out captioned radio at stations around the country through display-capable digital radios and other devices like the Insignia Infocast. The hope is that, one day, captioned radio could also be viewed on mobile apps and tablets.

“We’re trying to build this to work for all public radio and create a large enough model that it can be emulated by others,” Mike Starling, executive director of NPR Labs, told me.

The idea of captioning is much more obvious for television, where the visual medium provides a ready display for text. (Closed captioning dates back to the early 1970s at Boston’s WGBH.) But radio is just as critical a source of news reports and emergency information, Starling said. NPR has come a long way in offering transcripts of their programs online, but they still come with a delay. NPR Labs, which works on software and transmission technology, has been experimenting with captioned radio as digital broadcasting has expanded and radio has burst out of its audio-only bounds. As more radio signals became digital, it allowed for transmission of something like a speech-to-text algorithm that creates a captioning feed. A description from NPR Labs:

Audio recorded in any of NPR’s studios is sent to Master Control, which then routes this audio to both PRSS and to a captioner. The captioner can be either a stenographer or a re-speaker, like the BBC uses. Re-speakers listen to audio and re-speak what they hear into a voice recognition program that has been trained to their voice. This increases the accuracy greatly over speech-to-text programs that are untrained, and removes any background sounds from field reports that might confuse the program.

From there, captions would be sent to stations over the Internet or by satellite and available to read on display-enabled radios, on the web, or on Internet-enabled devices.

Starling said a big reason captioned radio is advancing now is because technology is making it easier to put screens in front of the estimated 25-30 million Americans who are deaf or hard of hearing. That may be part of the reason NPR was looking to make friends with Apple and Android developers at the most recent Consumer Electronics Show. Tablets like the iPad or Samsung Galaxy Tab are the right size for viewing live text, Starling said. But the price of those gadgets means they’re not widely available, which is why Starling considers something like the Infocast or Sony Dash good options that run less than $200. (NPR also developed a prototype car radio display with Delphi, that could act as a screen for turn-by-turn navigation or captions. See the video above for more.)

“It’s perfect timing for us to do the initial work on how to do captioning cost effectively,” he said.

More important than the technology is translating newscasts and other programs in a way that is faithful to content but also understandable for deaf audiences, Starling said. NPR worked with researchers at Gallaudet University to find the best ways to relay non-spoken information in stories, and what factors can interfere with reading captions. In one test, they found that people liked seeing avatars of NPR personalities like Robert Siegel or Michele Norris in captions, but that the extra visuals cut down on the retention of information from stories, Starling said.

“It’s like interpreting for a different language,” he said. “You have to figure out how to best translate this into something else so the full semantic impact is made in articulating a concept.”

The largest trial run of captioned radio was on election night in 2008 when 150 people at five member stations tested captions on a large display, an online stream, and a slide show. Starling said they now want to get a sense whether captioned radio can fit into everyday life and what problems may arise for listeners or stations. Though they’re just scratching the surface of what could be done with captioned radio, Starling said he can see a future where broadcasts could be visualized in different ways, possibly to incorporate images, graphic or video, made available anywhere on any device.

“We’ve got enough to bite off in doing faithful transcripts before we explore how this new artform could be more fully exploited for the intended audience,” he said.

POSTED     Jan. 31, 2011, 10 a.m.
SHARE THIS STORY
   
Show comments  
Show tags
 
Join the 45,000 who get the freshest future-of-journalism news in our daily email.
A big week for tech blowback: Regulation, broken promises, and Facebook victimhood
Among many weeks of bad press for the big tech companies, this week stands out.
The Honest Ads Act would force Internet companies to change their disclosure practices by January 2018
Plus: A former Russian troll speaks out; a definition of disinformation; Wikitribune’s preferred news sources.
From Nieman Reports: The powers and perils of news personalization
News personalization could help publishers attract and retain audiences — in the process making political polarization even worse.