This post was originally posted by me on December 20, 2008 in the Innovate blog, which has since disappeared: see the Internet Archive version saved on March 9, 2009. It was then reposted automatically on the etcjournal.com blog, when the content of the Innovate blog was transfered to it. I am reposting it as it was, except that bolded titles are replaced by H4 title styles and the broken pictures have been removed or replaced.
By Claude Almansi
- Sylvia Monnat, director of captioning at Télévision Suisse Romande (French-speaking Swiss television www.tsr.ch) for the explanations she gave me by phone on live captioning through re-speaking.
- Neal Stein, of Harris Corporation (www.harris.com), for the authorization to publish on YouTube the video excerpt shown below, and for his explanations on the US live radio captioning project.
Why Caption Radio?
Making radio accessible for deaf and hard of hearing persons is not commonly perceived as a priority. For instance, the new version of the Swiss law and ordinance on Radio and Television that came into force in 2007 does add several dispositions about accessibility for people with sight and hearing disabilities but does not mention captioning radio. See art. 7  of the law and art. 7  and 8  of the ordinance (in French). According to most non-deaf people’s “common sense,” deaf persons don’t use radio – just as many non-blind people still believe that blind people can’t use computers.
Yet deaf persons are interested in accessing radio content through captioning, as Cheryl Heppner, Executive Director of NVRC , explains in this video:
The video is from the January 8, 2008, I-CART introductory press conference at CES 2008. The full video can be downloaded from www.i-cart.net. Transcript of the above excerpt:
I’m one of 31 million people in the United States who are deaf or hard of hearing. A number that continues to grow. NPR Labs and its partners are on the verge of making many of my dreams come true. Beyond having that really crucial emergency information, captioned radio could also open up a world I’ve never had, because I lost my hearing before my seventh birthday.
When I am stuck in Washington’s legendary Beltway gridlock, I could check the traffic report and find out why, what my best route would be. I could check the sports scores and follow the games for all my favorite teams. I could know why my husband is always laughing so uproariously when he listens to “Car Talk.” And I could annoy him by singing along badly to the lyrics of his favorite songs.
I can’t wait. Thank you.
NPR’s Live Captioned Broadcast of Presidential Election
The work by NPR Labs and its partners, mentioned by Cheryl Heppner in this January 2008 conference, led to the broadcasting of live captioned debates on NPR during the US election campaign a few months later. The assessment by deaf and hard of hearing people of this experiment was extremely positive. According to the press release “Deaf and Hard of Hearing Vote Yes on New Radio Technology During NPR’s Live Captioned Broadcast of Presidential Election” (Nov. 13, 2008) :
- 95% were happy with the level of captioning accuracy, a crucial aspect for readability and comprehension
- 77% said they would be interested in purchasing a captioned radio display unit when it becomes available
- 86% indicated they would be interested in purchasing a “dual-view” screen display for a car (which would enable a deaf passenger to see the captioned radio text while the driver listens to the radio).
How Are Radio Captions Transmitted?
A digital radio signal can be divided to transmit audio and text, and the text can be read on the radio display. In fact, text messages are already being sent on car radio displays through Radio Data System (RDS). For instance, this is how the Swiss traffic information service Inforoutes updated drivers in real time – or almost – about the state of traffic jams due to work in the Glion tunnel in 2004. (See “Service,” in French, on page 4, in the May 2004 newsletter of Les Radios Francophones Publiques .)
The radio devices used in the experience conducted by NPR Labs and its partners that Cheryl Heppner mentions have a bigger display. For the exact technical explanation of how the captions work, see the presentations section of www.i-cart.net.
Stenocaptioning vs. Respeaking
The NPR experiment mentioned above used “stenocaptioned,” i.e., they were written with a stenotype  whose output gets translated into captions in normal English by computer software. Live stenocaptioning – whether for news broadcasts or for in-presence events in specially equipped venues – seems to be the preferred solution in countries such as the US and Italy that have a tradition of stenotyping court proceedings or parliamentary debates.
In most other European countries, according to Ms. Sylvia Monnat, director of captioning at Télévision Suisse Romande (French-speaking Swiss TV – www.tsr.ch), broadcasters tend to prefer “respeaking,” which works with speech-to-text technology: the software gets trained to recognize the voice of respeakers, and then converts what they repeat into captions.
Ms. Monnat further explained that, on the one hand, the advantages of respeaking involves training. In fact, countries without a stenotyping tradition do not offer courses for it, whereas existing interpretation schools can arrange respeaking courses since it is a normal exercise in the training of conference interpreters. Moreover, respeaking is easier to learn than stenotyping.
On the other hand, it takes time to, first, train the speech-to-text software to recognize the respeakers’ voices and, second, to add words not present in its basic thesaurus for each respeaker’s voice. Moreover, enough respeakers have to be trained so that one whose voice is recognized by the software will be available when needed. Whereas once a new word has been added to the thesaurus of the stenocaptioning software, it can be used by any stenocaptioner.
The fast evolution of technology makes it difficult to foresee the issues of live captioning, even in the near future. Radio and television are merging into “multimedia broadcasting.” And, in turn, the line between broadcasting and the internet is gradually fading (see the HDTV offer by internet providers). Speech-to-text technology will probably continue to improve. Mutimedia devices are also evolving rapidly.
However, the response of the deaf and hard of hearing people who participated in the NPR Live captioning experiment seems to allow one safe surmise: live radio captioning is here to stay, whatever the means it will use tomorrow.
Further information on live captioning can be found in the online version of the “Proceedings of the First International Seminar on Real-time Intralingual Subtitling” held in Forlì, Italy, on Nov. 17, 2006 .
This and other online resources mentioned here have been tagged “captioning” in Diigo and can therefore be found, together with resources added by other Diigo users, in www.diigo.com/tag/captioning.
Original category: Accessibility
Original tags: Almansi, Captioning, Car Talk, CES 2008, Cheryl, Claude, court, Deaf, digital, Diigo, European, Forlì, Glion, hard of hearing, Harris Corporation, HDTV, Heppner, I-CART, Inforoutes, interpretation, Italy, Labs, law, Les Radios Francophones Publiques, Live, Monnat, multimedia broadcasting, Neal Stein, NPR, NVRC, Proceedings of the First International Seminar on Real-time Intralingual Subtitling, Radio, Radio Data System, RDS, Respeaking, signal, software, speech-to-text, stenocaptioned, Stenocaptioning, stenotype, stenotyping, Swiss, Sylvia, Télévision Suisse Romande, television, thesaurus, training, tunnel, Washington, YouTube