An audio description track works by waiting for silence. A describer finds the pauses in a programme’s dialogue and fills them with short spoken notes on what is on screen. That is how Ira Miller of the Media Access Group at the Boston broadcaster WGBH described it to WNYC’s The Takeaway. The aim, he said, is to sound like a friend in the next seat, whispering what is happening.
The pauses can be tiny. Some last as little as one second, so writers need a firm command of language and grammar to land a description inside them.
What a description sounds like
The Takeaway played a scene from Sherlock to show it. Between lines, a describer’s voice notes that Watson glances at the phone in the Baker Street flat, frowns at a text message and carries it to Sherlock at the microscope. Then it reads the text aloud, from Jim Moriarty, inviting Sherlock to Tower Hill.
Describers are trained to pick out the visuals that matter rather than everything in the frame. The British accessibility charity AbilityNet says the hard part is a script that slots between dialogue and sound, and that careless description can hand blind viewers accidental spoilers.
How the track reaches a television
The idea came from theatre. In 1985 a WGBH technologist learned of a Washington DC group describing stage plays for blind audiences, just as television sets gained stereo sound with two audio channels. He saw that a description track could travel on the spare channel, known as the secondary audio program, or SAP.
WGBH ran its first tests on the PBS series American Playhouse, and its description service debuted in 1990. The station had already made television history with captions. Julia Child’s The French Chef, in 1972, was the first programme ever captioned.
How much television is described
In the UK the first audio-described programme arrived about 25 years before AbilityNet wrote its account, and is believed to have been Coronation Street. In the year before it wrote, the country had 150,000 hours of described television.
The gaps are in the details. Ofcom required only 10% of a channel’s output to carry description at the time, and some popular shows lost it on catch-up services or in a second series. Channel 4’s head of content, Rachel Yendoll, said making catch-up programmes on its 26 platforms accessible was expensive and complicated.
Part of the cost is the tooling. Nigel Megitt of the BBC, who chaired a W3C community group on audio description, said every tool used to make it was proprietary, with no way to exchange production data. The group set out to publish an open file format so more people could make description more cheaply.
