Future innovations in automating alternatives
From Steeple
Accessibility requirements report
- Definitions of terms
- Accessibility requirements for podcasting-related media
- Support for disabled students
- Future innovations in automating alternatives
- Accessibility evaluation
- Accessibility of iTunes
- References and other useful resources
You can look at all pages tagged with 'Accessibility requirements':
[edit] 1 Future innovations in automating alternatives
[edit] 1.1 Speech to text: speech recognition
The US National Center on Accessible Information Technology in Education reports that speech recognition technology is not yet advanced enough to allow conversion of speech in an audio track to be converted to text.
- “The greatest limitation with speech recognition technology is that it is only accurate in optimum situations, where the speaker has devoted time to training the software to recognize his or her speech patterns, where audio quality and the accoustics [sic] of the recording environment are excellent, and where distracting background noises are minimal. Few multimedia presentations meet these criteria.” 15
On the other hand speech recognition can be used in the production of a transcript. A transcriber, who has trained a speech recognition system to recognise their voice, can repeat the dialogue and narration of the video and the system can create text from their speech.
When a transcript is available, speech recognition technology can also be used to automatically create synchronised captions by recognising when a particular word was spoken (which is more achievable than recognising which word was spoken).
15 Can captions be generated automatically using speech recognition? http://www.washington.edu/accessit/articles?1209
[edit] 1.2 Sign language interpretation
There have been various projects investigating the possibilities of automatically translating speech or text into sign language using virtual signing avatars, but the technology is still at research stages 16 17.
As noted above, sign languages are different to spoken languages and so automatic translation is not straightforward.
16 IBM's SiSi virtually translates speech to sign language: http://www.engadget.com/2007/09/13/ibms-sisi-virtually-translates-speech-to-sign-language/
17 RNID: Virtual signing: http://www.rnid.org.uk/howwehelp/research_and_technology/communication_and_broadcasting/virtual_signing/
