Cool Speech Project Alert: Making Webcasts Available to the Deaf

IBM Research's CaptionMeNow project definitely falls into the "cool project" category. The idea is to automatically caption webcasts for people with hearing impairments by running the audio through a speech recognizer. Well, its almost automatic. They still have a human in the loop to make minor adjustments to the transcription produced by the speech recognizer.

We've reported on a similar approach in the past where speech technology is used to index video content, based on the audio. The CaptionMeNow article also suggests this technology could spill over into mainstream usage, as some people can more efficiently read and retain information in text vs. audible speech.

The author also claims TTS technology emerged from efforts to make written text more accessible to the visually. While I know TTS is commonly used by the blind, I didn't realize that it was the primary motivation behind early research in this area. In any case, I vividly remember my first encounter with a speech synthesizer as a teenager. My entire family assembled spellbound around my trust Commodore-64 when I first loaded up "Sam the talking computer" on the world's slowest disc drive (remember the Commodore 1541 floppy drive?). It was simply astonishing to realize the computer could intellibly say anything we typed in. Now everybody takes this stuff for granted.

Read the article about CaptionMeNow.