HUGE: Adobe Readies Voice-to-Text Metadata For Flash Video

PALO ALTO -- Beginning later this year, Flash video will contain rich metadata in the form of  automated transcripts.  Adobe is readying a system that uses speech-to-text technology directly into the production process. 

Adobe's plans were revealed yesterday by Adobe VP Jim Guerard, who was a featured speaker at the Beet.TV Executive Roundtable held on the Stanford Campus.

If the Adobe project is widely adopted, it will have a profound impact on how Flash video is searched and discovered.  Right now there is very little text data in Flash files.  Recently YouTube began including voice-to-text generated metadata into some of the clips in the presidential campaign.

At NAB in May, I interviewed Adobe's top Flash strategist Mark Randall on the topic of voice-to-text and its implications.  I've reposted that video here. Mark is speaking here at the AlwaysOn conference and I hope to catch up with him.

The Beet.TV event was great.  Here's a recap from CNET's Charlie Cooper. We have tons of videos from yesterday's conference.  Below, (L-R): Charles Tillinghast, president of  msnbc.com; Joan Walsh, editor-in-chief of Salon; YouTube's news and politics head Steve Grove; Andrew Heyward, former president of CBS News and now a consultant with the Monitor Group.  Andrew co-moderated the event with me.

--- Andy Plesser, Executive Producer