The transcription of recorded history

Yesterday TechCrunch revealed some interesting news about he WB, the popular teen-focused television station, who recently did a complete overhaul on its website. Powered by Digitalsmiths, the site's search engine features the ability to search the station's episode video library for character dialogue. Now, anyone can dig up a specific episode based on what the characters were talking about. And it's all powered by voice recognition software.
This technology will knock the socks off of any Seinfeld fan who can't remember in which episode they say "Schmoopy," "Serenity now!!!,"You're a very very bad man," or "The sea was angry that day, my friends."
I always pictured voice recognition software (VRS) to be a way of the future, but mostly with new media, never envisioning its impact on translating the world's recorded history. My Civic has VRS which I use to control my GPS hands-free while driving; I didn't picture it to go much futher than that. Google's mission is to organize the world's information and make it universally accessible and useful. So will their next project be to use their computing power to transcribe our diverse library of spoken information into searchable databases? They're already doing it with newspapers -- why not with news broadcasts?
The spoken word is found in so many places: movies, television, radio shows, the congressional floor, college lecture halls. How neat would it be if anyone could digitally record audio, upload it to an online translating engine, and receive an instant transcript? Translation errors could be manually corrected by the user, only making the engine smarter as time went on.
All of the sudden, instead of searching the internet for content that people have taken the time to sit down and write, you could search the things that people have stood up and said -- credible people, too. Imagine the technology's effect on college students who no longer need to take notes or go to class! All they'd need to do is nurse their hangovers at home and read a lecture's transcript.
This would probably be a nightmare for politicians. No longer would politi-sleuths need to dig through hours of video to find evidence of policy flip-flops. With a snap of the fingers, one could find every single thing that John and Barack said into a microphone about healthcare or Iraq...good luck, you two!
Some new technologies can be borderline creepy -- like cloning and killer robots. Despite their obvious differences, I would place mass automatic voice transcription right up there with the creepiest. Useful? Definitely. Comes along with limitless latent effects? Absolutely.
Labels: audio, creepy, microphone, news, radio, television, translation, voice recognition


1 Comments:
At August 27, 2008 11:33 AM ,
Steve said...
First of all, I love the 'creepy' label. Nice touch. I hope someone googles 'creepy' and finds your blog.
I would be really surprised if Google has not been considering this for some time. I think you're right - making all recordings that have ever been made searchable is probably going to happen in the relatively near future. I know Google and others have been working on improving picture and video search for some time - currently, search engines really just look at text around an image and assume there is some kind of connection there. This seems like a logical extension of their efforts to make the content of pictures and video searchable.
Steve
stevescookingjournal.blogspot.com
Post a Comment
Links to this post:
Create a Link
<< Home