October 20, 2008
Audio text editor update, or postdate, or something
In July, I blogged an idea for a system that would make it much easier to do quick an dirty edits of podcasts. The basic idea was that the software would convert the speech to text and then let you edit the text, using the revised text (with hidden time-codes) to cut and paste the original recording. I thought it was a good idea.
So did Ryan Shaw and Dan Perkel … three years ago. I just got an email from Ryan (responding to my talking about this idea in the issue of my newsletter I just sent out) saying that he and Dan put together a prototype for a class at Berkeley. He points us to a brief description, some slides, and a prototype that he says is “probably broken.” The description describes a different, and interesting, facet of the project, but, Ryan writes in his email, “Edits to a transcription text in a browser-based editor were translated into edits to an underlying audio SMIL file, playable in RealPlayer.”