From first-hand experience, I think voice recognition technology is ramping up to usable levels. Having to "train the system" has always been a painfully slow process. And, the results haven't been too hot, either.

On the other hand, Wizzard Software's conversational model approach has eliminated the need to train the system on a particular voice and is getting pretty decent recognition.

Even though Wizzscribe SI requires some fairly stout hardware to operate, I'm anxious to see how this technology progresses, with the new crop of duo-core, duo-memory processors, now coming on line.

Rob Reilly is a consultant, trend spotter, and writer. He is a contributing editor for LinuxToday. He advises clients on mobile business computing and presentation technology integration. You can visit his web page at http://home.earthlink.net/~robreilly.

« Back: Hearing Voices