Speech recognition programs have been around for years and have made great strides recently. IBM, Dragon, and Lernout & Hauspie have all offered programs for dictating to a PC.

Nobody uses those programs, for a lot of reasons – technical and social.

Microsoft has been spending gazillions of dollars on research into speech recognition but until now has not included very much of it in their products. (There’s some speech recognition in Microsoft Office XP, but Microsoft doesn’t particularly promote it.) Now Microsoft unexpectedly released a beta of “Speech Server,” a different approach to speech recognition. Instead of running on a desktop PC, it’s a back-end product running on big servers for big companies. And instead of being aimed at desktop PC users, it’s focused on cell phones, pagers and other hardware devices where keyboards aren’t convenient, as well as allowing call centers to reduce costs and provide a better experience before you reach a live person.

Here’s an article about Microsoft Speech Server, as well as a short discussion about why speech recognition is so darned difficult to get right.

Share This