Originally posted by: dakels
dammit I hit the clear button and deleted my post! 😛
short version:
Dragon Naturally speaking has a good recogniztion engine too and has low level consumer products. You will notice alot of the speech recognition software companies have different pricing level packages. Usually the low end packages are just a slimmed down version of the high end software. There are also special packages for medial or legal and such that will recognize special terms used in the respective field.
All decent speech recognition programs usually require a bit of "training" the proogram to get used to your voice. This can take a while depending on the package and how clear your speaking voice is. It also helps to have a decent microphone. Headsets are very handy, just make sure it has a decent mic reciever. In any event, you'll still find that these are nowhere near perfect yet and require alot of correction, again depending on the situation. Don't think it will be like some sci fi movie. Chances are you will have to carefuly watch what the program types in for a while. There are much better speech recognition systems that make very little mistakes and recognize a whole variety of vocal qualities but these are very expensive and are usually used by corporate, mdecial, or government uses. A good example of one you may have used already is say fedex or UPS's phone answering system. Those AI's are pretty good. I bet they'd even understand Marlin Brando 😛