David and Larry,

It may just be because of my 'souuthern' dialect, but I've never found any of these programs to work worth a hoot. Train, Train and more Training and I say hello, it spits out dskjhldhfldshlk. Tried both IBM and Dragon, wasted a ton of time. Just my $.01 :)

Jeff



>Larry,
>
>Most good speech-to-text systems are trainable so it understands each user. ViaVoice requires about a 30 minute training session where you read to it. The person just talks into the mike you don't have to record wav files.
>
>>These are all good ideas and we'll definitely look into them, but I was kind of hoping for a solution that did not involve playing the .wav file into a microphone.
>>
>>Also these .wav files are created by an Order Entry system that gets calls from all over the country, from people with all kinds of accents (shoot, I can hardly understand some of them myself!) Anyway, any other ideas would be greatly appreciated as well.