DRAGON

I used to work for the company that made it (it was Dictaphone, then they were bought by Nuance Communications).

Anyway - yeah - training is key. You need to spend some time with the software for it to get things right/right-er.

Men from the Midwest do the best. Women from the Northeast or South tend to do the worst. Be aware that you might elide words or add a little extra (um, uh, that sort of thing) to the ends of words or you may clip them. Try to allow a second or so between sentences, too.

It is hard to recognize speech
can turn into
It is hard to wreck a nice beach.

I have a speech pattern that is purposely simple. So far, with the exception of the techy words, its been relatively easy to find and correct any excursions in word. I want to recite the "Pledge of Allegiance" sometime and see what I get.

The earlier Vr software was pretty much garbage in that the programmers were often ESLers . My previous outfit had several versions of some SR programs and they were all lacking in accuracy in transcription. Consequently, many people who used the VR programs for report writing would often think they could transcribe a few pages and then take a break and come back later. Many times the transcription errors were so bad that the report writers couldnt reconstruct what their main ideas were.

DRAGON has a net "place hold" feature

Fastest thing I do in life is type so that I have no use for voice software on a computer but nobody enjoys typing on smartphone screens and Dragon is the way of the future for those.

Within my experience, Dragon is heads and shoulders above competing products. As far back as the late 90s when computers gained the power to run Dragon software reasonably, I've ended up setting it up for less computer-savvy friends and was fairly well astonished by it, a couple of the friends speak pure Brooklynese and the damned thing was picking THAT up close to 100%.

fman - you're using it right now? Some of our kids in school have tried using this program, but it doesn't tend to work out. I think the effort it takes to build a working balance takes too long and is too frustrating for kids.

probably. I have to recall to stop slurring my own speech when Im transcribing, otherwise I will have two different spellings and hosting of words like "probably" It will show up as both Prolly and Probably and its a bitch to get out when you are speaking at the speed of thought.
Still, its kinda cool for adults who are sitting there in front of a blank screen and no idea what to type first.

is there a word bank you can build? Like, could you make a dictionary of techy words that it can then use correctly?

yeh, you can compose a glossary of terms and anotherof unique .instruction/ They give you a tips cd

So you just have to spell them right once! Pretty cool.

They are in the process of rolling this out across our office. I'm not up for a while since I do more auditing than report writing.

I've been watching/listening the first batch of users and they're quite happy with it. I guess Dragon understands most Canajuns fairly easily.

What I've been finding entertaining as that each user starts with it - their volume gets much louder. A few weeks ago, we were all hearing about floods as that group started their use of Dragon.

With the version I use, I can set up multiple user definitions and give each one unique voice training. I use this with some of my repeat transcription clients. I trained each of their user definition profiles with their old audio files and text from past transcript sessions. All I need to do now is have Dragon do the first draft transcription of new audio files for a specific user and then I manually review and correct it while listening to the tape for the first time.

It is an older version, 9.0 Professional. Not sure if that feature is still available in the newer versions for home use. Let me look at their site and see what I can find.

Yep, the home edition of ver. 11 is a single-user version. I checked what they're calling their premium edition and don't see mention of multiple user definitions. I also don't see anything about a professional edition so they may have eliminated it from their selections for individuals.

Looking in their business section, their Legal version appears to offer the multi-user profile feature.

Geesh, their prices have gone through the roof...

I can only use Dragon for transcriptions of audio that does not have music in the background and only has one speaker. Have tried transcribing interviews and it took more time to correct Dragon's errors than to just manually transcribe it myself. Dragon absolutely hates music. It tries to transcribe each note into language, with sometimes hilarious results.

Ive got this baby set up for field work now. Ive been checking some "CREEP" stations in an arond the Plaski and Chatanooga areas and I jst walk around with a little netbook in my field vest and with my phones on I can dictate what I see and it gets almost 99% of what I say. (Ive learned to trim my beard and stache when I go out so I dont sound muffled).

I really dont see a need for multiple users cause its so cheap and its best to have the system configred to what Im doing rather than networking.

Awesome, glad it's working well for you.

I'm always interested in voice recognition software from the instant-captioning side of things.