Nuance announced Release 9 of Dragon NaturallySpeaking 9 in July 2006. DNS 9 Press Release
Nuance understands this is fundamental to customer adoption and satisfaction. Claimed is 20% fewer errors. In retrospect, this was probably exceeded for medical dictation using DNS 9 Medical. Other upgrading users did not generally see the same degree of change.
Creating a user includes a training session where Dragon
NaturallySpeaking learns how you speak. Initial training includes
reading aloud for several minutes from a list of available texts. You
can skip the initial training. However, completing the training
before you start dictating will increase your initial accuracy. You
can only skip the initial training of a users created with headset
microphones. This option is not available for users created with
array microphones, BlueTooth microphones, or portable devices.
See analysis, this basically means "Select and Say" works in more applications than before.
There is explicit support for Bluetooth microphones.
For applications that don't support Select and Say (and for special cases such as long Word documents) this provides much more flexibility in being able to dictate then transfer text into other applications. Nicely done!
This is done in several ways, mainly "under-the-hood." These improvements will help some more than others based on past experiences. There are more 2, 3, and now 4-word phrases (bigrams, trigrams, and quadgrams(?)) known to DNS 9. In some contexts, DNS 9 does a better job of "looking" at the document/email you are dictating and betting that you will use the same words that already exist in that document. But it is too early to tell how significant this change may be.
What is now obvious is that "out-of-the-box" accuracy is significantly better for medical dictation when using one of the Medical language models. This is not obvious when using the General English language model for medical dictation.
8 minutes is cut off the setup time. That's worth 8 minutes, perhaps repeated a couple of times per year. Nice, but not really a big deal and it is recommended to go ahead and enroll anyway. It will be nice for us selling the product as we'll be able to let people try it out quickly. One warning: if you choose not to enroll, you still have to read the scripts for the Audio Setup Wizard and you need to read them precisely as they serve as a mini-enrollment session.
This is a technical change but means that DNS will support Select and Say in many more applications. It won't help in Microsoft applications, but will make the experience in many EMR's and other applications smoother. It shouldn't hurt anyone.
TE Edit Control from Sub Systems, TX Text from TX Text Control, and Tmemo, Tedit, and TRichEdit from Borland are now supported. What does that mean? To most of you, it means maybe your applications will work better and maybe they won't. There is a new "Select and Say Indicator" that is one color in a Select and Say box, a different color when Select and Say isn't enabled. That's a nice added touch on the DragonBar.
This dictation box may allow the "Show Dictation Box" to work in many cases where it would not work correctly on Releases 7 or 8. You can adjust type font and adjust how to paste text into applications. You can have different paste methods or font for different applications. It probably solves the problem of moving the mouse while in the dictation box and having the paste go to the wrong place.
For some people, this will be a big deal if the added audio format support is used. WAVE PCM, MS ADPCM, IMA
ADPCM, a-law, mu-law, VOX and MP3 are now supported for transcription.
Higher RAM and CPU requirements are needed to support the changes. Minimum per Nuance is 1GHz, 512MB RAM and recommended is 2.4GHz, 1GB RAM. If using for heavy production use, I'd move to a dual core system with 1GB or even 2GB. After all, those systems are less expensive now than NaturallySpeaking Pro/Med/Legal products.
OK, this may not be an improvement. The manual has been shrunk, with the material formerly in Chapters 1-6 and 8 retained. It appears the rest is relegated to Help files. Some may like the smaller format, but I've already lost it once on my desk and have only had the product a few hours.
There are additional sections in several Help files, in some cases finally documenting features and functions that have been available at least since Release 6.
This book is available as of December 2006. Because the Help files included in the product are now more complete, parts of the book may be deleted or simply refer to the appropriate Help section. The Scripting for Dragon NaturallySpeaking 8 book remains applicable for DNS 9 scripting.
A conservative approach to upgrading is:
Compared to DNS 8; DNS 9 Preferred is same price, DNS 9 Professional increases price by $100 to $899, DNS 9 Medical and Legal increases price by $200 to $1199 (List Prices).
In early 2007 a DNS 9 Small (Medical) Practice Edition was introduced, list price as of November 2007 is $1199. November 2007 there was an additional price increase to $1599 for DNS 9 Medical.
Preferred 7/8 to Preferred 9 upgrades are $99.
For pricing on new products (full-version) see the Products section. For upgrade pricing, see Upgrades.
top