Talking to Yourself Productively with Dragon Naturally Speaking

By Frank Ohlhorst  |  Print this article Print

Version 10 of Dragon Naturally Speaking brings new features to those who like to talk themselves.

Although voice recognition and dictation software has been around for quite some time, it really has never taken off as one would expect. While the reasons may be hard to pinpoint, we can assume that it is not due to a lack of technological progress. Case in point, Nuance software has successfully released several versions of its Dragon Naturally Speaking software, each version successively better than the previous. Now, with version 10 hitting the market, Nuance is hoping that the latest iteration of Dragon Naturally Speaking will do the trick.

Released on Aug. 7, Dragon Naturally Speaking version 10 includes several improvements and enhancements that advances the state of speech-recognition software. The company is banking on the fact that potential users have gotten used to speaking to their devices. After all, we now have voice dialing on cell phones, voice recognition built into new automobiles and voice recognition used by automated call-in centers.

Arguably, one of the biggest frustrations with using voice recognition software has been the need to train the application to properly recognize your voice. That usually meant that a user would have to go through a long-winded process before even being able to use the voice-recognition technology. That is one area where the new version of Dragon has made significant improvements. The product claims to be more than 20 percent more accurate than previous versions and in turn allows users to get started right away. In most cases, if you have a clear speaking voice, Dragon Naturally Speaking version 10 is able to perform dictation chores right away and is able to learn the speaker's voice the more the product is used.

That enhancement allows users to become immediately productive with the product. Although in our testing the product did seem to perform slightly better if the user went through some of the simpler training chores. That said, we also had great success using the product right out of the box and were impressed with the accuracy and the product stability to recognize a new voice without any training.

The improvement in accuracy is only one small part of the story with Dragon Naturally Speaking 10. Nuance software has also improved the performance of the product, which has greatly reduced the latency found in previous versions. That performance enhancement allows users to speak faster and see the results much more quickly on the screen. That performance improvement also has an unrealized benefit--most speakers wait for the words to appear on the screen before continuing, which introduces an artificial delay in how quickly a user can create a document by dictation.

With version 10's increased performance, that delay is eliminated. While speed and accuracy are very important enhancements, Nuance did not end just there. Version 10 also supports a new interface that makes the product more intuitive. There is also an improved help system, extensive tutorials and the ability to recognize regional accents. Those accents are broken down by either geographical locations or cultural influences on how English is spoken.

The new look and feel of the product also incorporates better voice-command support for many popular applications. Users can now use Dragon to process e-mail, initiate searches on the Web, execute commands in many Microsoft office products and perform enhanced voice-formatting chores. For example, it is now very easy to select text in a document and reformat it, change it, replace it or even change fonts and other elements. The product does this all in a more natural fashion. Users can now select phrases, individual words, paragraphs or other text elements in a more natural fashion and make any changes verbally.

The new voice commands and capabilities allow users to surf the Web, create e-mails, launch applications and sell on with much more ease than ever before. Version 10 seems to just be more intelligent than any other voice-recognition product to date.

Users will find installation of the product straightforward, but be forewarned--the product does need to be activated and registered for the system that it is going to be installed on. The company does allow four separate activations, so that users can install the product on their work PC, notebook computer and home computer. Once installed, users can choose to go through the recommended training process or just dive right in. Untrained, the product does a bang-up job of recognizing the user's voice. The product does offer several training methods, where the user reads provided text and controls how in-depth the training session is.

Dragon Naturally Speaking version 10 supports several different input methods and the product includes a headset which can be plugged directly into the soundcard ports on a PC, which in most cases will be how the typical user runs the product. Users can also record dictation to a voice recorder or MP3 player that supports recording, or other device and then import the data file into Dragon Naturally Speaking for conversion. The product also supports a batch mode, which allows several MP3 (or other sound files) to be placed into a directory and then automatically be converted into text. That feature comes in handy for those who like to dictate into an electronic device.

The product is offered in several versions, some of which are for specific markets, such as legal or medical and some of which are intended for general markets. The two primary version of the product includes the $899 Professional version, which supports multiple network users and includes the new series of "natural commands" for controlling applications, and a $199 preferred version, which is aimed at the home- or small-office user. The company also sells bundles that include the wireless Bluetooth headset; those are differentiated by the word "Wireless" being added to the product name and add about $150 to the price. The company also offers discounted upgrade prices for users of previous versions of Dragon Naturally Speaking.

All things considered, the latest version of Naturally Speaking is a step in the right direction and may just get people talking to themselves and, more importantly, to their computers!

Frank Ohlhorst Frank J. Ohlhorst is the Executive Technology Editor for eWeek Channel Insider and brings with him over 20 years of experience in the Information Technology field.He began his career as a network administrator and applications program in the private sector for two years before joining a computer consulting firm as a programmer analyst. In 1988 Frank founded a computer consulting company, which specialized in network design, implementation, and support, along with custom accounting applications developed in a variety of programming languages.In 1991, Frank took a position with the United States Department of Energy as a Network Manager for multiple DOE Area Offices with locations at Brookhaven National Laboratory (BNL), Princeton Plasma Physics Laboratory (PPL), Argonne National Laboratory (ANL), FermiLAB and the Ames Area Office (AMESAO). Frank's duties included managing the site networks, associated staff and the inter-network links between the area offices. He also served at the Computer Security Officer (CSO) for multiple DOE sites. Frank joined CMP Technology's Channel group in 1999 as a Technical Editor assigned to the CRN Test Center, within a year, Frank became the Senior Technical Editor, and was responsible for designing product testing methodologies, assigning product reviews, roundups and bakeoffs to the CRN Test Center staff.In 2003, Frank was named Technology Editor of CRN. In that capacity, he ensured that CRN maintained a clearer focus on technology and increased the integration of the Test Center's review content into both CRN's print and web properties. He also contributed to Netseminar's, hosted sessions at CMP's Xchange Channel trade shows and helped to develop new methods of content delivery, Such as CRN-TV.In September of 2004, Frank became the Director of the CRN Test Center and was charged with increasing the Test Center's contributions to CMP's Channel Web online presence and CMP's latest monthly publication, Digital Connect, a magazine geared towards the home integrator. He also continued to contribute to CMP's Netseminar series, Xchange events, industry conferences and CRN-TV.In January of 2007, CMP Launched CRNtech, a monthly publication focused on technology for the channel, with a mailed audience of 70,000 qualified readers. Frank was instrumental in the development and design of CRNTech and was the editorial director of the publication as well as its primary contributor. He also maintained the edit calendar, and hosted quarterly CRNTech Live events.In June 2007, Frank was named Senior Technology Analyst and became responsible for the technical focus and edit calendars of all the Channel Group's publications, including CRN, CRNTech, and VARBusiness, along with the Channel Group's specialized publications Solutions Inc., Government VAR, TechBuilder and various custom publications. Frank joined Ziff Davis Enterprise in September of 2007 and focuses on creating editorial content geared towards the purveyors of Information Technology products and services. Frank writes comparative reviews, channel analysis pieces and participates in many of Ziff Davis Enterprise's tradeshows and webinars. He has received several awards for his writing and editing, including back to back best review of the year awards, and a president's award for CRN-TV. Frank speaks at many industry conferences, is a contributor to several IT Books, holds several records for online hits and has several industry certifications, including Novell's CNE, Microsoft's MCP.Frank can be reached at frank.ohlhorst@ziffdavisenterprise.com