Dragon NaturallySpeaking 12 Shows Why It Rules Voice Recognition Market
The latest version of Nuance's speech recognition software impresses with its Windows integration, application support and transcription capabilities. But which version is right for your business?
By Paul Mah
Speech recognition software is of tremendous value to many end users, including those with physical disabilities or injuries that make it difficult or impossible to use a keyboard and mouse. If nothing else, most people can speak much faster than they type, which makes speech recognition a great tool to get words into a document.
Nuance Communications’ Dragon NaturallySpeaking is the company’s flagship voice recognition product for PCs. Dragon NaturallySpeaking 12 Professional, released late last year, is the newest version. Here’s a closer look at this speech recognition software, its new features and how they could benefit your business.
According to Nuance, Dragon 12 is 20 percent more accurate out of the box compared to version 11, with more natural text-to-speech reading. Moreover, performance has improved thanks to support for computers with multiple cores.
Usability improvements include additional personalization options. For example, you can set Dragon 12 to require users to say “click” to denote a user interface action. Individual users who don’t like this feature can disable it.
In addition, Dragon 12 now works with common applications and devices. For starters, it comes with full text control and commands for WebMail services such as Gmail and Hotmail. It also recognizes the most frequent commands for the Internet Explorer, Mozilla Firefox and Google Chrome browsers. Posting updates to Facebook or Twitter with simple voice commands is supported, too. Finally, the free Dragon Remote Microphone app is available to users with an iOS or Android device.
I evaluated the Professional edition (599.99) of Dragon NaturallySpeaking. Along with the voice-recognition improvements and overall enhancements listed above, this version comes with various features not found in the Home ($99.99) or Premium ($199.99) editions. You can see a full breakdown of the differences here.
The Professional edition supports profile administrative capabilities geared towards enterprise deployments; it will also function over Remote Desktop Connection from a local Windows computer on which Dragon is not installed.
Individual power users, though, will be more interested in its capability to create custom commands to automate frequent or repetitive tasks. The toolset includes Macro Recorder, VBA-style scripting or the organizing of custom commands in custom groups. Most importantly, Dragon Professional comes with transcription tools such as an auto-transcribe folder agent capable of performing transcription in the background. (More on this later.)
Dragon NaturallySpeaking 12 Installation and Initial Setup
On the hardware front, Nuance recommends a PC equipped with 2GB of RAM and a 2.2 GHz Intel dual core processor. A minimum of a 1GHz Intel Pentium or 1.66 GHz Intel Atom processor is required; the installation software will not install if the minimum requirements are not met. The specifications are probably not an issue with newer laptops—just don’t expect to use it on a first-generation netbook or 5-year-old laptop.
The boxed review set I received from Nuance came with DVD installation media, a Quick Reference Card and a wired microphone with single-sided ear cup. Dragon requires 3.2GB of storage space for the English version. As using the laptop’s built-in microphone with Dragon was explicitly discouraged, so I used the bundled USB-based microphone for all the tasks in this review.
Installation was a relatively straightforward affair that entails popping in the DVD disc, keying in the serial number and letting the installation run its course. Once installation is completed, Dragon creates a user profile for you. This entails some reading exercises that allow the voice-recognition software to better adapt to your speech. More importantly, the user profile will allow the application to associate words unique to you, such as nicknames, jargons and capitalized phrases. Users who have used Dragon 10 or 11 will see an option to have their user profiles upgraded.
Configuration isn’t a task that should be rushed or skipped. I spent slightly more than 10 minutes completing some reading passages, after which Dragon took a few more minutes to process the captured recording.
Using Dragon NaturallySpeaking 12 Professional
Instead of trying to identify each spoken word in isolation, Dragon improves accuracy by listening one sentence at a time in order to better guess the context of the spoken words. This may be somewhat disorientating to first-time users, though it does provide dramatically better results.
Some mistranslations do appear, such as the numeral “12” appearing as “trial.” (It may be better to read it out as “one” and “two.”) Overall, I found the speech recognition to be very good, with common mistranslations avoidable with some experience.
I was more interested in testing Dragon’s Windows-level integration, which offers the promise of great productivity. On that front, I found it possible to get the hang of common commands relatively quickly.
For example, the “Switch to” and “Start” commands will switch to the appropriate running application and launch the specified application within the Start Menu, respectively; examples include “Switch to Excel” or “Start Word.”
Commands are generally intuitive and include “Show Desktop,” “Maximize Window,” “Minimize Window” and “Close Window.” Standard dialog boxes are also easily navigated with “Click OK” or “Click Cancel” commands.
On the other hand, the sheer number of commands to remember can make mastering Dragon nontrivial. That’s why the Quick Reference card, which highlights common sample commands for working with Dragon, is especially useful. Another handy tool is the “Dragon Sidebar,” which includes a comprehensive list of usable commands that can be referenced on screen.
Finally, there are tasks where using Dragon is actually slower or not usable. For example, while it is possible to use voice to move the mouse around the screen with the “Mouse Grid” command, it’s probably much faster for able-bodied users to use a physical mouse. In addition, multi-monitor users will find it challenging to reposition their windows over more than one monitor.
Overall, it is entirely possible to control Dragon without touching the keyboard and mouse on a standard non-multi-monitor PC, though it does require due diligence to learn—and practice—the various commands. Ultimately, the learning curve means that it will take a while to actually do things more quickly with voice commands. Nuance provides the comprehensive reference materials to make things easier as easy as possible.
One powerful feature found only in Dragon Professional is its offline transcription capability. In a nutshell, Dragon can transcribe the contents of prerecorded audio files into text. This is useful for folks who spend a significant amount of time commuting by car. They can now record their thoughts on a digital recorder and have them transcribed upon reaching the office.
Besides the capability to transcribe single audio files, Dragon can also perform batch conversion of multiple files. This can be selected under “Tools > AutoTranscribe Folder Agent&” From here, it’s possible to assign multiple input folders for the software to monitor for new audio files to transcribe.
Input folders can reside on shared drives on the network, with each input folder tagged with a specific Dragon user profile. This makes Dragon Professional an excellent tool for large projects with multiple employees that may involve transcribing large amount of audio recordings to text. In the latter case, transcribed text will appear as a continuous flow of text unless the audio recording includes vocalized punctuation such as “full stop” or “comma.” Aside from that, I found the quality of transcribed text to be stellar.
Of course, Dragon does take a significant amount of processing capability when transcribing. I was able to continue using an Intel i5 laptop while transcription was in process, though it did become noticeably warmer during the process, and battery life also took a hit.
Conclusion: Weigh Options Before Upgrading to Dragon 12
Dragon NaturallySpeaking 12 continues to push the envelope as the best full-fledged speech recognition software on the market. Users looking to implement voice recognition as part of their work won’t be disappointed in its breadth of capabilities, while its deep integration with Windows and popular business-critical applications has been further broadened.
For all its appeal, though, Dragon 11 users may find it a stretch to justify upgrading to Dragon 12 unless they use Dragon heavily as part of their daily workflow and need the new features found only in Dragon 12. For these users, Nuance offers an upgrade package of $299.99 for users on Dragon NaturallySpeaking Pro v10 and up.
Ultimately, individual users may want to evaluate if they need advanced features, such as the capability to create custom commands or access to offline transcription tools, before settling on Dragon 12 Professional. If not, then Dragon Premium 12, which offers almost the same feature set and voice recognition engine, may be a better bet.
Paul Mah is a freelance writer and blogger who lives in Singapore. Paul has worked a number of years in various capacities within the IT industry. Paul also enjoys tinkering with tech gadgets, smartphones and networking devices. You can reach Paul at firstname.lastname@example.org and follow him on Twitter at @paulmah.