 |
The install was trivial; I was able to install and configure it on two separate computers, a laptop and a desktop, within a few minutes. Both machines are running Windows 7 Professional Service Pack one. The desktop computer that I’m using the software on, with a reliable if unspectacular AMD Opteron 1300, is in fact a little under powered given the 4 GB recommended RAM for Windows 7, but it is quite usable with only 2 GB. The laptop I am using, an aging HP PavilionDV9000, has 4 Gb of RAM but it’s feeble Turion64 Processor is hard pressed to keep up with the demands of the software. This was really the only negative experience I had with the software, and in my opinion, the difficulties I was experiencing were an artifact of the laptop I installed it on, and not the software itself.
Because I’m able to use the remote microphone to control my computer from anywhere within the same wireless network, so not only am I not limited to my home office, I am not even limited to my home. I can wander anywhere within the limits of the broadcast range of my wireless network access point. I did notice that in environments with multiple wireless LANs to connect to, it is pretty easy to have your iPhone on a different network than your laptop. Especially if they are cached profiles for the network on each device.
Training the computer to recognize one’s voice is a relatively straightforward process. The key to the process is the creation and development of a user profile. These are profile is a data set that characterizes the way an individual user speaks and uses languages. There are even settings to account for sex, age, and regional speech and dialect differences. The first step in the creation of a profile involves having the software determine the user’s microphone’s volume and clarity. Once the software has calibrated itself to one’s particular microphone and speaking style, the user is offered a set of readings designed to teach the computer how one speaks, the way that one uses language, and the manner in which the different phonemes are formed by a given user. For the laptop, I chose an excerpt from “2001 Space Odyssey” by Arthur C Clarke, and for the desktop, I used “Dogbert’s Top Secret Management Handbook” by Scott Adams. This process took approximately 15 minutes to complete including 10 minutes to search through my documents and e-mails. I further invested another half hour or so reviewing a decade or so’s worth of writing in order to further enhance the software’s accuracy and efficacy.
After training the software to recognize my voice and speaking patterns, learning to use this version was relatively straightforward given the intuitive interface and the massive improvements made in the speech recognition algorithms. This software can run in several different modes: Normal, Dictation, Command, Numbers. In Normal Mode the software recognizes speech as both commands and spoken text to be transcribed. In Dictation Mode, the software interprets all spoken data as if it were text to be captured and transcribed. In Command Mode, the software interprets all spoken data is commands to control one’s computer. One can switch between the modes by choosing modes from the software’s menu using the mouse or one’s voice, or one can simply say the command “blank mode on”, were blank is replaced by the name of the mode one wishes to use. It takes a little bit of getting used to, but once one gets the hang of it, it becomes almost second nature and is actually quite a bit of fun.
In my experience using the software for voice recognition in order to capture voice status text, is significantly easier than attempting to use the software to control one’s computer. That being said the most challenging implementation for this reviewer was using the software to create and manipulate spreadsheets using Microsoft Excel version 2010. Another challenge is using the remote microphone application in noisy environments, this despite the fact that the iPhone contains a built-in secondary microphone to measure and attempt to ameliorate background noise.
In order to test Dragon NaturallySpeaking version 11.5, and compare and contrast its speech recognition capabilities with that of its junior iPhone app cousin, Dragon Dictation and the built-in Windows speech recognition software bundled with Windows 7, the reviewer used Lincoln’s Gettysburg address as a sample text.
My attempt to transcribe this text by hand using “touch typing” took 12 minutes and resulted in 21 errors. The following is the result from using the iPhone Dragon Dictation App. The dictation took 31/2 minutes, and required stopping four times to permit the software to analyze and transcribe the voice data.
As can easily be seen, the iPhone app doesn’t easily allow for formatting the paragraphs, although it does permit one to use punctuation marks in the same manner as other Dragon products, by saying the name of the punctuation mark; for example, “comma” and “period”. There are 9 recognition errors (in UPPERCASE) in this attempt.
The next example is from the built-in voice recognition software that is bundled within Windows 7. The Windows software was not given as extensive an opportunity to with my voice, this is in no small part due to the fact that Windows lacks the ability to import documents and e-mails that give the software some insight into the speech patterns and vocabulary of the individual user. I did 3 rounds of training with the Windows Speech Recognition.
“Four score and seven years ago…
There are 222 recognition errors including 219 missing words. This is out of a total of 271 words. The main problem seems to be that the Windows speech recognition software uses an intermediate application to capture the voice data and then paste it into whatever application you’re using, in this case Microsoft Word version 2010. The result is that the Windows software was completely unable to compete even with the iPhone app, let alone does Dragon NaturallySpeaking version 11.5. The dictation took a total of 2 min. and 17 seconds.
The final examples are a test of Dragon NaturallySpeaking version 11.5, the first, using the remote microphone iPhone app., the second the Andrea headset. I did not evaluate the digital recorder functionality of the application.
First the remote microphone iPhone.
“IN four score and seven years ago our fathers brought forth on this continent a new nation, conceived in liberty, and dedicated to the proposition that all men are created equal.
There are 5 recognition errors including 2 missing words. The dictation took a total of 2 min. and 19 seconds.
Next, using the Andrea headset provided with the software.
“Four score and seven years ago our fathers brought forth on this continent a new nation…”
There are 2 recognition errors including 1 missing words. The dictation took a total of 2 min. and 21 seconds.
Conclusion:
Bottom line Dragon NaturallySpeaking version 11.5 is awesome software, not only does it have phenomenally good speech recognition, it is also extremely fun and relatively easy to use and quite honestly the title “NaturallySpeaking” is incredibly apt. I would strongly recommend this software to everyone from nerds who like cool things, to students, teachers, businesspeople, lawyers, and physicians, in short, anyone who generates large volumes of text data. And of course it is perfect for people with actual limitations using a standard keyboard and mouse. For example, those suffering from repetitive motion injuries, or those unable to use their hands for whatever reason, amputation, paralysis or temporary incapacity. On the downside, at least on my ancient and groaning laptop, it was a serious resource hog, at one point using almost a gigabyte of RAM and 53% of the available CPU cycles. It can cripple a machine that only marginally meets the minimum requirements. That being said, with a reasonably robust PC, the benefits of utilizing the software vastly outweigh its reasonable costs in terms of effort, outlay and system resources. This software is rated “Editor’s Choice”
A giant pain in the ass is an understatement. Give the tech another 10 years and maybe the giant part of the pain will be gone. Enjoy that software, after the therapy session dealing with your newly out of control rage.