DNS 11.5 / Dragon Medical Practice Review

DNS 11.5 Review
(Updated on 7-18-11)

The Following Is a KnowBrainer Review of
NaturallySpeaking Ver. 11 (available at www.KnowBrainer.com)

 

Pre-introduction:

We originally intended to cover all the new features in our introduction but soon realized that even if we were to list the new features, without explaining them, it would take nearly a page because this version of NaturallySpeaking was almost entirely rewritten from the ground, up barely preserving 20% of the original code and incorporating more changes than all previous versions of NaturallySpeaking combined. Because there is so much ground to cover, we decided to create a short review, consisting of the following introduction and conclusion. If you haven’t fallen asleep by the end of the Short Review, feel free to continue with our more in-depth review which begins by covering the pros and cons as we experienced them. Please also note that this review will be subject to updates but whenever we make a change, we will revise the date at the top of the file. Additionally note that Dragon Medical 11 is not due to be released until early 2011. The following review currently covers DNS Home, Premium, Pro, and Legal Ver. 11 Editions and will be updated when Dragon Medical 11 is released.

Short Review

Introduction:

NaturallySpeaking Ver. 11 includes more new features and enhancements then all previous versions of NaturallySpeaking put together. For those who are accustomed to NaturallySpeaking Standard and Preferred, the names have been changed to NaturallySpeaking Home and Premium respectively. Besides a virtually new interface, this version of NaturallySpeaking includes amenities such as: a new pop-up “application/window-aware” Sidebar that displays a list of the most popular commands (which changes depending on your environment), significantly improved accuracy with some microphones (more info), notably improved learning from corrections, Best Match IV modeling, 22.5 kHz sampling rate (as opposed to the previous 11.025 kHz sampling rate), Office 2010, WordPerfect X5 and OpenOffice Writer support, serious revamping of the help files including numerous videos,  scripting support (for the professional versions of DNS), dual core multi-threading and a slew of additional features that you can view in the Pros section in our extended review.

Conclusions:

Windows Speech Recognition (WSR) is still the best bang for your buck because it is freely included with Windows 7 but if you required higher accuracy and notably greater efficiency NaturallySpeaking 10 was the only game in town. However, with the release of Ver. 11, the ante just got upped!

We highly recommend that everyone (including users of WSR), other than Windows 2000 users, upgrade to DNS Ver. 11 for both accuracy and productivity reasons. Ver. 11 adds numerous new features such as being able to simultaneously format every instance of a specific word in a document, faster Natural Language Command deployment, dual core multithreading (allows you to fire off a string of commands without having to wait for the 1st one to deploy), being able to train words from within the Spell That Correction Dialog Box (a feature that was previously available but didn’t actually work) and improved correction features. However, perhaps the most significant improvement is the one that Nuance themselves may have failed to fully appreciate.

Nuance claims up to a 15% fewer mistakes which we were unable to verify on low to mid-range microphones (including the manufacturer OEM microphone that is included with the Home and Premium Editions) but we were nearly shocked by a 30% decline in errors when using a high-end microphone such as the Samson Airline 77  or Sennheiser MD 431 II. We were only able to appreciate a marginal 10% or less error rate reduction on mid-quality microphones such as the Andrea Electronics ANC 700 and NC-181 (DNS Pro included manufacturer microphone). We were unable to appreciate any improvement with consumer grade low-end microphones and in one test, even experienced a decline in accuracy. Utilizing a high-quality microphone in previous versions of NaturallySpeaking made a significant difference in accuracy but in Ver. 11, it may be doubly important. If NaturallySpeaking is important to your workflow, we don’t recommend using the manufacturer included microphone or typical over-the-counter VoIP/gaming microphones such as those made by Plantronics, Logitech, Cyber Acoustics, Labtec and numerous others that claim to be adequate for speech recognition.

Although we still recommend the BlueParrott Xpressway bluetooth wireless microphone, we now have to qualify that recommend to situations where no other microphone would be appropriate, such as wireless mobile solution for a tablet computer. An alternative situation might be where an end-user requires an inexpensive wireless microphone that can pull double duty with their cell phone or an inexpensive solution for a netbook which can only handle Best Match III. Because Bluetooth microphones are limited to an 8000 Hz frequency response, by their very nature, they cannot take advantage of Best Match IV or the new 22.5 kHz sampling rate (12,000 Hz). If you don’t HAVE to use a Bluetooth microphone in Ver. 11, don’t. Because of the new sampling rate, we only recommend microphones with a fairly flat frequency response between 500 and 12,000 Hz.

Purchasing Note: If, after reading this review, you are as excited about NaturallySpeaking Ver. 11 as we are, we recommend purchasing NaturallySpeaking and speech recognition peripherals (microphones, digital recorders etc.) exclusively from Nuance certified/licensed VARs (value added resellers). When you purchase all of your speech recognition tools from a single Nuance certified source, you’re typically dealing with an expert whose combined knowledge of microphones, digital recorders and NaturallySpeaking supersedes other specialty vendors who are either limited to selling one type of speech recognition item, such as microphones or recorders or a jack of all trades box pusher like Best Buys, Office Depot etc. who are not qualified to answer NaturallySpeaking questions.

 

Another disadvantage of purchasing NaturallySpeaking from large resellers such as CDW, PC Connection, Amazon, eBay or directly from the manufacturer is that you only have the manufacturer’s technical support. When you purchase from a Nuance certified reseller, you not only maintain your right to manufacturer support but additionally obtain an extra layer of technical support from your certified reseller/trainer along with various other amenities your reseller may be offering ranging from step-by-step installation guides to virtual on-site assistance . It is in your best interest to support your local Nuance certified reseller.

For those of you who didn’t get enough information from the previous Short Review, we offer a more detailed review below:


Extended Review

Pros:

● Best Match IV with 22.5 kHz sampling rate as opposed to the Best Match III 11.025 kHz sampling rate of previous versions of NaturallySpeaking.

● Referring to the new interface as an enhancement would be a gross underestimation. The new interface has been rewritten almost entirely from the ground up featuring numerous easy to find menu improvements such as a new Audio menu.

● We experienced up to 30% fewer errors when using a high quality speech recognition microphone such as the Samson Airline 77 and Sennheiser MD 431 II. We also noted accuracy improvements with other high-end microphones such as the Philips SpeechMike AirGrundig Digta SonicMic II and Sennheiser MD 431 II but mid-quality microphones such as the ANC 700 and TalkPro Xpress turned in more modest gains of about 10% while over-the-counter microphones and consumer grade microphones (like those found in DNS Standard/Home and DNS Preferred/Premium Editions generated no accuracy improvements and in one test, even yielded lower accuracy.

● Introducing dual core multithreading which gives you the ability to deploy a string of commands without having to wait for NaturallySpeaking to catch up or dictate several paragraphs nonstop without overrunning the internal buffer. Multithreading also increases your accuracy by giving NaturallySpeaking more choices to choose from when comparing what you said to what it thinks you said.

● See following speech utterance history enhancements:

● Significantly improved accuracy when transcribing digital recordings. Instead of conducting this test the right way, we chose to take a DS2 recording, from our Olympus DS-5000 digital recorder, and transcribe it through an untrained microphone user profile which is a combination of 2 no nos. In theory, you should always create and train a specialized digital recorder user profile. Nevertheless, we were still able to appreciate 99% accuracy from the 3 paragraphs in our recording. Conducting the same test in DNS 10 produced about 97% accuracy. A 2% improvement in accuracy may not sound like much until you realize that you have to correct 3 out of every 100 words as compared to 1 out of every 100 or to put that another way, only having to correct 1/3 as many errors.

● In previous versions of NaturallySpeaking it took a minimum of 15 min. to create a user profile but typically took closer to 25 min. Ver. 11 requires 4 min. and typically takes about 6.

● Pop-up help sidebar (notable improvement over previous What Can I Say command) displays a list of the most popular commands for a specific application.

● Smart installation detects slower computers, such as netbooks and reconfigures the installation to Best Match III.

● Add Word now includes a spoken form option.

● Numbers each word in a document when you say <select, delete, bold, underline, insert before> <word/s> which gives you the option of choosing any single instance of the word you want to format or applying the formatting to each instance of the word/s.

● New and improved revamped interface will be great for new users but may throw up some temporary hurdles to previous DNS users who have memorized the previous interface.

● Seriously beefed-up help files including videos and even command writing assistance (for those using the professional versions of DNS).

● Easier to understand text-to-speech playback.

● DNS 11 improves your language model even when you correct your documents by typing. It’s also smart enough to discern a correction from a rewrite/edit.

● The Correction Dialog Box learns when you use the Train button. Previous versions of DNS included a Train button but only worked for training words from within the vocabulary editor.

● When training a correction, you no longer have to waste time training the misrecognition.

● New Spanish accent support (with US dialect).

● Improved scheduling of accuracy tuning (no longer requires Windows password).

● Faster at saving user files.

● Support for WordPerfect 15, OpenOffice Writer and Microsoft Office 2010 (32-bit and 64-bit).

● Improved training text that allows you to learn about NaturallySpeaking while reading the training. You’ll find more information on this new feature in our KnowBrainer Training Guide which we include in all of our NaturallySpeaking bundles.

● Improved Scrolling in applications such as Microsoft Word, Outlook, Excel and Internet Explorer.

● Less obtrusive Results Box options.

● Auto punctuation appears to be improved over previous versions of DNS but most end users will probably prefer to avoid using it. However, the new enhancements may help in specialty areas such as transcription.

● Smart punctuation such as going back to properly insert a comma to the left side of the last word even after manually pressing the spacebar. If you dictate a period in the middle of a sentence, NaturallySpeaking is smart enough to locate the next word and capitalize it in any Full Text Control (formerly Select-&-Say) enabled application.

● Smart Edit. Ver. 11 is smart enough to know that when you accidentally select part of a word with your mouse that you want the entire word to disappear when you re-dictate.

● Multithreading increases accuracy by simultaneously feeding your dictation through 2 acoustic models.

● Slightly improved Bluetooth support.

● Ver. 11 let’s you know if your accuracy falls below a minimum which is an excellent idea but about the 2nd time we experienced this error message, we put a checkmark in the box to never show the error again. Although we acknowledge that this is a good idea, from a personal perspective, we’re on the fence.

● You can start most apps with the word Open or Start plus the application name but you can now optionally say Launch. Example, if you wish to launch Microsoft Word you can now say Launch Microsoft Word. Of course if you’re using KnowBrainer 2010 you can shorten that to just Start Word.

● New Dictation Box with a host of new features including the ability to accept HTML input.

● Premium 11 includes Microsoft Outlook support. This feature used to be limited to the professional versions.

Cons:

● Nuance claims “faster recognition response for commands and dictation” but we could not verify this claim. Although the Natural Language Commands appear to deploy faster, we couldn’t detect any speed increase of dictation on our Windows 7 64-bit i7 965 chip with 12 Gb of RAM. We actually noted a decline of dictation speed on our 2.0 GHz core 2 duo Windows XP notebook computer with 2 Gb of RAM. NaturallySpeaking is smart enough to drop its settings down to Best Match III on slower computers such as netbooks but on midrange computers, such as the previously mentioned 2.0 GHz core 2 duo computer, we recommend moving the Speed vs. Accuracy bar to the 50% position (under DNS Tools/Options/Miscellaneous) which is the default setting but you may even have to move it lower. This may be a good time to purchase that new i7 computer you’ve been thinking about.

● Unfortunately, when Nuance added the new smart punctuation feature, they reintroduced an old problem that was fixed in Ver. 10 and will need to be fixed again. When attempting to insert text, make a correction or just dictate in front of punctuation, you will frequently wind up with an additional space which was aggravating Ver. 9 but eliminated in Ver. 10. In our opinion, this problem requires immediate attention and should be fixed with a service release patch as soon as possible.

● If NaturallySpeaking unexpectedly exits and you re-launch it but forget to re-launch other Full Text Control applications such as Microsoft Word, Outlook and Internet Explorer, after re-launching NaturallySpeaking you may find yourself in Samurai Command Mode. The problem can also occur when attempting to dictate into a non-Full Text Control application. When we attempted to dictate into Microsoft Word (after leaving Microsoft Word running and re-launching NaturallySpeaking), NaturallySpeaking insisted on converting our dictation into searching the Internet (including adding a new unwanted IE 8 tab), searching our desktop or creating an e-mail message which had to be deleted. If NaturallySpeaking unexpectedly closes on you, after re-launching NaturallySpeaking you will have to remember to close and re-launch your other Full Text Control applications. You may also experience Samurai Command Mode in other non-Full Text Control applications.

● Although some over-the-counter microphones, like the Plantronics DSP 400 and the Nuance HS-Gen-0 (Andrea Electronics NC-181) included with the Pro versions of DNS, worked reasonably well as starter microphones in previous versions of NaturallySpeaking, we found these types of microphones to be only marginally acceptable in DNS 11. We believe this is directly related to the new 22.5 kHz sampling rate. If NaturallySpeaking 11 is going to play an important role in your workflow, don’t settle for OEM or over-the-counter microphones which are equivalent to putting bicycle tires on a race car. If you want the best accuracy from DNS 11 a high-quality microphone with a reasonably flat frequency response from 500 Hz to 12 kHz is all but a requirement.

● Because Bluetooth is limited to 8 kHz, Bluetooth microphones will not be able take advantage of the new 22.5 kHz sampling rate (12 kHz range) used for other types of microphones. This probably shouldn’t be listed as a con because Bluetooth microphones actually work a little better in Ver. 11 but because of the new higher sampling rate, you will only be able to experience your best accuracy with other types of microphones.

● Pop-up sidebar is rather sluggish at changing its command list.

● Elimination of the Results Box hang time where words would hang in the Results Box for about 3 seconds before dropping into your document. Although we agree that this is an impressive improvement, it was implemented in Ver. 10 and is not a new Ver. 11 feature.

● Spell That Correction dialog box is no longer resizable.

● We had hoped to see a variation of the WSR “Show Numbers” feature which numbers every control on a page. Unfortunately, Nuance was unable to implement this refinement but the feature is available as a third-party utility add-on from ShowNumbers Plus.

● Being able to Delete, insert Before/After (text) and Selecting text to edit is both a blessing and curse. Being able to choose which specific word in your document you wish to capitalize can be a blessing but at the same time, deleting an extra word such as “the” because you said it twice (between pauses) requires an extra step. We’re on the fence on this feature.

● We found the new Dictation Box to be less stable than the previous version.

● NaturallySpeaking now supports direct transcribing of the Olympus DS2 and DSS files. Although we agree that this is also impressive feature, our problem is that this feature was initially implemented in Ver. 10 and is not new to Ver. 11.

Additional Notes:

Let DNS 11 import your old user files but don’t use them as they will be lethargic. Our customers will be receiving a hard copy of our new pictorial step-by-step upgrade guide which includes 15 pages to assist them in the procedure.

The Text-to-speech engine includes numerous new adjustments but as with previous versions of NaturallySpeaking, the manufacturer always sets the speed a little too low. If you use the text-to-speech engine, you’ll probably want to increase the speed.

No matter how fast your computer is, NaturallySpeaking will set the Speed vs. Accuracy slider at the 50% mark. If you’re using a fairly new processor with plenty of RAM, we recommend changing this setting to Most Accurate.

Be carefull not to press {Alt+Spacebar} in IE8 as it can still freeze IE8 as in DNS 9/10.

The Audio Setup Wizard is now referred to as Microphone Check

Select-&-Say is now referred to as Full Text Control

Nuance has changed a number of command names which are probably more practical for newbies but may take previous end users a while to get accustomed to. KnowBrainer 2010 users won’t have to deal with the new command names until they are ready because KnowBrainer 2010 includes aliases to the old command names and all the new NaturallySpeaking DragonBar menus.

Addendum: In July of 2011, Nuance released Ver. 11.5.
The following includes a list of Ver. 11.5 additional amenities:

1. IE9 support
2. Adjustable sidebar
3. Social media commands (such as the ability to instantly post to Facebook through an IM type of interface)
4. The ability to reverse the damage after utilizing the
choose all command (Select-&-Say correction) by saying undo all
5. Resizable Spell Correction Window
6. Turn your iPad or iPhone (not currently available for Android) into a wireless Wi-Fi microphone via a downloadable Dragon Remote microphone application
7. Additional iTunes command capability with Bonjour
8.
Quote That and Bracket That which will add quotes or brackets to any selected text. UNFORTUNATELY the NaturallySpeaking developers left out other important commands such as Paren That, Parenthesis That, Brace That and Angle Bracket That but we have added those commands to KnowBrainer 2011 along with commands like Quote Back which will surround highlighted text with quotes and return the cursor to the end of the current paragraph or line.

Previous Reviews:

If you would like to see some of our previous reviews check out the following:

KnowBrainer TableMike Face-Off
KnowBrainer Handheld Microphone Face-Off
KnowBrainer Headset Microphone Face-Off
KnowBrainer Telephone Microphone Face-Off
KnowBrainer Digital Recorder Face-Off
KnowBrainer Wireless Microphone Face-Off
KnowBrainer USB Soundcard Face-Off (coming soon)

 

 

 

 

Current Gold Standards:

The following is a list of microphones that we consider to be #1 in their microphone categories we:

Best Wireless Microphone = Samson Airline 77
Best Combo Telephone Wireless Microphone = Jabra Pro 9470
Best Handheld USB Microphone = Dictaphone PowerMic II
Best USB TableMike = SpeechWare 3-in-1 / SpeechWare 6-in-1
Best Wired Headset Microphone = Audio-Technica Pro 8HEmW
Best Bluetooth “like” Wireless Microphone = Plantronics Savi 440*
Best Digital Recorder = Olympus DS-7000
Best Speech Recognition Application =  NaturallySpeaking 11.5 & DMPE

 

*Bluetooth Note: The Plantronics Savi 440 is not a Bluetooth microphone but comes seemingly close by emulating the features of a Bluetooth microphone because it includes a completely wireless headset and wireless receiver. We do not recommend any Bluetooth microphones for use with Dragon Medical 11, NaturallySpeaking 11, DragonDictate 2.5 or WSR because Bluetooth microphones are limited to a 4 kHz frequency response while the previously mentioned speech recognition applications can process frequencies as high as 11 kHz. Bluetooth also suffers from syncing issues and that’s why we no longer offer or rate Bluetooth microphones.

Lunis Orcutt – Developer of KnowBrainer 2011 and founder of KnowBrainer Speech Recognition Forums (the world’s largest speech recognition forum)
Chat Live
w/a Nuance Gold Certified Speech Recognition Solutions Provider

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s