Dragon NaturallySpeaking and citrix
A White Paper from Nuance Communications March 2009
DRAGON NATURALLYSPEAKING - CITRIX

Introduction

As the number of deployed enterprise applications increases, organizations are seeking solutions that simplify application delivery without sacrificing performance security or cost. Citrix Systems has been providing innovative software solutions that enable the best delivery methods for all applications. The Citrix ICA thin client architecture provides a centralized, scalable and secure application delivery platform that can benefit organizations of all sizes. Citrix is one of the most trusted names in Application Delivery Infrastructure. More than 215,000 organizations worldwide rely on Citrix to deliver any application to users anywhere with the best performance, highest security and lowest cost. Citrix customers include 100 percent of the Fortune 100 companies and 99 percent of the Fortune Global 500, as well as hundreds of thousands of small businesses and prosumers. The latest release of the award-winning Dragon NaturallySpeaking voice recognition software from Nuance Communications supports deployments within a Citrix Presentation Server environment. This reference guide shows how customers can benefit from using Dragon NaturallySpeaking in a Citrix environment. This white paper supplements the Dragon NaturallySpeaking IT Administrators Guide which explains how to install and configure Dragon NaturallySpeaking on Citrix servers.
2009 Nuance Communications, Inc. All rights reserved.

- Page 2 -

Citrix Overview
Citrix provides a server-based centralized architecture to deliver applications to end users. In Citrix, all application processing and logic occur on the server; the client PCs only display data and allow the user to interact with the graphical user interface of the program. This approach is ideal for highly distributed environments where the administrative costs of maintaining a consistent PC desktop image are problematic or in scenarios where bandwidth and or client PC processing power is limited. Citrix also simplifies the process for updating software functionality by eliminating the need for end users to reboot the client PC. Nuance Communications understands the value that Citrix brings to large organizations and it is for this reason that Dragon NaturallySpeaking supports deployment in Citrix thin-client environments.
Benefits of Running Dragon NaturallySpeaking on a Citrix Server
Dragon NaturallySpeaking supports installation on a Citrix Presentation Server, enabling users to dictate from workstations that do not have Dragon NaturallySpeaking installed. Deploying Dragon NaturallySpeaking in a Citrix environment delivers the following features and benefits

- Page 3 -

FEATURES
Standard Application Access Centralized Management & Administration Secure

DESCRIPTION

Access to Dragon NaturallySpeaking functionality via consistent and homogeneous methods Deploy and manage all Dragon NaturallySpeaking users through Citrix management console Allows the administrator to control which end users have access to Dragon NaturallySpeaking Eliminates the time and effort of installing Dragon NaturallySpeaking on client workstations Permits the use of client workstationsthat do not meet the system requirements for Dragon NaturallySpeaking Users are able to utilize the same dictation functionality through Citrix that is available in the standard Dragon NaturallySpeaking desktop version. Leverage proven Citrix support for automatic fail-over and load balancing

BENEFITS

Simplifies end-user training for consistent end-user experience Reduced administrative overhead lowers total cost of ownership Ensures that only authorized users have access to Dragon NaturallySpeaking Lower administration costs; enables fast application deployment even for highly distributed or mobile organizations Enables deployment of complete Dragon NaturallySpeaking functionality on PCs that do not meet standard Dragon NaturallySpeaking desktop requirements Enables the application integration features of Dragon NaturallySpeaking, such as Select-and-Say, Say-what-you-See, and Natural Language commands, to be used with applications running on a server Deploy Dragon NaturallySpeaking in distributed high transaction volume environments with high availability security.
Simplified Application Deployment
Reduced PC Client-Side Hardware Requirements
Full Dragon NaturallySpeaking Functionality

Scalable and Reliable

Specifications and Deployment
Nuance provides support for deploying and running Dragon NaturallySpeaking in a Citrix environment. Customers of Dragon NaturallySpeaking must purchase a license for each speaker that creates a user profile in Citrix. Note that Citrix support is only available in the Dragon NaturallySpeaking Professional, Legal and Medical editions. Deploying Dragon NaturallySpeaking requires the following Citrix components: Server: Citrix Presentation Server 4.0 on either Microsoft Windows Server 2003 Microsoft Windows 2000 Advanced Server

- Page 4 -

Client: Citrix ICA client 8.x or 9.x on either Microsoft Windows 2000 Service Pack 4 Microsoft Windows XP Service Pack 1 Microsoft Windows XP Service Pack 2
Sizing and Configuring Dragon NaturallySpeaking in Citrix Key Issues to Consider
Organizations considering deployments of Dragon NaturallySpeaking in a Citrix environment need to consider several items for configuring the Citrix environment and scaling client usage on Citrix servers. This section provides some basic guidelines on how to address these issues to ensure maximum performance. Please note these are guidelines only and not definitive specifications-actual performance will vary from customer to customer. Deploying Dragon NaturallySpeaking in a Citrix environment must consider the following three items 1. Projected user base and usage metrics 2. Server-side hardware 3. Network bandwidth

1. Projected User Base

Before considering any deployment of Dragon NaturallySpeaking in Citrix, organizations need to answer key questions about the user base and potential growth. These metrics are essential since they are needed to size the appropriate network and hardware requirements. How many users do you plan to enable access to Dragon NaturallySpeaking through a Citrix environment in the first 6 months? 12 months? What is the projected distribution of simultaneous user access to Dragon NaturallySpeaking in a Citrix environment? Where will these users be physically distributed? On the same corporate LAN, WAN, remote access or combination?

2. Server-Side Hardware

Dragon NaturallySpeaking has been tested to run in a Citrix enterprise environment configured with Citrix Presentation Server 4.0, and clients running Citrix ICA thin-client software. Based on answers to question #1, your organization will need to scale and deploy an appropriate number of Windows servers to run Citrix Presentation Server 4.0 to support all your Dragon NaturallySpeaking users. In the appendix, we provide some baseline performance data using standard Windows hardware.

3. Network bandwidth

The request for network bandwidth when running Dragon NaturallySpeaking derives primarily from requests issued on the virtual audio channel. Nuance recommends and checks for use of the high-quality sound on Citrix to ensure the highest quality of accuracy for speech recognition. Based on the user population you intend to serve, you must account for and allocate the appropriate amount of network bandwidth for users to be able to utilize Dragon NaturallySpeaking 9 from a Citrix client. For information about setting sound quality on Citrix ICA clients, refer to the Dragon NaturallySpeaking IT Administrators Guide in the supplied documentation.

- Page 5 -

APPENDIX
Preliminary Test Results of Running Dragon NaturallySpeaking in Citrix
Listed below are the results of Nuances internal testing of Dragon NaturallySpeaking in a Citrix environment.This information stated is only meant to provide guidance for setting up your Citrix environment-it is not a definitive specification. Your experience using Dragon NaturallySpeaking with Citrix may vary, depending on many factors that may not be addressed in this brief overview. This appendix should help you accurately size what is required within your Citrix environment for Dragon NaturallySpeaking. The internal Nuance testing of Dragon NaturallySpeaking with Citrix utilized the following components: SERVER SPECIFICATIONS Server Hardware Dell PowerEdge 2850 Processor: Dual Intel Xeon single-core processor, 3.16 GHz/1MB cache, 800 MHz FSB Memory: 4GB DDRMHz (4x1GB), single ranked DIMMs Hard Drive: 146GB 10K RPM Ultra 320 SCSI Hard Drive Network Card: Dual Onboard NICs Dell PowerEdge 6850 Processor: Quad Intel Xeon single-core processor, 3.16 GHz/1MB cache, Redundant Memory: 8GB DDRMHz (8x1GB), single ranked DIMMs Hard Drive:146GB 10K RPM Ultra 320 SCSI Hard Drive Network Card: Dual Onboard NICs Server Hardware Windows Server 2003 Published applications: Dragon NaturallySpeaking Professional, Medical and Legal editions Dragon NaturallySpeaking Audio Client Update Microsoft Outlook Microsoft Word 2003 Client Software Citrix client user interfaces: ICA 32-bit clients, version 8.x and 9.x Program Neighborhood Program Neighborhood Agent Web Client PC Client Specifications Dell Optiplex GX 260 with Windows XP 512 RAM and 2 GHz CPU Sound card: Sound Blaster Live! And also on-board sound systems Network Specifications Network speed: 100 Mbps Fast Ethernet Network environment and active software run on the server: Latest Citrix Presentation Server 4.0 service packs were installed; available from www.citrix.com Citrix tools for CPU and memory management were activated to optimize server performance.

- Page 6 -

An unspecific amount of corporate network traffic was present during testing that was not generated by the Citrix transactions created by running Dragon NaturallySpeaking. NOTE: no additional high-memory or CPU-consuming applications were active during testing

INITIAL RESULTS

CPU Consumption and Memory Usage Running Dragon NaturallySpeaking
The following tables show the CPU time and memory usage observed on the Citrix servers with Dragon NaturallySpeaking running. The CPU time represents the total time (expressed as a percentage) used by a single client session; the percentage is a total of available CPU. Memory usage is also represented for a single-user session as a percentage of total RAM available. Peak CPU consumption is reached when Dragon NaturallySpeaking user profiles are being opened, saved and closed. Acoustic training appears to consume the most amount of CPU time for any process executed with Dragon NaturallySpeaking on Citrix. As such, administrative procedures that run any training process need to take this into consideration CPU consumption, with one user dictating and correcting in the DragonPad
CPU time (Dual processor) CPU time (Quad processor)

Average

10% 5%

25% 13%

CPU consumption, performing General Training in New User Wizard, Additional Training or running the Acoustic Optimizer

25% 12%

30% 18%
Memory consumption, with one user dictating and correcting in the DragonPad
RAM required for both Dual and Quad processors

130 MB

160 MB
When other applications are running concurrently with Dragon NaturallySpeaking: Dragon NaturallySpeaking consumes the same amount of CPU time as it does when it runs alon e Memory usage for Dragon NaturallySpeaking will increase

- Page 7 -

For example, memory usage for Dragon NaturallySpeaking goes up when running together with Microsoft Word 2003.
RAM required for a single session, Dragon/Microsoft Word 2003 When Dragon NaturallySpeaking is not being used: 160 MB

190 MB

CPU time consumed is 0. The same amount of RAM is required Network bandwidth is 1.3 Mbps bandwidth if the microphone is active; 0 Mbps if the microphone is turned off.
Network Bandwidth The network bandwidth setting for a single Dragon NaturallySpeaking Citrix client MUST be at least 1.3 Mbps.
For information about setting sound quality on Citrix ICA clients, refer to the Dragon NaturallySpeaking IT Administrators Guide in the supplied documentation. NOTE: If the microphone is turned off for an active DNS session running on Citrix, network bandwidth on a channel drops almost to 0.

Summary

Based on the foregoing observations made while testing a single Dragon NaturallySpeaking session running on Citrix, it is reasonable to conclude that deployment of Dragon NaturallySpeaking will scale linearly in a Citrix Presentation Server environment. CPU consumption emerges as the most important limiting factor, as collective CPU time per session will determine how many Dragon sessions can be active at the same time. Using the hardware and software configurations detailed in the appendix, here is the approximate range of Dragon NaturallySpeaking sessions that can be expected to run on the Citrix Presentation Server 4.0 utilizing the hardware and software configurations outlined in the appendix: Maximum Load: In a realistic scenario, where multiple sessions are performing different types of activities (opening, saving and closing users, dictating, and correcting dictation), it can be expected that 6 concurrent Dragon NaturallySpeaking 9/Citrix sessions can be executed. Minimum Load: Expect to run 4 concurrent Dragon NaturallySpeaking/Citrix sessions in the case when all Dragon NaturallySpeaking sessions are running acoustic training.

- Page 8 -

About Nuance Communications, Inc.
Nuance (Nasdaq: NUAN) is the leading provider of speech and imaging solutions for businesses and consumers around the world. Its technologies, applications and services make the user experience more compelling by transforming the way people interact with information and how they create, share and use documents. Every day, millions of users and thousands of businesses experience Nuances proven applications and professional services. For more information, please visit www.nuance.com.
About Citrix Systems, Inc.
Citrix Systems, Inc. is the global leader and the most trusted name in Application Delivery Infrastructure. More than 215,000 organizations worldwide rely on Citrix to deliver any application to users anywhere with the best performance, highest security and lowest cost. Citrix customers include 100 percent of the Fortune 100 companies and 99 percent of the Fortune Global 500, as well as hundreds of thousands of small businesses and prosumers. Citrix has approximately 8,000 partners in more than 100 countries. Annual revenue in 2008 was $1.6 billion.

- Page 9 -

Dragon NaturallySpeaking Dragon Medical

Version 10

User Guide
Version 10 Professional, Medical, Legal, Contact Center, Preferred, Standard, and Essentials editions. Nuance Communications, Inc. has patents or pending patent applications covering the subject matter contained in this document. The furnishing of this document does not give you any license to such patents. This software is protected under the following patents, among others: 5,027,406; 5,168,548 5,202,952; 5,231,670; 5,388,183; 5,428,707; 5,465,318; 5,526,463; 5,680,511; 5,715,367; 5,754,972; 5,765,132; 5,794,189; 5,799,279; 5,818,423; 5,822,730; 5,850,627; 5,890,181; 5,909,666; 5,915,236; 5,920,836; 5,920,837; 5,949,886; 5,960,394; 5,970,448; 5,970,460; 5,983,179; 6,029,124; 6,058,366; 6,064,959; 6,073,097; 6,088,671; 6,092,043; 6,092,044; 6,101,468; 6,125,342; 6,125,347; 6,138,098; 6,151,575; 6,163,768; 6,167,377; 6,212,498; 6,260,013; 6,292,779; 6,349,282; 6,424,943; 6,456,972. 6,393,399, 6,360,237, 6,456,972, 6,490,555, 6,601,027; 6,912,498, and 7,315,818. No part of this manual or software may be reproduced in any form or by any means, including, without limitation, electronic or mechanical, such as photocopying or recording, or by any information storage and retrieval systems, without the express written consent of Nuance Communications, Inc. Specifications are subject to change without notice. Copyright 2002-2008 Nuance Communications, Inc. All rights reserved. Nuance, ScanSoft, the Nuance logo, the Dragon logo, Dragon, DragonBar, NaturallySpeaking, NaturallyMobile, RealSpeak, Nothing But Speech (NBS), Natural Language Technology, Select-and-Say, MouseGrid, and Vocabulary Editor are registered trademarks or trademarks of Nuance Communications, Inc. in the United States or other countries. All other names and trademarks referenced herein are trademarks of Nuance Communications or their respective owners. Designations used by third-party manufacturers and sellers to distinguish their products may be claimed as trademarks by those third-parties. Pentium is a registered trademark of Intel Corporation. Adobe and Acrobat are registered trademarks of Adobe Systems Incorporated. Corel and WordPerfect are registered trademarks of Corel Corporation. Lotus and Lotus Notes are registered trademarks of Lotus Development Corporation. Microsoft, Outlook, Windows, Windows NT, Visual Basic, and PowerPoint are trademarks or registered trademarks of Microsoft Corporation. Mozilla, Mozilla Firefox, and Mozilla Thunderbird are trademarks or registered trademarks of the Mozilla Foundation. Voice It, the Voice It logo, and Voice It Link are trademarks or registered trademarks of VXI Corporation. AMD is a trademark of Advanced Micro Devices, Inc. America Online is a registered trademark of America Online, Inc., a division of AOL Time Warner. Sony and Memory Stick are regitered trademarks of the Sony Corporation Panasonic is a registered trademark of the Matsushita Electric Industrial Co., Ltd iPAQ is a registered trademark of the Hewlett-Packard Company Palm OS is a registered trademark of of PalmSource, Inc. or its affiliates. Macromedia Flash is a trademark of Macromedia, Inc. Dave Barry in Cyberspace, Copyright 1996 by Dave Barry, published by Crown Publishers. 3001: The Final Odyssey, Copyright 1997 by Arthur C. Clarke, published by HarperCollins Publishers. Dogbert's Top Secret Management Handbook, Copyright 1996 by United Feature Syndicate, Inc., published by HarperBusiness, a division of HarperCollins Publishers. Success Is a Journey: 7 Steps to Achieving Success in the Business of Life, Copyright 1999 by Jeffrey J. Mayer, published by McGraw Hill. Charlie and the Great Glass Elevator, Copyright 1972 by Roald Dahl, published by Alfred A. Knopf, Inc., and Penguin Books. The Captain of Battery Park, Copyright 1978 by Eugene Pool, published by Addison-Wesley.

50-A11A-10511

Contents
C H A P T E R 1 Introduction 1 About this guide 2 What should I expect from Dragon? 3 C H A P T E R 2 Installation and User Creation 5 System requirements 5 Plugging in the microphone 7 What you should know before installing 7 Before installing the software 8 Choosing type of installation 9 Installing software Typical Installation 10 Installing software Custom Installation 12 Modifying application settings for all users 16 Modifying administrative settings 17 Setting formatting options 17 Creating a new user 19 Training a new user 24 Adapting to your writing style 25 Scheduling periodic optimizations 26 Upgrading Existing Users from Versions 8.x and 9.x 27 Version 10 File Structure 29 Accessing log files, samples, and tools 31 Setting command, hot key, backup options 33 C H A P T E R 3 Starting to Dictate 37 Starting Dragon 37 Turning on the microphone Starting to dictate 39 Getting Help 43 Troubleshooting 45 The DragonBar 45 Using QuickStart 48

4 Enter your User Name and Organization, and the Serial
Number supplied to your installation.
5 (Optional) When the Setup Type page appears, click the
Change button and choose where to install the product.
If no earlier versions of Dragon are installed on your system, the default directory is: C:\Program Files\Nuance\NaturallySpeaking10 For a list of directories created by installation, see Version 10 File Structure on page 29.
6 While you are still on the Setup Type page, select Typical. 7 On the Ready to Install the Program page, you can choose:
Enable QuickStart mode for the current userIf you enable QuickStart, the product launches at system startup time and adds the Dragon QuickStart icon to the Windows task bar. Upgrade existing speech files to work with the installation If you select this option, either immediately after you install (or after you reboot when you are prompted to reboot), Dragon starts the user upgrade process (Upgrading Existing Users from Versions 8.x and 9.x on page 27).
8 Click Install to start the installation.
9 When prompted, click Finish to complete the installation. (If
you are prompted to restart your computer, restart it now.)
10Select Start > All Programs > Dragon NaturallySpeaking 10 >
Dragon NaturallySpeaking 10. The DragonBar appears on your desktop.
11(optional) If you have users from Version 8.x or 9.x that you
would like to upgrade before proceeding, refer to Upgrading Existing Users from Versions 8.x and 9.x on page 27.
12If you have no previously existing users, the New User
Wizard starts immediately after you run the product. Proceed to Creating a new user on page 19 to set up a user, then continue as instructed, to train the user. You are now ready to create a user, as explained under Creating a new user on page 19.
Installing software Custom Installation
To install portions of Dragon or to install it for multiple users:
NOTE: Windows Vista: When you start the installation on Windows Vista, if you see a message saying A program needs your permission to continue, click Continue to start the installation.
3 After the Wizard begins click Next to proceed to the License
Change button and choose where to install the product. If no earlier versions of Dragon are installed on your system, the default directory is: C:\Program Files\Nuance\NaturallySpeaking10 For a list of directories created by installation, see Version 10 File Structure on page 29.
6 While you are still on the Setup Type page, select Custom for
the type of installation.
7 Click Next and you see a tree where you can select any
particular feature of the product and click the down arrow to its left to choose where/when to install it:
Install now, on local hard drive Install now, with all subfeatures on local hard drive Install when it is needed on a just-in-time basis These options are particularly useful for installing multiple languages and/or multiple vocabularies.

2 If you re-read the same words and the computer still doesnt
get it, just click Skip. Otherwise, click Next to continue.
3 Choose a text to read aloud and click OK.
You only need to read for about five minutes to train Dragon to recognize your voice. You can take breaks during this training by clicking Pause. Dont worry if you make mistakes or laugh. You should try to read exactly what you see on the screen, but its okay if you read something incorrectly. The computer either ignores the mistake or positions the yellow arrow at the beginning of the text for you to reread it. The progress bar shows how much text is left to dictate.
During training, dictating punctuation is not necessary.
4 When youve read enough, the New User Wizard displays a
congratulations message. Click OK and Dragon starts adapting to your voice.
5 Proceed to Adapting to your writing style on page 25.
Adapting to your writing style
After it adapts to your voice, the New User Wizard prompts you to adapt Dragon to your writing style. Dragon can analyze your writing style to increase your overall recognition accuracy. This tool scans documents in your My Documents directoryMicrosoft Word, Corel WordPerfect, text files, html files, and rich text files (.rtf). Adapting to your writing style can add 5 to 30 minutes to user training. Close all other programs when you run this step.
NOTE: Though we recommend that you run this part of the New User Wizard, if you plan to skip this step, select Skip this step rather than Cancel.
1 Click Start to continue. While scanning your files, the New
User Wizard displays its status.
2 When Dragon has finished adapting to your writing style,

click Next.

3 Proceed to Scheduling periodic optimizations on page 26.
Scheduling periodic optimizations
After it adapts to your writing style, the New User Wizard presents the Periodic Optimizations page, where you select the type of language and acoustic model optimizations Dragon should carry out and how often:
1 Check off the optimizations you want carried out. 2 To the right, click the Configure button to set when, how
often, and what time of day to run each optimization. Instead of Daily, Weekly, and Monthly, you can run the optimization At System Startup, At Logon, or when the system is idle.
3 You might be asked to enter your Windows password. 4 Under Data Collection you set whether or not to collect data
and schedule when it should be collected. If you choose to turn it on, this process collects 500 MB of acoustic data from your dictation sessions. You have the option of sending that data to Nuance to help improve the accuracy of future versions of Dragon. No personal information is ever sent to Nuance and participation in data collection is completely voluntary. Click Configure to schedule acoustic data collection to occur and be sent Daily, Weekly, or Monthly, and to indicate the date and time it should start.

NOTE: You can start Dragon over a Remote Desktop connection (Windows XP Professional and Windows XP Home), but interactive dictation is not supported. Using a Remote Desktop connection, you can access the DragonBar and transcribe dictation but you cannot use a microphone attached to the remote machine to dictate.
Turning on the microphone
Before you can dictate, you need to turn on the microphone. You turn on the microphone by:
Click the microphone icon on the DragonBar. You can click this icon again to turn it off. Pressing the plus (+) key Microphone icon on the numeric keypad to turn the microphone on, and then press it again to turn the microphone off. Clicking the microphone icon in the Windows task bar.
The button and the volume meter on the DragonBar change to show if the microphone is off or on. Shows that the microphone is off Shows that the microphone is on
Once the microphone is off, you cannot turn it on again by

voice.

Sleeping and waking up
To make Dragon stop listening temporarily:
1 Say Go to Sleep or Stop Listening. Then Dragon
ignores everything except the Wake Up or Listen To Me commands.
2 To reactivate the microphone, say Wake Up or Listen To
Me. You can also press the numeric plus (+) key to turn on/off the microphone.

Starting to dictate

To begin dictating, start a word processor (such as Microsoft Word or Corel WordPerfect) and begin a new document. Make sure your text insertion point is at the start of the new document.

DragonPad

You can use the DragonPad, a simple word processor included with Dragon. To open the DragonPad, from the Tools menu on the DragonBar, click DragonPad.
TIP: Remember to click in the window you want to talk to before you speak.
Start talking. As you talk, text displays in the Results Box while Dragon figures out what you said. The Results Box is a small yellow window that appears on-screen as you dictate. For example:
The words in the Results Box may change as the program considers different interpretations of what it heard. The Results Box displays a line that changes from yellow to green as you speak, indicating the volume of your dictation.
When Dragon recognizes what you said as a command, the Results Box displays a bold blue border.
You can also anchor the Results Box permanently in one place. From the Tools menu of the DragonBar, click Options, then click the View tab and enable the Anchor option. Dont worry about mistakes at this point; Dragon improves as you use it.
Using Natural Punctuation
Dragon can automatically add periods and commas at the appropriate places in your dictation without you having to explicitly speak that punctuation. The Natural Punctuation feature can be useful in helping you get used to dictation by

When the Correction menu list appears, you can also choose to:
Spell the word (for example, say Spell That q-u-i-c-k). Selecting a large amount of text and then saying Spell That can produce unpredictable results. Press the correction hot key () while the Correction menu is open will display the Spell dialog box. Listen to a recording of what you just said (say Play That Back).
Capitalize it (say Cap That, in this example, to get Clinic) Say Unselect That to cancel the selection and close (dismiss) the Correction menu Ignore the Correction list and keep dictating (in the example above, just say quick to replace the selected text)
TIP: If you prefer not to see the Correction menu every time you select text by voice, you can turn it off. In the Options dialog box, click the Correction tab to display the Correction options, then clear the check box for Select commands bring up Correction menu.
To help you quickly find the alternative you want, the recognition choices in the Correction menu appear in boldface where they differ from the word youve selected (applicable only when you select multiple word at one time).
Words that are changed in the selection appear in bold, to help you find the correct alternative quickly.
If you select any leading or trailing spaces (including paragraph marks) along with the text you want to correct, choosing one of the alternatives in the Correction menu deletes those leading and trailing spaces.

Correcting Punctuation

Note the following when correcting punctuation:
When you select a phrase (more than one word) that has punctuation, be sure to dictate that punctuation. For example, say you initially dictated, Today is Thursday I have a meeting at 11 oclock and the text was transcribed as Today
is Thursday, I have a meeting at 11 oclock. If you want to change the comma to a period, say Select Thursday comma. If the correct punctuation appears in the Correction menu, select it. Otherwise, you can dictate over the selection by, for example, saying Thursday period.
When you select punctuation, the Correction menu might list alternate punctuation (see example to the right)

If Natural Punctuation is turned on, when you select a phrase that Dragon has added punctuation to, the choices will include punctuation changes (see example here).
NOTE: When you select a word that has adjacent punctuation the Natural Punctuation feature added, the selection will extend to include that punctuation.

Selecting text by voice

Using Select-and-Say
You can revise your dictation without correcting it by selecting the text using the Select command and then saying new words to replace the selected text.

To use Select-and-Say:

1 Dictate the sentence below: US/Canada: Lets meet for lunch on Tuesday [period] Other Dialects: Lets meet for lunch on Tuesday [full stop] 2 Say Select lunch on Tuesday. The words lunch on Tuesday
should be highlighted on the screen.
3 Say dinner on Wednesday. These words should replace
lunch on Tuesday. Since Tuesday and Wednesday sound completely different, Dragon will know you are not correcting a recognition mistake but rather revising your dictated text.
You can also select punctuation marks.
4 Say: Select period (US/Canada) or Select full stop
(Other Dialects). If theres more than one period or full stop, you can say Select Again to select a different one.
5 To replace the period or full stop with an exclamation mark,

say exclamation mark.

TIP: It is often easier for Dragon to find the matching text if you select a short phrase rather than individual words. If you select some words that are already correct, just say them again along with the ones you want to change.
Selecting the same text again
If the words youre trying to select appear more than once on the screen and Dragon selects the wrong ones, just say Select Again. The program then looks for another instance of the same word or words. You can also say Select Again if Dragon selects a word that sounds like the word you want, but is not the correct one (for example, two instead of too).

Unselecting words

If the wrong text is selected, say Unselect That.
You can also unselect words by moving your insertion point (by mouse or voice) to another part of your document. For example, say Go to End of Line or click somewhere else in your document.
Selecting a longer phrase
You can select a longer phrase by saying Select [text] Through [text] (US/Canada) or Select [text] To [text] (Other Dialects). For [text], substitute the actual word or words at the beginning and the end of the range of wrong words. For example, you could correct the underlined words in the following sentence: With a little practice, who will develop a habit of dictating an unclear, steady voice, and the computer will understand you better. by saying: (US/Canada) Select who Through unclear or Select who will Through an unclear (Other Dialects) Select who To unclear or Select who will To an unclear Then dictate the correct text: you will develop the habit of dictating in a clear The final corrected sentence reads: With a little practice, you will develop the habit of dictating in a clear, steady voice, and the computer will understand you better.

NOTE: If youre correcting more than one word, the words must all be in sequence (next to each other). You cant use a single command to correct words that are in different parts of your document.
Selecting your whole document
To select all the text in your document, say Select Document or Select All. This command is useful when you want to change the font or the way text is aligned.
When you want to copy all the text in a document to another window, the easiest way to do it is with the Copy All to Clipboard command.
NOTE: When a lot of text is selected, the Scratch That and Cut That commands dont work, nor can you overwrite the selection by dictating new text. This prevents you from accidentally deleting a large part of your document. To remove a large selection, you can say Delete That instead.
Selecting an entire paragraph or line
You can select the current paragraph by saying Select Paragraph. To select the current line, say Select Line. You can also select a number of paragraphs or lines (up to 20). For example, you can say Select Previous 5 Paragraphs.

SAY THEN THEN

Select
Next Previous Forward Back Last
Paragraph 2.20 Paragraphs Line 2.20 Lines
Selecting a word or character
You can select the current word by saying Select Word. To select a character, say Select Next Character or Select Previous Character. You can also select a number of words or characters (up to 20). For example, say Select Previous 2 Words.

Next Previous Forward

Word 2.20 Words Character

T HE N

Back Last

2.20 Characters

Moving around in a document
When youre editing a document, you can move around in it by voice. After you place the insertion point where you want it, you can dictate more text, select text, copy and paste, or apply formatting.
Going to the top or bottom of a page or document
You can move to the top or bottom of the current page by saying Page Up (equivalent to pressing the PAGE UP key) or Page Down (equivalent to pressing the PAGE DOWN key). You can move to the top or bottom of your document by saying Go to Top or Go to Bottom.

Go to Move to

Top Bottom Top of Document Beginning of Document Start of Document Bottom of Document End of Document
Going to the beginning or end of a line
You can move to the beginning or end of the current line by saying Go to Beginning of Line or Go to End of Line.

Beginning of Line

Move to
Start of Line End of Line
Placing the insertion point before or after a specific word
You can place the insertion point before a specific word by saying Insert Before and then the word or words. You can place the insertion point after a word by saying Insert After and then the word or words. After you move the insertion point where you want it, you can dictate more text, paste text, add punctuation, and so on.

If you have Dragon Naturally Speaking Professional or Medical or a higher edition, you can save your dictation with your document for later playback (see Dictate now, correct later on page 85). Otherwise, you cannot play back dictation after closing a document. If dictation is not available, the Play Back button is dimmed (grayed out). When playback is not available, you may find text-
to-speech useful for checking your work. See Notes on correcting with another authors user files on page 87.
Playing back dictation in a document
To help you check your work for mistakes, you can play back a line, a paragraph, a selection, or the whole document. After playback starts, you can stop it as soon as you notice a mistake and automatically open the Correction menu. The Playback toolbar is an extra section of the DragonBar that is normally hidden. To see the Playback toolbar, click the double chevron icon on the DragonBar to open the Extras toolbar.

To play back dictation:

To play back dictation, do any of the following:
Select the text you want to play back, and say Play That Back. Click the Start Playback button on the Playback toolbar. Move the insertion point to the text you want to play back and say any of the following commands:

SAY TO

Play Back Line Play Back Paragraph
Play back dictation for the current line. Play back dictation for the current paragraph.
Play Back Document Play back dictation for the whole document. Play Back Window Play back dictation for the text in view.

Play Back to Here

Play back dictation from the top of the document window to the insertion point.
Play Back from Here Play back dictation from the insertion point to the bottom of the document window.

To stop playback:

To stop playback, do any of the following:
Click the Stop Playback button on the Playback toolbar. Click anywhere in the document window. Press the ESC key.
(Its not possible to stop playback by voice, because the computer cannot hear speech input when it is playing back dictation.)
To skip backward or forward:
You can skip backward or forward in your document by a few words by clicking the Skip Backward and Skip Forward buttons.
To stop playback and correct a mistake:
To stop playback and correct a mistake, do any of the following:
Click the Correction button on the Playback toolbar. Press the minus () key on the numeric keypad.
This stops playback and simultaneously opens the Correction menu. There you can correct the text for the last phrase played back.

NOTE: By default, Dragon stores about 30 minutes of dictation (40 MB). If you want to be able to store more dictation, you can change the amount of disk space thats set aside for storing it. On the Tools menu, click Options, and then click the Data tab. Increase the number in the Disk space reserved for playback box.
Dictate now, correct later
When you dictate into the DragonPad, you can save your dictation with your text so either you or someone else can correct it later. You must create and edit your file in the DragonPad to be able to play back dictation. This feature is available in Dragon Professional, Medical, and higher editions. You can save dictation with your document if you select the Save recorded dictation with document box on the Data tab of the Options dialog. The first time you save a document in DragonPad during an editing session, Dragon asks you whether you want to save your speech data. If you save your speech data, you can reopen the document at a later time and play back your dictation. If you do not save your speech data, your dictation is stored only during the current editing session.
Correcting your own dictation
If you save your dictation, you can open your file later and play back and correct the text as if you had just dictated it.
Correcting someone elses dictation
You can play back someone elses dictation and correct the text to match the dictation. You can do this in two ways: Correct the dictation using your own user files. Correct the dictation using the document authors user files.
Correcting with your own user files
You should use this correction technique if you regularly correct another persons text by voice and it is not important to maximize the authors recognition accuracy. When you use your own user files to correct someone elses text, you can correct the dictation just as you would correct your own dictation, using any combination of voice commands and keyboard typing. Even though you are correcting someone elses dictation, your work wont reduce recognition accuracy for either you or the person who dictated the text. But your corrections to the other persons dictation wont improve his or her recognition either.
Correcting with the document authors user files
You should use this correction technique if you do not need to correct by voice and it is important to maximize the authors recognition accuracy.
TIP: This technique works if all work is done on a single computer. Consult your Dragon reseller about ways to dictate on one computer and correct on another.

Set Font Arial Set Font Arial 12 or Set Font Arial 12 Point Set Size 12 Bold Set Font Arial Bold Set Font Bold
NOTE: If youre changing only the font size, use the Set Size commands, not the Set Font command.

Changing the font later

You can go back and change the font face, size, or style of text by selecting it and then using the Format That commands. Format That works on selected text with the same combinations of font face, size, and style as the Set Font commands (see the tables in the previous section.)

To change the font:

1 Select the text you want to change. 2 Say Format That and then the font attributes you want to
apply as described in the previous section. For example, say Format that Arial 18.
Bold, italics, and underlining
You can apply formatting to text in your document by selecting it and then saying Bold That, Italicize That, or Underline That. You can also say Strikeout That.
To add bold, italics, and underlining:
1 Select the text you want to change. 2 Say Bold That, Italicize That, Underline That, or

Strikeout That.

To remove formatting:
1 Select the text you want to change. 2 Say Restore That to remove formatting.

Aligning text

You can change how text is aligned by placing your insertion point in the text and saying Center That, Left Align That, or Right Align That.

To align text:

1 Move the insertion point to the text you want to align.
2 Say Center That, Left Align That, or Right Align
That. You can also use the following commands on selected text: Format That Centered, Format That Left Aligned, and Format That Right Aligned.

Bulleted text

To add bullets to text, place your insertion point in the text and say Format that Bullet Style. To remove bullets, say Format that Bullet Style again.
Dictating Numbers, Punctuation, and Special Characters
his chapter describes how to dictate numbers, punctuation, Web addresses, and a few other special items. For complete information on this topic, see the online help.

Dates tab.

3 Click the check box labeled Use roman numerals for diabetes

types.

Postal and ZIP Codes
Dragon can automatically formats postal codes. This feature is controlled from the Formatting tab of the Options dialog box. For more information, see the online help.

US ZIP Codes

You can dictate US five-digit ZIP codes just as you dictate any group of numbers. When dictating nine-digit ZIP codes, you must say the hyphen.

01886 94704-1150

oh one eight eight six nine four seven oh four [hyphen] one one five oh
UK and Canadian Postal Codes
You can dictate UK postcodes by saying Postcode followed by the letters and numbers that make up the postcode. For Canadian postal codes, say Postal code followed by the letters and numbers that make up the postal code. Spacing and formatting will happen automatically.
NG3 2HX E10 7BD EC2Y 4LK K1A 0M5 X0A 0H0
Postcode n g three two h x Postcode e ten seven b d Postcode e c two y four l k Postal code k one a zero m five Postal code x oh a oh h oh
To indicate that you want United Kingdom and Canadian postal codes formatted automatically in Dragon Medical, refer to Automatic formatting of dates, times, telephone numbers on page 102.

Currency and coin

You dictate your own currency as you would normally say it.
Currency in US/Canada (US English dialect)
$58.00 $1.75 $5.25 $3.9 billion 45 99.50 2.million
fifty eight dollars and zero cents one dollar and seventy five cents five dollars and twenty five cents three point nine billion dollars forty five euros ninety-nine euros and five cents pound sterling sign two point two oh pound sterling sign five million
Currency in Other Dialects (UK, Australian, Indian, and Southeast Asian English)
$58.00 $1.75 4.25 3.9 billion 45 99.50 2.million
dollar sign fifty eight dollar sign one point seventy five four euros and twenty five cents three point nine billion euros forty five pounds ninety nine pounds and fifty pence two pounds twenty five million pounds
NOTE: Dragon uses the currency symbol ($, , and so on) specified in your Windows Regional Settings as your default currency.
Dictate other currencies by first saying the currency symbol followed by the digits.

open paren (US/Canada) left paren (US/Canada) left parenthesis (All Dialects) open parenthesis (All Dialects) close paren (US/Canada) right paren (US/Canada) right parenthesis (All Dialects) close parenthesis (All Dialects) apostrophe apostrophe-ess hyphen minus sign dash space bar

s -- [double hyphen]

You can use the Vocabulary Editor to view or change the formatting properties of the punctuation.

Words that end with s

To dictate a word ending with apostrophe S (s), just say it as you normally would. Dragon will often add it. If Dragon doesnt include the s, you can add it later.
We took Marys car thats enough its time to go
Including s as you dictate
When you want to make sure that Dragon types a word with s, say apostrophe ess after saying the word. For words that
already end in s, say apostrophe after the word, as in the second example shown here.

TO EN TER SA Y

We took my brothers car I met my friends children
We took my brother [apostrophe ess] car I met my friends [apostrophe] children
TIP: You can add s to a word by selecting it and then saying it again with apostrophe ess.
Dictating hyphenated words
Many hyphenated words and phrases are already in the Dragon vocabulary. To dictate a word or phrase that is usually hyphenated, just say it as you normally would.
long-lasting up-to-date schedule Tokyo-based company nine-year-old boy
long lasting up to date schedule Tokyo based company nine year old boy
Including hyphens as you dictate
To hyphenate words that Dragon doesnt hyphenate automatically, just say hyphen wherever you want a hyphen.
speech-recognition software power-sharing agreement Elizabeth Walker-Smith
speech [hyphen] recognition software power [hyphen] sharing agreement Elizabeth Walker [hyphen] Smith
You can hyphenate words you just dictated with the command Hyphenate That. Select the words you want to hyphenate and say Hyphenate That. If you say this command when no words are selected, Dragon will usually hyphenate the last few words that you said.

Removing hyphens

You can remove a hyphen by selecting it and replacing it with a space.

To remove a hyphen:

1 Say Select hyphen. 2 Say Space Bar.

Preventing hyphens

You can prevent Dragon from entering a hyphen by pausing where the hyphen would normally be. For example, to type long lasting (normally hyphenated) say long, then pause for a moment, and then say lasting. Or you can say long space bar lasting without pausing, to insert a space in place of the hyphen.
Hyphens and dashes in Medical Edition
In the Dragon Medical edition, to indicate that you want a hyphen between a number and the word French or that you want a hyphen between a single digit and point when no other digit follows:

Dictating abbreviations and acronyms
Dragon knows many common abbreviations (such as NYC and BBC) and acronyms (such as NATO). To dictate an abbreviation or acronym, just say it as you normally would.
US/Canada: Dr. Other Dialects: Dr UK RSVP
Doctor U K (say each letter) R S V P (say each letter)
HTML 8 cm US/Canada: pp. 2733 Other Dialects: pp 2733 NATO NASDAQ
H T M L (say each letter) eight centimeters pages 27 hyphen 33 NATO (say as one word) NASDAQ (say as one word)
If Dragon types the full word instead of the abbreviation or acronym, enters the wrong word, or includes incorrect punctuation, just correct it in your document by voice or by keyboard.
Abbreviations and spelled-out words in Dragon Medical
You can set up Dragon Medical to automatically use special abbreviations or spell out words that would otherwise be abbreviated or replaced by symbols but should not be in a medical context. To set up these types of special formats in Dragon Medical:
3 Click the check boxes that indicate the usage you would like:
Rewrite cc after a number as mL Rewrite U after a number as units Write degree and degrees after a number as Insert a space between mm and Hg Write % as percent Write out positive and negative before numbers and for blood types and Rh factors Expect temperatures to be dictated in Celsius rather than Fahrenheit
4 Click the Abbreviations tab.
5 Click the check boxes that indicate the usage you would like.
Here are some examples of check boxes on the tab:
Expand ACL to anterior cruciate ligament Expand ADHD to attention deficit/hyperactivity disorder or
Click Select All to select all check boxes on the tab or click Deselect All to deselect all check boxes on the tab.
6 Click the Miscellaneous tab. 7 Click the check boxes that indicate the usage you would like.
new paragraph should insert one new line, not two Dosage abbreviations with period Insert commas into oncology T N M expressions Replace & with and Treat bracket as equivalent to parenthesis Write accented characters with their non-accented forms Write AP before the word diameter as A-P Spell orthopedics and related words with ae Rewrite q.d., q. day, or q.daily as daily
8 Click OK to save the settings.
Dictating e-mail and Web addresses
Dictate e-mail and Web addresses as you would normally say them. Dragon automatically formats them for you.
Virginia@aol.com info@samplecompany.com http:// www.nuance.com
Virginia at a o l dot com [No Caps On] info at sample company dot com [No Caps Off] [No Caps On] h t t p w w w dot nuance dot com [No Caps Off]