Knowledge Base

Frequently asked questions about the speech recognition feature for Office

Article ID: 826853

Article Last Modified on 3/23/2006

APPLIES TO

Microsoft Office Excel 2003
Microsoft Office FrontPage 2003
Microsoft Office InfoPath 2003
Microsoft Office OneNote 2003
Microsoft Office Outlook 2003
Microsoft Office Outlook 2003 with Business Contact Manager
Microsoft Office PowerPoint 2003
Microsoft Office Project Professional 2003
Microsoft Office Project Server 2003
Microsoft Office Project Standard 2003
Microsoft Office Publisher 2003
Microsoft Office Publisher 2003 with Digital Imaging
Microsoft Office SharePoint Portal Server 2003
Microsoft Office Standard Edition 2003
Microsoft Office Visio Professional 2003
Microsoft Office Visio Standard 2003
Microsoft Office Word 2003
Microsoft Excel 2002 Standard Edition
Microsoft PowerPoint 2002 Standard Edition
Microsoft Word 2002 Standard Edition

This article is a consolidation of the following previously available article: 283159

SUMMARY

You can use the speech recognition feature to speak to your computer through a microphone. If your computer meets the system requirements, you can dictate text by using any one of the products that are listed in the "Applies to" section Or, with a mouse click or a spoken term, you can switch the input mode to Voice Command. Therefore, you can control these products without clicking the menus. This feature is available in the Simplified Chinese version, the English (U.S.) version, and the Japanese version of Microsoft Office XP and Microsoft Office 2003.

MORE INFORMATION

About Speech Recognition

You can use speech recognition to dictate text by using any Microsoft Office XP or Microsoft Office 2003 program. You can also select menus, toolbars, and dialog box items by using your voice. Speech recognition is not designed for complete handsfree operation. You may obtain better results if you use a combination of your voice and the mouse or the keyboard.

Questions and Answers

What can speech recognition do for me?

The new speech recognition technologies can save you significant time if you enter text by using Dictation mode and if you control menus by using Voice Command mode. However, speech recognition is not designed to completely free you from the keyboard. You do have to use some keyboard interaction.
Where can I learn more about speech recognition?

For the most up to date information about speech recognition developments at Microsoft, visit the following Microsoft Web site:

For general questions, examine the public newsgroup to see whether your question has been answered. If your question has not been answered, post the question to the newsgroup. The newsgroup likely can provide the quickest answer to your question.
A Speech API 5 voice program is already installed on my computer. I subsequently installed either Microsoft Office XP or Microsoft Office 2003 and received one of the following error messages:

There was an error in the speech recognition engine, reinstall the engine

The speech recognition engine did not initialize. try another engine, or try to reinstall the malfunctioning engine

Most likely, an earlier version or a beta version of Speech API 5 is installed. To remove an earlier version of Speech API 5, follow these steps:
1. In Control Panel, click Regional and Language Options.
2. On the Languages tab, click Details in the Text services and input languages area.
3. In the Installed services area, click Speech Recognition under the language that you are using, and then click Remove.
Note You may also have to delete the following key in the Windows registry:

HKEY_CURRENT_USER\Software\Microsoft\Speech
How can I tell whether speech recognition is turned on?

Speech recognition is turned on when your microphone is turned on and you can see that Dictation or Voice Command is selected on the Language bar. Because the Language bar can be minimized, remember that you can also view the Speech command on the Tools menu. If the Speech command is selected, speech recognition is turned on.

To turn on or to turn off speech recognition, use one of the following methods:
1. On the Language bar, click Microphone.
2. Or, click Tools, click Speech, and then click to select either the Dictation check box or the Voice Command check box.
Note Remember to turn the microphone off when you are not using speech recognition. Speech recognition continues to process sounds until the microphone is turned off. Unexpected behavior may occur if you leave the microphone turned on.
Why is my company name not recognized when I say it?

You can add your company name and other words to the speech recognition dictionary. To do this, click Speech Tools on the Language bar, and then click Add/Delete Word(s).
I installed speech recognition and ran Microphone Wizard. What else must I do for speech recognition to work?

When you install speech recognition and run Microphone Wizard, speech recognition may not work. Speech recognition may not work because of any one of the following reasons:
- You may not have clicked to select the Dictation check box or the Voice Command check box on the Language bar.
- Or, your current keyboard layout may not support the speech recognition engine that is currently selected. For example, the English (U.S.) speech recognition engine requires that you work with a U.S. keyboard layout.
Why is the new keyboard that I added not available in Microsoft Word ?

If you added a keyboard by using the Text Services Control Panel tool while Word was open, you must exit Word and then restart it.
Why does the computer not run a command when I say the command?

This problem may occur for any one of the following reasons:
- Your computer may be in Dictation mode. To verify that this is the problem, click Voice Command on the Language bar, and then try speaking again.
- The microphone may not be turned on. To verify that this is the problem, click Microphone on the Language bar to turn on the microphone.
- You may be working with a read-only file.
- Your microphone may have a mute button. If so, check to see whether it is turned on.
- You may have to change the default speech recognition engine.
- Examine the connection on your microphone. To test the microphone in Sound Recorder, follow these steps:
  1. To start Sound Recorder, click Start, point to All Programs, point to Accessories, point to Entertainment, and then click Sound Recorder.
  2. Click the Record button, and then speak into the microphone.
  3. Click the Stop button, and then click the Rewind button.
  4. Click the Play button. If the sound is played back, your microphone is working.
- Examine the dimmed volume settings on your microphone. See the Help topics in your operating system for more information.
After I adjust the microphone, why am I not able to hear anything when I play back dictated text?

This problem may occur for any one of the following reasons:
- The speaker volume may be too low.
- The speaker may be turned off.
- The microphone may not be correctly connected.
- The microphone may be muted.
- Microphone Setup Wizard may have incorrect information about whether your headset has a speaker.
Why does speech recognition not work with the Office Assistant?

We do not recommend using speech recognition with the Office Assistant.
When I work in another language, why does speech recognition not work the way I expect it to?

If you use speech recognition for a language other than Simplified Chinese, English (U.S.), or Japanese, voice commands do not work. However, Dictation mode might still work.
When I say "right-click" for a spelling error, why does the shortcut menu not contain spelling suggestions?

To view the spelling suggestions, you must move the insertion point to the misspelled word by using your mouse or keyboard before you say "right-click."
Why do I see a blue box on the screen when I dictate?

As you dictate, a blue box appears on the screen while the computer processes the sound of your voice. After the computer processes the sound of your voice, your words appear in your document and the blue box disappears. This is an expected behavior.

Note You may see a dimmed box if your display settings are set to 256 colors or lower.
When I am not using speech recognition, why does text appear in my document and why do menus open?

Most likely, speech recognition was turned on during an earlier session of Word. To turn off speech recognition, use one of the following methods:
- On the Language bar, click Microphone.
- On the Tools menu, click Speech.
Note When speech recognition is turned off, the Dictation and Voice Command buttons on the Language bar do not appear. Also, remember to turn the microphone off when you are not using speech recognition. Speech recognition continues to process sounds until the microphone is turned off.
Sometimes when I say a command, it appears in my document. Why does this occur?

Your computer may be in Dictation mode. To change to Voice Command mode, click Voice Command on the Language bar, and then say the command again.
Why is the file size of my document so large after I save it?

There are two options that save information in your Word document and increase the file size of the document: the Save Speech Data option and the Embed linguistic data option.

The Save Speech Data option saves speech recognition input for recognition processing and corrections. To turn off this option, click Speech Tools on the Language bar, and then click to clear the Save Speech Data check box. The next time that you save your Word document, this data is removed from your Word document.

The Embed linguistic data option affects both speech recognition and handwriting recognition. The Embed linguistic data option can cause you to lose conversion information about non-ink objects. This means that handwriting that appears as text in the document becomes text without handwriting or speech attributes that permit conversion or correction. Also, speech data is a subset of the linguistic data. This means that if you turn off the Embed linguistic data option, you cannot save speech data for recognized text. However, handwriting information is retained. To turn off this option, click Options on the Tools menu. On the Save tab, click to clear the Embed linguistic data check box, and then click OK.
Why does Word not save speech input data in my document?

This problem occurs because of one of the following reasons:
- If you saved the document in an earlier version of Word or in a different format, Word does not save your speech input in the document. Speech recognition is not backward compatible. Word can only save speech input data when you save your document in the Word Document (.doc) format.
- The Save Speech option may not be turned on. To turn on the Save Speech option, follow these steps:
  1. On the Tools menu, click Options.
  2. On the Save tab, click to select the Embed linguistic data check box, and then click OK.
  3. On the Language bar, click Speech Tools, and then click to select Save Speech Data.
Why does my microphone not work in other programs while speech recognition is running?

You cannot use other programs that use a microphone while you use the microphone for speech recognition. (For example, you cannot use Sound Recorder.) A microphone is not a resource that can be shared by multiple programs. You cannot perform multiple tasks with a microphone in different programs.
What is the biggest problem that people experience when they use microphones and speech recognition, and what can I do to resolve that problem?

One of the biggest problems that people experience with speech recognition is noise. Electromagnetic noise that is also referred to as electromagnetic frequencies and background noise are sounds that are not speech data. Electromagnetic noise causes the humming sound on your home amplifier. Electromagnetic noise also causes the humming sound on public address systems when the volume is turned up but nothing is playing. Computers generate some electromagnetic noise, but electric motors and fluorescent lamps are the main causes of electromagnetic noise. Electrical devices that are not grounded correctly and that are on the same electrical circuit as your computer may also cause electromagnetic noise. You can reduce the effects of electromagnetic noise by isolating your microphone in the following ways:
- Use a universal serial bus (USB) microphone.
- Position your microphone as far from the computer as possible.
- Make sure that cables and printers are not close to your computer.
You can also reduce background noise by using a high-quality noise-cancellation microphone. Also, remember to speak as clearly as possible and at a steady rate. Jumbled speech or very fast speech is frequently not recognized and causes inaccurate results.
Are USB microphones better for speech recognition?

Universal serial bus (or USB) microphones have the advantage of having the analog-to-digital converter in the microphone. Therefore, many of the microphones and the converters are far away from the computer. It is good to keep the microphone away from the computer because the computer creates electromagnetic noise. When you maintain distance between the computer and the USB microphone, the USB microphone should, in theory, provide a higher-quality digitized sound. However, for analog uses, such as Sound Recorder, the USB microphone may seem inferior because of the clipping of the digital signal that the microphone produces. A USB microphone uses electricity from the CPU instead of electricity from the processor on your sound card. This can additionally reduce noise and increase accuracy, but it also increases the processing load on the CPU. If you have a high-quality microphone and sound card, but a slower CPU, you might experiment with your system before you purchase a USB microphone. In most situations, the benefits of a USB microphone outweigh the cost. However, if you have a high-quality late-model sound card and a high-quality noise-cancellation microphone, you do not have to have a USB microphone.
Is it important for my sound card to be full-duplex or half-duplex?

No, it is not important. Full-duplex systems let you hear sounds while you create speech input. Most sound cards permit this two-way transmission of information, but you do not have to have a full-duplex sound card for speech recognition.
My hardware is fine and I did the speech recognition training, but I still do not receive consistent results. What might be wrong?

Microphone placement is very important. You must also make sure that your voice and your voice level are consistent from session to session. For example, you do not want the distance between the microphone and your mouth to vary, because the volume of your voice varies with the distance. Therefore, handheld microphones, monitor microphones, keyboard-mounted microphones, and boom microphones do not work well for speech recognition unless you consistently remain in one position. Headset microphones work the best, because you can adjust the mouthpiece so that it is always at the same distance from your mouth. The recommended position for the mouthpiece is a thumb-width away from the corner of your mouth. You do not want the microphone to be directly in front of your mouth.
I have a different speech program that I installed and trained. How does installing the Microsoft Office speech recognition components affect that program?

If the third-party speech program uses a Speech API 5 engine that supports both Dictation, and Command and Control, the third-party speech program is supported in Office XP and Microsoft Office 2003. For example, Word 2003 turns off the third-party speech program if the third-party speech program supports either Dictation, or Command and Control, but not both. And, if the program is the default speech engine in Control Panel, this is also the case. If a non-Speech API 5 program controls the microphone when the operating system starts, speech recognition may not work because the microphone is not a resource that can be shared. Speech training that you completed in a different speech program can only be used in the Microsoft speech engine if the program is Speech API 5-compliant. If the program is Speech API 5-compliant, the speech training that you already completed can be used.

What characters can I insert when I am dictating, and what controls are available in Voice Command mode?

You can say the names of buttons on the menus and toolbars. To determine a button name, point to the button. The ToolTip that appears indicates the name of the button.

Word also supports voice mapping to the keyboard. For example, you can map the following keys to your voice commands: ENTER, ESCAPE, TAB, LEFT ARROW, RIGHT ARROW. In Dictation mode, you can insert frequently used characters with a voice command. The following table lists some of the characters that are supported in Dictation mode:

Spoken word	Inserted Character	Spoken word	Inserted Character
Ampersand	&	Less than	<
Asterisk	*	New line	Carriage return
At sign, at	@	New paragraph	Two carriage returns
Backslash	\	Paren	(
Open bracket	[	Percent	%
Caret	^	Period	.
Close paren	)	Plus, Plus sign	+
Colon	:	Number sign	#
Comma	,	Question mark	?
Dollar sign	$	Single quote, single quotation mark	'
Double dash	--	Quote, open quote	"
Ellipsis	...	Close brace	}
Close quote, End quote	'	Right bracket	]
End quote	"	Semicolon	;
Equals	=	Slash	/
Exclamation point	!	Tilde	~
Greater than	>	Underscore	_
Hyphen, dash	-	Vertical bar
Open brace	{

For more information about voice commands that you can use, click Microsoft Office Word 2003 Help on the Help menu, type voice commands in the Search for box in the Assistance pane, and then click Start searching to view the topic.

Are there any incompatibilities between Speech API 5 and chipsets such as Advanced Micro Devices's 3D Now technologies?

There are no known incompatibilities between Speech API 5 and chipsets such as Advanced Micro Devices's 3D Now technologies. Any processor supports speech recognition to the extent that it supports the Intel standard x86 architecture and Speech API 5.
Are the speech recognition training files backed up by Save My Settings Wizard and Profile Wizard? If not, is there a way to do this manually and to restore to a new system?

The Save My Settings Wizard and the Profile Wizard do not save the speech recognition training profile. It is not very useful to restore user profiles to a different computer where audio channel may have a different character.
Does Word support Text-to-Speech like Microsoft Office Excel does?

Word does not support Text-to-Speech (TTS).

For additional information about workarounds, click the following article number to view the article in the Microsoft Knowledge Base:

287120 How to: automate Excel text to speech in another Office program
Which languages does speech recognition support?

The Microsoft speech recognition engine supports Simplified Chinese, English (U.S.), and Japanese. Additional languages might also be available from various third-party vendors.
Does speech recognition take advantage of a multiprocessor system?

The speech recognition feature benefits from multiple processors in a multiprocessor system. However, there is no explicit support for a multiprocessor system in Microsoft Office 2003, in Microsoft Office XP, or in Microsoft Word 2002. The only support for a multiprocessor system is the support that your operating system provides. For additional information about multiprocessor support, click the following article numbers to view the articles in the Microsoft Knowledge Base:

234558 How to add support for multiple processors in Windows 2000

193645 Windows Me/98/95 Do Not Use Multiple Processors

156358 How to manually add support for a second processor

186562 Terminal Server Multiprocessor Support

REFERENCES

For additional information, click the following article numbers to view the articles in the Microsoft Knowledge Base:

278927 General information on the installation and configuration of speech and handwriting recognition

283160 Handwriting Recognition Frequently Asked Questions

295947 How to troubleshoot speech recognition problems in Word 2002

Additional query words: speech recognition AMD

Keywords: kbfaq KB826853

Microsoft KB Archive/826853

Contents