Support

Technical support: frequently asked questions

Click on the question number to see the answer.


Section 1 - Microphone Setup

1.1

What is Microphone Setup?

1.2

When should I run Microphone Setup?

1.3

New inexpensive ThinkPad adapter now available.

1.4

Do I need to use the microphone adapters included with VoiceType?

1.5

Can I use the microphone that shipped with my computer (or the built-in microphone.if there is one)?

1.6

Can any other microphones be used?


Section 2 - General Dictation

2.1

What is isolated-word speech?

2.2

How does the system learn as I correct mistakes?

2.3

Is it possible to add an alternate pronunciation for an existing word in the vocabulary?

2.4

My recognition was good, but now it has deteriorated. Why?

2.5

How do I dictate numbers?

2.6

How do I improve my recognition performance with VoiceType?


Section 3 - VoicePad

3.1

Once I've dictated a document in Voicepad, how do I transfer it to another application?

3.2

What's the difference between dictation macros and navigation/control macros?


Section 4 - VoiceCenter

4.1

How do I change users if more than one user has enrolled?

4.2

What is the Recognition Sensitivity used for under the VoiceCenter Properties Voice page?

4.3

What is the Recognition Performance used for under the VoiceCenter Properties Voice page?

4.4

What is the difference between Current and Always macros?

4.5

What are Select Lists used for in What Can I Say macros?



Section 5 - VoiceType Direct

5.1

When should I use VoiceType Direct?


Section 6 - Quick Dictation

6.1

When do I use Quick Dictation?


Section 7 - Vocabulary Manager

7.1

How do I see my added word list, and when would I want to delete a word?


Section 8 - Enrollment

8.1

What is enrollment and is it required?

8.2

How many sentences should I read for enrollment?

8.3

What if I do not speak with a native accent for the language version I have installed?


Section 9 - Hardware Requirements

9.1

What are the minimum hardware requirements for VoiceType Dictation 3.0?


Answers to questions

Answers section 1 - Microphone Setup

1.1

What is Microphone Setup?
Microphone Setup is a setup wizard, located in the VoiceType Dictation 3.0 folder, that helps you connect the IBM VoiceType microphone headset to your computer, and helps you configure your audio settings. Microphone Setup does the following:
- shows you how to connect the microphone headset to your computer;
- tests your existing audio setup;
- adjusts the microphone volume;
- performs a speech recognition test.

1.2

When should I run Microphone Setup?
You should run Microphone Setup:
- before you begin your first dictation session;
- if your recognition accuracy has deteriorated;
- if you've changed your sound card;
- if you've changed your microphone;
- if you've switched users.

Setting up your microphone correctly is the single most important thing you can do to enhance the use of VoiceType. Proper microphone set up, and correct audio settings are required to achieve the highest recognition accuracy.

1.3

New inexpensive ThinkPad adapter now available.
Some IBM Thinkpads can now use an inexpensive adapter to provide power to the microphone which does not require any batteries. This would replace the battery adapter that you are currently using. These adapters can be purchased from Andrea Electronics. Please be aware that these adapters will work on all models of the following families of Thinkpads: 380, 385, 560, 760, 765 and 770.

1.4

Do I need to use the microphone adapters included with VoiceType?
The Microphone Setup program gives you instructions on what microphone adapters to use, depending on the sound card it detects.

- Power adapter (black): if you have one of the VoiceType microphones Andrea NC-50 and ANC-500, you may require power. Most sound cards provide power to the microphone, but not all sound cards provide this power in the same way. This is why the power adapter (black) is necessary for some sound cards. For instance, the Sound Blaster(**) cards usually require the power adapter (black). However, the other VoiceType microphones (Andrea NC-50u and NC-80u) have the power adapter built in to their tips and is therefore not shipped as a separate item in your microphone package.

- Battery adapter: A few sound cards and some laptop computers (including the IBM ThinkPad range) do not provide microphone power. For those, you must use the battery adapter. If you require a battery adapter, please contact Andrea Electronics (information included with the headset). Try the microphone both with and without the black power adapter, following the instructions in the Microphone Setup program, before deciding to use the battery adapter. Don't forget to install two fresh AAA batteries in the correct orientation into the battery adapter before using it. Even if it seems that your sound card needs the battery adapter, you may not, in fact, need it. Some sound cards have the ability to provide microphone power, but this feature is disabled. To enable the feature, carefully read the documentation for your sound card; you might need to set a jumper or run a setup utility. If you do not have documentation for your card, contact your computer or sound card manufacturer and ask if and how your particular sound card can provide power to microphones so that you can avoid using the battery adapter and having to buy batteries.

- Signal adapter (light gray): this adapter reduces the microphone signal. It is used with sound cards that require a weak microphone signal. The Microphone Setup will give you instructions if this adapter is required.

1.5

Can I use the microphone that shipped with my computer (or the built-in microphone.if there is one)?
We recommend using the microphone supplied with your VoiceType product. This microphone has noise cancelling capabilities and was specifically designed for speech recognition. However, if you have another microphone for your computer, you can use it to make sure that the computer's sound system works before you plug in the microphone for VoiceType. For this pretest, make a test recording of your voice in the Windows Sound Recorder. Again, use this microphone only for testing the sound system -- not for VoiceType speech recognition.

When plugging the VoiceType microphone into your computer, do not use the microphone-in jack on the side of the monitor. Instead, make sure that you plug the microphone directly into your sound card's microphone-in jack. This will help avoid getting any interference from your monitor through the microphone.

1.6

Can any other microphones be used?
You can use many microphones for speech recognition; however, not all microphones will produce satisfactory results. Several microphones have been tested and found to produce dictation accuracy consistent with the microphone included with the product. Please click here to be linked to an up to date list of tested microphones.


Answers section 2 - General Dictation

2.1

What is isolated-word speech?
Isolated-word speech is speaking with small pauses between your words. This form of speech helps VoiceType detect where one word ends and another begins. The Enrollment program presents a recorded example of isolated-word speech to give you a model of correct dictation. You must use isolated-word speech when you dictate memos, letters, and other text using VoiceType. Isolated speech is not required for command and navigation control.

2.2

How does the system learn as I correct mistakes?
As soon as you correct a word, the correction is saved in memory and is available during your dictation session. Whatever dictation method you choose to use, once you close your dictation session, the correction is saved to disk and your speech data is updated with the correct pronunciation of a new word and its usage of the misrecognized word.

2.3

Is it possible to add an alternate pronunciation for an existing word in the vocabulary?
Yes. If a word is misrecognized and you attempt to correct the word, a pop-up message may appear stating that the word is already in the vocabulary. The pronunciation that you supplied is compared to the known pronunciation. If the new pronunciation is similar to the known pronunciation, your personal language model is updated. If the two pronunciations are not similar, a pop-up message will be displayed prompting you for a sounds-like spelling. VoiceType then maintains both pronunciations for the word.

2.4

My recognition was good, but now it has deteriorated. Why?
Several factors could account for this problem:

- The microphone settings might have been changed either manually or by some other sound application. Please run Microphone Setup. This will readjust your microphone settings so that VoiceType will work correctly.

- You might not be using your own speech data. Make sure you use the correct user name that contains your own personal language model. You can use the IBM VoiceCenter Properties in the Control Panel to check and change users.

- You might be positioning your microphone incorrectly. The microphone boom should be angled toward the corner of your mouth (not up or down) so that the flat part on the microphone element under the foam faces the corner of your mouth.

- The Microphone Setup Wizard diagrams the correct positioning of the microphone.

- Make sure that you are not breathing directly into the microphone mouthpiece when dictating. Also, make sure the foam microphone wind screen is attached and in place.

2.5

How do I dictate numbers?
There are two ways you can dictate numbers in VoiceType:

1. To dictate large numbers, like a serial number, you should break them up into two digits at a time. For instance, to dictate the following serial number, 521-44-9655, say fifty-two one hyphen forty-four hyphen ninety-six fifty-five. Initially, the numbers will have spaces between them, but once VoiceType understands the context of this number, by saying another word or ending the sentence, it will glue the digits back together.

2. To dictate numbers as digits (0-9) rather than allow VoiceType to spell out the numbers, say begin numbers. Begin Numbers mode gives you higher recognition accuracy since you are restricting the vocabulary to numbers only. Please note, that this voice command will not be acted upon until you check the 'Enable "Begin numbers" box' under Dictation Options in VoicePad. To end numbers mode and resume dictation, say Resume Dictation. Saying Change that while in numbers mode after a digit has been spoken, will delete the digit. You cannot use the correct error pop-up menu for digits dictated in numbers mode, so we advise making your corrections before returning to dictation mode.

2.6

How do I improve my recognition performance with VoiceType?
VoiceType is speaker independent, so you can use VoicePad to dictate without first training it to recognize your voice (enrollment). Let's assume that you are using the US English version of VoiceType Dictation. If you have set up your microphone correctly, and if you have a native US English accent, you should get more than 90% recognition accuracy for words in the VoiceType vocabulary without enrolling. To gauge your true recognition accuracy, dictate a few paragraphs, say, from a business letter, skipping words, such as names, that are not likely to be in the vocabulary. After you finish dictating, you might find that there are a number of misrecognized words in your text. If you find your recognition performance is lower than 90%, then we recommend going through enrollment. Refer to the enrollment section in these frequently asked questions for further information.

Here are some tips on how to improve your recognition performance:

- Use speech correction to correct a word that VoiceType misrecognizes and teach it how to display the word. Speech correction adjusts your personal speech files so that VoiceType remembers the correction each time you say the word. ONLY use speech correction to correct misrecognized words, do not use it for text editing. To use speech correction, just double-click on the misrecognized word and select the correct word from the correct error pop-up menu. If the correct word is not on the list, simply type in the word.

- DO NOT say your words too s-l-o-w-l-y. Pronounce them naturally as you would in conversation--just don't forget to pause briefly after every word! Enunciate the words to further increase your accuracy. As you become more proficient in pausing briefly between words, you can increase your speaking rate. Experienced VoiceType users can dictate at more than 100 words per minute.

- It is important to speak in a normal tone of voice. If the audio level is low (not reaching the green in the microphone input level), move the microphone closer to the side of your mouth, or rerun the Microphone Setup program to readjust your microphone settings.

- The microphone position is very important for good recognition accuracy. Be sure that the microphone boom bends toward your mouth. The microphone should be about 20 mm (3/4 inch) away from your face. It should be a little to the side of your mouth so that you do not breathe into it.

- If you have enrolled, use the same microphone and sound card that you used for enrollment when you dictate. Although you can move your speech data to another computer, your accuracy will never be as good as on the system you used to enroll. So if you switch computers, sound card, or microphone, and you want the highest possible dictation accuracy, rerun Microphone Setup and re-enroll on the new system!

- If your system has at least a Pentium 133MHz processor, then you can increase recognition accuracy by setting recognition performance to "accurate" under VoiceCenter Properties in the Voice page. Refer to the VoiceCenter Properties section in these Frequently Asked Questions for further information.

- Don't forget that you must say dictation commands such as STOP-DICTATION and NEW-PARAGRAPH continuously as a single word, without pausing between the words in the command!

- If you want to interrupt your dictation, for example in order
to answer the phone or talk to someone, first stop your dictation by saying "Stop-Dictation" or by clicking Stop Dictation in the Dictation menu.


Answers section 3 - VoicePad

3.1

Once I've dictated a document in Voicepad, how do I transfer it to another application?
After you have completed a dictation session in VoicePad and corrected any errors, you can transfer the text to any application that accepts text. With the left mouse button, click the Transfer icon in the smart icon toolbar and drag your mouse pointer to the open application window where you want to transfer the dictated text. The formatting of the document in VoicePad WILL NOTbe preserved in the targeted application. You should format the text after transferring it.

3.2

What's the difference between dictation macros and navigation/control macros?
Dictation macros are used to substitute frequently used text for a short expression, e.g. "business address" could print off the details of your office address. Navigation/control macros are voice-command shortcuts for working with applications and objects on your Windows desktop, e.g. "show calendar" to open the Lotus Organizer diary.

Note: The on-line help gives further information and tips on creating dictation macros as well as navigation/control macros.

- The Dictation Macro Editor allows you to create dictation macros and voice templates which are only active during dictation sessions. Note that although dictation macros are active in all methods of dictation (VoicePad, Quick Dictation, VoiceType Direct and direct dictation into MS Word), voice templates are only active in VoicePad and Quick Dictation.

- The navigation and control macros, shown in the "What Can I Say?" window, are not available during dictation. The IBM VoiceCenter must be running to activate the "What Can I Say?" command and control macros.VoiceType comes with pre-defined macros for commonly used applications, as well as the mini-applications and accessories provided by Windows. You can create additional command and control macros. For example, you could create a macro in which you select the style, font and type size for a document in a word processor.

Answers section 4 - VoiceCenter

4.1

How do I change users if more than one user has enrolled?
Start IBM VoiceCenter Properties from within the Control Panel folder. You must click the User tab and select a new user from the drop down menu. If one user enrolled twice (for example, once in a noisy environment and once in a quieter environment), you can switch enrollments from this menu as well.
Note: If any speech applications are active, you will be prompted to close them.

4.2

What is the Recognition Sensitivity used for under the VoiceCenter Properties Voice page?
The Recognition sensitivity slider does not affect the recognition for dictation, but is used to relax or tighten the acceptance of the pronunciation of words read during enrollment. If the pronunciation doesn't match the expected pronunciation of the word read during enrollment, VoiceType will reject the word by hi-lighting this word in red. VoiceType will be more likely to reject words when this slider is moved closer towards "Exact Match". If during enrollment you get a lot of red words, you can move the slider closer to "Best Guess". This will help reduce the number of red words, by relaxing VoiceType's expected pronunciation of words read during enrollment. This is particularly important for people with heavy accents. This setting is the same as the Match Word to Sound setting under Options in enrollment.

4.3

What is the Recognition Performance used for under the VoiceCenter Properties Voice page?
The Recognition Performance setting is used to increase performance or accuracy depending upon your system. In general, if your system has a lower speed Pentium processor (90MHz, for example), we recommend that you set the recognition performance to "fast". This will speed up recognition processing at the slight expense of accuracy. If your system has a higher speed Pentium processor (166MHz, for example), you should set the recognition performance to "accurate". This will improve recognition accuracy with a minimum degradation of speed. The optimal setting is "balanced".

4.4

What is the difference between Current and Always macros?
- Always macros are always available when IBM VoiceCenter is running independently of the active application. These macros are commands that send a series of keystrokes to the active application or highlighted object. You specify these keystrokes by recording (capturing) them. The Always macros you create are added to the Always page in the "What Can I Say?" Window.
- Current macros are commands that carry out a series of actions in a single application program. These macros are named "current" because they apply only to the application window that was active when the macro was created. You specify these actions by recording (capturing) the actual keystrokes and mouse clicks within the application's windows. Current macros you create are added to the Current page in the "What Can I Say?" window whenever the application has focus. When you say a macro command, Voice Center sends the keystrokes and mouse clicks in the captured sequence to the application.

4.5

What are Select Lists used for in What Can I Say macros?
Macro select lists add useful function to command and control macros. For example, you can use macro lists in the "What Can I Say?" window to create a macro for typing telephone numbers or even currency into a document. To create an always macro for typing currency:
- Say or click Macro and the Macro window appears.
- Say or click Capture.
- Say or click Lists.
- Select $ and cents and then click on OK.
- Say or click Done.
- Insert the word value before the macro list, <$ and cents> in the voice command field then click on OK.
- Train the macro named value and click on Done.
The macro named value <$ and cents> just created should now appear under Customized Always Macros in the What Can I Always Say window.

Now, if you open an application like Wordpad and say in a continuous phrase (without pausing between the words), value two hundred three dollars and forty two cents, 203.42 should display in your active application.

The $ and cents list is identical to the Amounts list, except that it also includes the words dollars and cents. The digits list provides numeric values with high recognition accuracy where the values are spoken as separate digits. For example: for 17.32, say one seven point three two.

If you click on the + sign before the macro list name, it will show you an example and list the available words for this macro.

Answers section 5 - VoiceType Direct

5.1

When should I use VoiceType Direct?
If you choose not to dictate directly into IBM VoicePad or into Microsoft Word, then you may want to use VoiceType Direct to dictate into your favorite application. Since the infirm words do not display in your target application, you will not see the words display in your application until VoiceType has determined, using a trigram algorithm, the correct words. As a result, you will notice a 3-4
word delay, but you will see a status bar showing that it's working.

Note: You must correct your dictation after you say Stop Dictation, but before starting another dictation session. After starting another dictation session, you cannot go back and make changes to previously dictated text.

Answers section 6 - Quick Dictation

6.1

When do I use Quick Dictation?
Quick Dictation is useful for dictating quick notes and brief pieces of information, such as spreadsheet figures, to transfer to another application, the destination application. It does not have the more extensive features of a word processor, such as file (save, open, print), deferred correction, formatting (align text left, center, right), or font-changing functions.

Your destination application must be open and active. With VoiceCenter open, position the cursor in the destination application where you wish to place your dictated text. Then, after you say Begin dictation, the Quick Dictation window appears (unless you are in another dictation window).

Answers section 7 - Vocabulary Manager

7.1

How do I see my added word list, and when would I want to delete a word?
The Vocabulary Manager utility is used to display words
that were added to your personal added word list. On occasion, a word will be added to your personal added word list that will take the place of an already existing word. For example, if the word comma was inadvertently added during word correction, the word "comma" might appear instead of the punctuation mark ",". In your added word list, you can delete the word "comma" and the punctuation mark should again appear during dictation. You might also want to delete misspelled words and words such as proper names or cities that you know you will not use again.

Note: The Vocabulary Manager is included in our latest corrective service (CSD1) for IBM VoiceType Dictation 3.0.0.

Answers section 8 - Enrollment

8.1

What is enrollment and is it required?
Enrollment is the process of teaching VoiceType your voice so your dictated words can be recognized with greater accuracy. Although enrollment is not required (you can simply start using VoiceType right out of the box), doing either the short or full enrollment will help improve your recognition accuracy.

8.2

How many sentences should I read for enrollment?
Although you can start using IBM VoiceType right out of the box, we recommend that you do the short enrollment to improve recognition accuracy. The short enrollment requires that you read 50 sentences. After reading the 50 sentences, you will be prompted to select Train. Unless you have a heavy accent, the short 50 sentence enrollment should suffice. If you decide not to train and continue reading beyond the 50 sentences, you must then complete the full enrollment of 283 sentences before you can process this enrollment. For this reason, we suggest that you choose to train after reading 50 sentences. You still have the option to do the full enrollment by going back and reading the remaining sentences. Even after you have processed your enrollment, IBM VoiceType continues to update your personal language model helping to improve your recognition performance. Whenever you use the Correct Error function to correct recognition mistakes, IBM VoiceType will learn the usage of the corrected word in reference to its language context and will give this word higher rank next time this word is spoken. Essentially, doing word correction allows VoiceType to learn from its mistakes and recognize the corrected word next time you speak it.

To help improve command and control recognition accuracy, we suggest you do the enrollment for commands. You will also be prompted to select Train after reading 50 command phrases. Doing the full enrollment for commands, 190 phrases, will help if you're having recognition problems for commands and control, or have an accent.

8.3

What if I do not speak with a native accent for the language version I have installed?
If you do not have a native accent for that language, or if you have a strong regional accent, and are frequently getting "red" words while reading the enrollment sentences, decrease the recognition sensitivity as follows:
- Click "Options" in the Enrollment panel.
- Move the slider to the left in the "Match word to sound" section by clicking the "Approximate" arrow 3 to 5 times. Click "OK", then click "Yes".

Answers section 9 - Hardware Requirements

9.1

What are the minimum hardware requirements for VoiceType Dictation 3.0?
To run VoiceType Dictation 3.0, you must be running Windows(**) 95 and have at least the following hardware:

- Pentium(**) 100MHz, or equivalent AMD-K5(**) or Cyrix 6x86(**) processor.
- 16MB of memory.
- 30+MB of available (non-compressed) hard file space. Roughly, an additional 30MB of space is required for enrollment.
- Creative Labs(**) Sound Blaster 16 (or 100% compatible) or Mwave(*) sound card.

Refer to our list of systems tested with VoiceType for additional information.

THE INFORMATION PROVIDED IN THIS TECHNICAL SUPPORT WEB SITE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. You assume the entire risk as to the accuracy and the use of the information contained herein.


(*) IBM ®, Mwave ®, Aptiva ®, ThinkPad ® and VoiceType ® are IBM registered trademarks.

(**) Microsoft ®, Windows ®, and NT ®, are registered trademarks of Microsoft Corporation.

(**) All other products are trademarks or registered trademarks of their respective companies.


Back