Customizing Speech Recognition

After you've tried using speech recognition to dictate and issue voice commands, you might decide that you want to customize it a bit. For example, you could create an additional profile for another location in which you work, customize audio settings, add words to the speech recognition dictionary, or perform additional voice training.

TIP

If you're the only person using your computer, you might not think multiple profiles would be useful. But if you're mobile and work in a variety of locationshome, office, in the field, and so forthcreating a profile for each location can help increase accuracy. Even if you're working in a very temporary location such as a hotel room, it's worthwhile to create a new profile if you're planning to use speech recognition.

Working with Speech Recognition Profiles

A recognition profile stores the speech datarecords and analyzes when you do voice training. When you set up speech recognition for the first time, PowerPoint automatically records your speech in the default user profile. If you're the only person using speech recognition on your computer and you use your computer in only one location, this might be the only recognition profile you need.

You can, however, create multiple profiles in PowerPoint. This is useful if more than one person uses your computer, but it can also be useful if you want to use speech recognition with different microphones or in different environments (for example, if you have a notebook computer and work in a variety of locations). Creating a separate profile of your speech under each of these conditions helps ensure that the speech recognition tools take these environmental changes into consideration as well.

Creating a New Speech Recognition Profile

If you want to create a recognition profile other than the default profile, follow these steps:

1. Click the Tools button on the Language bar, and choose Options from the submenu that appears. The Speech Input Settings dialog box opens. Click the Advanced Speech button. The Speech Properties dialog box opens, as shown in Figure 8.12.

   Figure 8.12. Adding a new speech recognition profile is just one of the tasks you can perform in the Speech Properties dialog box.

   

2. On the Speech Recognition tab, click the New button. Figure 8.13 illustrates the Profile Wizard dialog box that opens.

   Figure 8.13. Give your new profile a name .

   

3. Enter your name in the Profile text box. If you're creating multiple profiles of your speech in different environments, you could enter something like "Pat SmithOffice."

4. Click Next to continue to the first step of the Microphone Wizard.

5. Continue on from step 3 in the earlier section "Setting Up Speech Recognition" to complete setup, which is nearly identical to setting up your first speech recognition profile (the default user).

Modifying Recognition Profile Settings

To modify recognition profile settings, click the Settings button on the Speech Recognition tab of the Speech Properties dialog box. The Recognition Profile Settings dialog box will open , shown in Figure 8.14.

Figure 8.14. Customizing settings is another option.

In this dialog box, you can set specifications on pronunciation sensitivity and accuracy versus recognition response time, as well as choose to have the system automatically adapt to your voice to increase accuracy.

To restore your default settings, click the Restore Defaults button.

Deleting a Speech Recognition Profile

To delete an existing recognition profile, return to the Speech Recognition tab of the Speech Properties dialog box. Select the profile you want to delete from the Recognition Profiles group box and click the Delete button.

Performing Additional Voice Training

You can perform additional voice training when you set up speech recognition by clicking the More Training button before you complete the setup process. You also can return at a later time to do additional training. The more samples Office has of your voice, the better it recognizes your speech. You don't need to do all the training at once, however. It's often a good idea to do the initial training, see how accurate your results are when you use the speech recognition tools, and then go back and do more training.

To continue with additional voice training at a later time, click the Tools button on the Language bar and choose Training from the submenu. The Voice Training dialog box shown in Figure 8.15 opens for the current user. Here you can choose from additional training sessions.

Figure 8.15. Additional voice training helps increase your accuracy.

NOTE

To change the current user, click the Tools button on the Language bar and then choose a user from the Current User menu.

At first, your accuracy might only be around 85%90%, but after additional training it should increase to around 95%.

Adding Words to the Speech Recognition Dictionary

You can add unusual words that might not be in Office's speech recognition dictionary. This is particularly useful if you want to add proper names, company names , or industry termsanything unusual that might not be in the standard dictionary. To add and delete words to this dictionary, follow these steps:

1. Click the Tools button on the Language bar.

2. Click Add/Delete Word(s) to open the Add/Delete Word(s) dialog box, shown in Figure 8.16.

   Figure 8.16. Add uncommon or unfamiliar words to the dictionary.

   

3. Enter the word you want to add to the dictionary in the Word text box.

4. Click the Record Pronunciation button and speak the word into the microphone. When the speech recognition engine recognizes the word, it adds it to the dictionary list.

5. To delete a word from the dictionary, select it in the list and click the Delete button.

NOTE

To spell out an unfamiliar word, say "spelling mode" when you're in Dictation or Voice Command mode and then spell the word letter by letter.

Changing to Another Speech Recognition Engine

If you plan to use speech recognition in more than one language, you need to switch speech recognition engines. To do so, follow these steps:

1. Click the Tools button on the Language bar and choose Options from the submenu that appears. Then click the Advanced Speech button. The Speech Properties dialog box opens (refer to Figure 8.12).

2. On the Speech Recognition tab, choose the alternative engine from the Language drop-down list. If you're speaking U.S. English, the default engine will be Microsoft English (U.S.) v6.1 Recognizer.

3. Click OK to close the dialog box.

You'll need to restart PowerPoint for this change to take effect.

Changing to Another Speech Recognition Profile

If you've created multiple speech recognition profiles, you need to switch to the appropriate profile before using PowerPoint's speech recognition tools. To change profiles, click the Tools button on the Language bar, choose Current User, and then select the profile you want to use. The submenu displays all profiles you've recorded (default, those for other users, those for other environments such as office or home, and so forth).

Adjusting Audio Input Settings

On the Speech Recognition tab of the Speech Properties dialog box, you can adjust your microphone's audio input settings by clicking the Audio Input button. Figure 8.17 shows the Audio Input Settings dialog box.

Figure 8.17. Adjust audio input settings if you like.

On this dialog box, you can choose the preferred (default) audio input device or select an alternative device from the drop-down list.

NOTE

To choose an audio input line, click the Properties button. To adjust the volume, click the Volume button. Unless you have advanced knowledge of audio systems or settings, you'll probably want to use the defaults.