The Acoustic and Language Model Optimizer dialog box

Acoustic Optimization and Language Model Optimization together are referred to as Accuracy Tuning. Use the Acoustic and Language Model Optimizer to perform Accuracy Tuning and increase your recognition accuracy. The tool adds up to 180 minutes of accumulated acoustic data from your dictations and corrections and any acoustic training you performed to your user profile. It also adds word sequences you often used to the Language Model.

The Optimizer requires data to be present in the acoustic archive to perform Accuracy Tuning. If you attempt to run the Optimizer before the data has accumulated, a message indicating that there is not enough data may appear.

By default, Dragon retains 180 minutes of acoustic data to add to your user profile. You can change this amount on the Options dialog box—Data page.

If you enabled the Roaming feature, the Acoustic and Language Model Optimizer is disabled. For information on running Accuracy Tuning on your Roaming user profiles, see Scheduling Accuracy Tuning for Roaming users.

Because optimization can take some time to complete, you can schedule the Optimizer to run at a convenient time. You can schedule Accuracy Tuning from Administrative Settings or using the Acoustic and Language Model Optimizer Scheduler. For more information, see Scheduling Accuracy Tuning in Administrative Settings and Scheduling Accuracy Tuning with the Acoustic Optimizer Scheduler.

For more information on the Accuracy Tuning process, including prerequisites, see About Accuracy Tuning.

If you enabled Correction Only mode, you must disable it before running Accuracy Tuning.

To open the Optimizer:

From the DragonBar, select Tools>Accuracy Center>Launch Accuracy Tuning.

To view when you last ran the Optimizer:

  1. From the DragonBar, select Settings>Profile>Manage User Profiles.

    The Manage User Profiles dialog box opens.

  2. Select a user profile, and then click Properties.

    The Acoustic optimizer field indicates when it last ran for this user profile.

Option Description

Perform Acoustic Optimization

If selected, updates your user profile with accumulated acoustic data from any dictations, corrections and training you have done.

If deselected, Acoustic Optimization does not run.

Default value: Selected

Perform Language Model Optimization

If selected, updates your Language Model, which contains statistical information that predicts the words that are most likely to occur in your dictation. Language Model Optimization uses text extracted from your user profile's DRA files to add word sequences you used often to the Language Model.

If deselected, Language Model Optimization does not run.

Default value: Selected

Status

Progress of the optimization process.

Estimated time remaining

Estimated time to finish the adaptation of your user profile.

 

Related procedures Additional information
Scheduling Accuracy Tuning in Administrative Settings About Accuracy Tuning
Scheduling Accuracy Tuning with the Acoustic Optimizer Scheduler About acoustic models
  About the language model
  Improving accuracy