About Accuracy Tuning

Accuracy Tuning is a process that increases your recognition accuracy. This process consists of the following:

• Acoustic optimization—Updates your user profile with the following:

  • Up to 180 minutes of acoustic data from dictation and any corrections you make, including corrections with the keyboard.
  • Words on which you trained Dragon.
  • Acoustic training you performed by reading text.
• Language model optimization—Updates your language model with frequently dictated word sequences extracted from your user profile's DRA files.

You run Accuracy Tuning using the Acoustic and Language Model Optimizer (Audio>Launch Accuracy Tuning).

• Set the following options on the Options dialog box—Data page:

  • Store corrections in archive—Select to enable.
  • Archive size—Optionally change the amount of acoustic data Dragon retains to add to your user profile. Default value is 180 minutes.

• Optionally select the following option on the Options dialog box—Data page:

  • Run Accuracy Tuning at the time scheduled by your Administrator—Select to enable. For more information on scheduled accuracy tuning, see Scheduling Accuracy Tuning in Administrative Settings.

• Optionally select the following options on the Administrative Settings dialog box—Scheduled Tasks tab:

  • Enable scheduled Accuracy Tuning—Select to enable.
  • Configure—Click to configure intervals and time.

• Optionally select the following option on the Administrative Settings dialog box—Miscellaneous tab tab:

  • Disable automatic acoustic model selection in the Acoustic Optimizer—Select to prevent Dragon from choosing a different acoustic model for your user profile during Accuracy Tuning.

Additional information