InvTool is a program that allows a person to capture his or her own voice for use in speech synthesis. Using InvTool, users record a carefully selected inventory of "sentences" (actually sentences, phrases, and isolated words) designed to cover almost all of the different combinations of speech sounds found in naturally occurring English. Once the person is finished recording as many of the sentences as he can, InvTool compiles the speech into a database that is used by InvTool’s sister program, ModelTalker. ModelTalker can then be used to synthesize sentences that sound like they were spoken by the original individual.

This document is an illustrated tutorial for how to use InvTool. We recommend that you read at least Sections 1-7 below before you begin any recording.


BUG: Currently, there is a minor bug in the InvTool program. Oftentimes, if you move the InvTool window, or if another window is moved over the InvTool window, the writing in the Performance Meter boxes will disappear. This can be alleviated by minimizing and then re-opening the InvTool window.

NOTE to users of older versions of InvTool: Versions of InvTool made available after March 4, 2002, use named inventory files and speech directories which it creates with the File->New menu. If you have already started recording a speech inventory with a version of InvTool downloaded before 3/4/2002, read this to convert your old inventory into the new format.


Updated July 20, 2002.



  1. Installation
  2. Microphone Set-Up And Calibration
  3. Creating A New Inventory File
  4. How To Record Sentences
  5. Listening To And Re-Recording Your Sentences
  6. Stopping, Re-Starting, And Tracking Your Progress
  7. General Recording Guidelines
  8. Advanced Feature: Editing the Inventory
  9. Advanced Features: The 'Settings' Menu
  10. When You Are Finished Recording: Making The Synthetic Voice
  11. Uploading your Recordings to the Speech Research Laboratory