System for transferring personalize matter from one computer to another

Abstract

This invention combines methodologies that enhance voice recognition dictation. It describes features for moving speaker voice files eliminating redundant training of speech recognition dictation applications. It defines how to create synthetic voice models reducing speaker dependency. It combines accuracy and performance into a single measure called RAP Rate. Moreover, the invention describes enhancing voice recognition applications and systems by measure / adjusting hardware and software features for optimal voice recognition dictation incorporating methodical processes based on RAP Rate. Using these approaches and tools the invention includes a method for constructing a handheld transcriber that immediately translates audio speech into text with real-time display. The invention describes a method for applying RAP Rate and synthetic voice models to applications like voice mail to text. With the ability to move and translate voice models the invention describes new services that could be provided for a fee.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]This application is a reissue application of U.S. Pat. No. 7,689,416, which was derived from U.S. patent application Ser. No. 10 / 763,966, filed on Jan. 23, 2004, which claims the benefit under 35 U.S.C. §119(e) of U.S. Provisional Patent Application No. 60 / 156,638, filed Sep. 29, 1999, and U.S. Provisional Patent Application No. 60 / 214,504, filed on Jun. 28, 2000, and claims the benefit under 35 U.S.C. §120 as a continuation-in-part of U.S. patent application Ser. No. 09 / 676,328, filed Sep. 29, 2000, now abandoned of which applications are hereby incorporated herein by reference.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT[0002]This invention has been created without the sponsorship or funding of any federally sponsored research or development program.BACKGROUND OF THE INVENTION[0003]1. Field of the Invention[0004]This invention relates to computer voice recognition enhancements. It explains methodologies for measuring ...

AI Technical Summary

Benefits of technology

[0014]This invention includes several components that provide enhancements and ease of use features to voice recognition systems and applications. In accordance with this invention, it is possible to reliably measure accuracy and responsiveness of a voice recognition system used for dictation purposes. With the ability to measure these key metrics other enhancements can then be added to voice recognition systems with a quick and easy determination of system improvement or degradation. One such enhancement described is the ability to move speaker voice models (Voice Modeled Mobility) between systems with the ability to quickly determine the success of a quick user enrollment versus a full training session of a voice recognition system. The measurements can also be applied to a new type of handheld transcriber with internal voice dictation software eliminating the need for a two-step process of recording the dictation notes and then transferring them to the voice recognition software for a speech to text translation. Further advantages can be achieved by applying the RAP Rate measurement techniques to engineering and manufacturing processes resulting in products that have a common reference and relationships providing a known value to industry, users, and customers prior to purchasing the voice dictation product. Applying the RAP Rate measurement techniques with other techniques for determining voice recognition user speech patterns (described in detail later) enables the creation of a new type of speaker voice model or a (Super Voice Model) that can be applied to many people without the prerequisite of training or voice recognition system enrollment. In overview this invention includes components that measure voice recognition metrics (RAP meter), provide ease of use for the movement of speaker voice models (Voice Model Mobility), a handheld transcriber that includes voice recognition software for dictation internal to the transcriber (Powerful Handheld Device), a process for the manufacturing and verification of systems used for voice dictation purposes (RAP Rate Manufacturing Process), a methodology for creating speaker independent voice models (Super Voice Model), and applying these techniques of RAP Meters, Voice Model Mobility, Super Voice Model, and Powerful Handheld Devices to a Audio Voice Mail to Text Translation system. These components and their related features and advantages are described in the description of the preferred embodiment using the drawings as reference.

Problems solved by technology

When using Large Vocabulary Voice Recognition (LVVR) applications in the clone PC environment, two problems are experienced: machine dependency and speaker dependency.

While this approach is typically used throughout the computer industry, it often leaves users frustrated with accuracy and performance of the voice recognition applications.

These approaches have many problems including: No direct feedback while the dictation is taking place, it was not real time large vocabulary voice recognition, training for the voice recognition was cumbersome to accomplish resulting in poor accuracy and user frustration, and training required redundant work since a separate voice model is needed from the desktop speaker voice files.

Moreover, updating the voice parameters and training was typically not possible or very difficult to accomplish resulting in the accuracy level not getting better over time.

Since training can be time consuming and ongoing task and typically results in speaker dependency other inventions have avoided confronting the training and speaker voice models issues needed to accomplish speaker independent and / or mobility between voice recognition dictation systems.

U.S. Pat. No. 6,477,491 Chandler et al. describes needing training for voice recognition applications but does not provide any specific means to accomplish this task and is focused on providing identity of a specific person by the specific microphone they are speaking into.

This investment of time and effort is a per machine cost adding to machine dependency.

When trying to determine and obtain the highest level of system accuracy and performance, one can spend much effort, time, and money trying to determine the best options, (performance and accuracy versus components, effort, and cost).

This has led to frustration and funds wasted with the result being that the speech recognition system is left sitting on the shelf or discarded.

When more than 1 computer is used for voice recognition, accuracy and performance may not be consistent due to different levels of training accomplished for each system.

Images