Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Speech Recognition in Games and Simulators

IP.com Disclosure Number: IPCOM000116118D
Original Publication Date: 1995-Aug-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 2 page(s) / 45K

Publishing Venue

IBM

Related People

Cohen, PS: AUTHOR [+2]

Abstract

Disclosed is the use of a speech recognition system to provide realistic simulations and games by simultaneously using speaker-independent voice models, full-duplex signal processing and sound systems, with on-screen visual prompts. This method is particularly useful, since it frees the hands of the user for other tasks.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 73% of the total text.

Speech Recognition in Games and Simulators

      Disclosed is the use of a speech recognition system to
provide realistic simulations and games by simultaneously using
speaker-independent voice models, full-duplex signal processing and
sound systems, with on-screen visual prompts.  This method is
particularly useful, since it frees the hands of the user for other
tasks.

      With this method, a realistic simulation is achieved as the
system microphone and speaker are left open, while the speaker
projects sound effects and action.  Visual prompts are displayed
on-screen to make the user aware of the currently active vocabulary
and sentences, which may be used with the speech recognition system.
These prompts may be displayed optionally, as on a pull-down display
segment or window to be opened during the first few times the user
operates the simulator.  The vocabulary and grammar of the speech
recognition system varies with the context of the simulation,
eliminating ambiguity in the user's responses.  Since
speaker-independent
speech models are used, the user does not have to train the system to
recognize his speech patterns.

      High-speed context switching is used to change the
probabilities that certain words and phrases are recognized, with the
closest acoustic match being selected as the system "snaps" to the
best fit.  Background and spurious noise is eliminated through the
use of confidence intervals and noise-cancelling microphones.  Also,
the s...