Browse Prior Art Database

Discrete And Continuous Control Using Stable Speech Sounds

IP.com Disclosure Number: IPCOM000119948D
Original Publication Date: 1991-Mar-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 2 page(s) / 54K

Publishing Venue

IBM

Related People

Destombes, F: AUTHOR

Abstract

The device disclosed here allows a user to command actions in a continuous way or a discrete way, or both, by using stable speech sounds and relying on their measurable, continuous aspects such as duration or intensity, as well as on discrete aspects. This is different from the way speech recognition is used in various applications to perform actions via spoken commands, which is only discrete, since the recognition techniques used are based on isolated words, each possible word in the vocabulary used corresponding to a given action. Stable speech sounds are defined here as those which can be pronounced in a sustained way for a few seconds or more, e.g., vowels (/a/, /i/, /o/, /u/, etc), including nasalized vowels in languages like French or Portuguese, unvoiced fricatives (/f/, /s/, /sh/...), voiced fricatives (/v/, /z/, /zh/..

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 59% of the total text.

Discrete And Continuous Control Using Stable Speech Sounds

      The device disclosed here allows a user to command
actions in a continuous way or a discrete way, or both, by using
stable speech sounds and relying on their measurable, continuous
aspects such as duration or intensity, as well as on discrete
aspects. This is different from the way speech recognition is used in
various applications to perform actions via spoken commands, which is
only discrete, since the recognition techniques used are based on
isolated words, each possible word in the vocabulary used
corresponding to a given action. Stable speech sounds are defined
here as those which can be pronounced in a sustained way for a few
seconds or more, e.g., vowels (/a/, /i/, /o/, /u/, etc), including
nasalized vowels in languages like French or Portuguese, unvoiced
fricatives (/f/, /s/, /sh/...), voiced fricatives (/v/, /z/,
/zh/...), and possibly other sounds not usually considered as speech
(whistling...).

      The device consists of:
1.   An acoustic component to measure continuous parameters, such as
intensity, pitch and duration, and other parameters for phonetic
recognition (e.g., a spectrum).
2.   A phonetic recognizer to identify which sound is pronounced
(e.g., which vowel in a set of possible vowels).
3.   A control mechanism to initiate discrete actions (based on the
sound recognized by the phonetic recognizer) or continuous actions
(based on the parameters provided by the acoustic comp...