Data processing: speech signal processing – linguistics – language – Speech signal processing – Application
Reexamination Certificate
1999-02-02
2002-06-11
Korzuch, William (Department: 2641)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Application
C704S257000, C379S071000
Reexamination Certificate
active
06405171
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to the field of voice communication applications. More particularly, this invention relates to dynamically loadable phrase book libraries that allow voice applications to “speak” in different languages.
BACKGROUND OF THE INVENTION
Voice communication applications often include voice mail systems, automated menu systems, or voice information retrieval systems. A voice mail system may “speak” to callers by replaying a stored greeting message when a telephone call is first answered by the voice mail system. For example, a greeting message in English may take the form of “Welcome to your voice mail messaging system.” The voice mail system may then present the caller with an option menu that may allow callers to retrieve messages or record new greetings.
Automated menu systems is another example of voice applications. These are often employed by companies to help callers reach the right department or person to meet their needs or resolve their problems quickly and efficiently. For example, these systems may greet callers with the message “To speed your call, press ‘1’ for Sales, press ‘2’ for Customer Service, or wait and an operator will assist you.” Use of automated menu systems allows companies to handle increased calling volumes without having to hire additional operators.
And yet another example of voice applications is voice information retrieval systems. These allow callers to use their telephones to retrieve specific information without human assistance. Banks often use these systems to allow customers to access their banking account information 24-hours a day. The system may prompt users for input and provide them with options to guide users in retrieving the desired information. These options may allow retrieval of account balance, amounts and dates of withdrawals, or amounts and dates of deposits.
Considering these various voice communication applications then, it becomes clear that many times the “speaking” capabilities of these applications must extend beyond simple greetings to encompass more complex phrases. A voice mail system, for instance, must inform the caller as to the number of new messages by speaking the phrase “You have <n> new [message/messages],” where <n> is the number of new messages at that instant time, and the word “message” is used when <n>=1, but the word “messages” is used when <n> does not equal to 1.
It is usually not feasible to store a recording of this phrase for every possible value of <n>. Instead, voice recordings in a particular language, such as English, would typically be made for the phrases “You have,” “new message,” and “new messages,” as well as for the words “no,” “one,” “two,” “three,” etc. These recordings are called “utterances.” A programmer with knowledge of the English grammar would then write software to select the correct utterances that are replayed, as well as the order in which they are replayed, based on the value of <n>.
If the voice application must speak in a language other than English, the recorded utterances are typically replaced one-for-one with the new language. This may work for similar languages, like two Romance languages that share the same syntax and grammar. However, this one-for-one translation will not work for languages that have different grammars. These may require that different utterances be recorded or that the order in which the utterances are replayed be changed. For instance, in the prior voice mail application example, the application will select the word “messages” if <n>=0, to speak the phrase “You have no new messages.” However, if this sentence were to be translated in Spanish, a word-for-word substitution would not result in an accurate translation because, among other things, the Spanish equivalent of the word “message,” and not “messages,” must be utilized when <n>=0, to create the sentence “Usted no tiene ningun mensaje.”
The application software, therefore, must be changed in order to support new languages. Such modifications take time and incur additional expenses, which thereby limit the multinational capability of existing voice applications in an increasingly international marketplace. For example, for applications like voice mail systems which aspire uptime percentages of greater than 99.9%, it is undesirable to have to “waste” any downtime for loading new or updated software to add or modify the language of the system.
Therefore, it would be highly desirable to have a system which would allow voice applications to add or modify languages used by the application without stopping.
REFERENCES:
patent: 4365315 (1982-12-01), Jamnik
patent: 5384701 (1995-01-01), Stentiford et al.
patent: 5561800 (1996-10-01), Sabatella
patent: 5745881 (1998-04-01), Tada et al.
patent: 5765131 (1998-06-01), Stentiford et al.
patent: 5884247 (1999-03-01), Christy
patent: 5983182 (1999-11-01), Moore
patent: 6044337 (2000-03-01), Gorin et al.
patent: 6058166 (2000-05-01), Osder et al.
patent: 6061646 (2000-05-01), Martino et al.
patent: 6173266 (2001-01-01), Marx et al.
patent: 0 121 071 (1984-10-01), None
patent: WO 94/11811 (1994-05-01), None
Azad Abul K.
Korzuch William
Rode Lise A.
Starr Mark T.
Unisys PulsePoint Communications
LandOfFree
Dynamically loadable phrase book libraries for spoken... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Dynamically loadable phrase book libraries for spoken..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dynamically loadable phrase book libraries for spoken... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2934826