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

Idiosyncratic Phonetic Representations for Electronic Directories

IP.com Disclosure Number: IPCOM000105034D
Original Publication Date: 1993-Jun-01
Included in the Prior Art Database: 2005-Mar-19
Document File: 4 page(s) / 100K

Publishing Venue

IBM

Related People

Fitzpatrick, GP: AUTHOR [+2]

Abstract

Users of electronic directories or address books may often be able to identify the unique instance of a given individual in the directory (by having known the correct spelling of a person's name, perhaps via delivery of an electronic mail item, for instance) but be unsure as to the "correct" pronunciation of that name. (The correct pronunciation is that determined solely by the individual identified by the entry.) Further complicating the situation is the fact that, because of personal preferences, ethnic idiosyncrasies and family traditions, there is no way to unambiguously determine (from the spelling of the name) that pronunciation.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 47% of the total text.

Idiosyncratic Phonetic Representations for Electronic Directories

      Users of electronic directories or address books may often be
able to identify the unique instance of a given individual in the
directory (by having known the correct spelling of a person's name,
perhaps via delivery of an electronic mail item, for instance) but be
unsure as to the "correct" pronunciation of that name.  (The correct
pronunciation is that determined solely by the individual identified
by the entry.)  Further complicating the situation is the fact that,
because of personal preferences, ethnic idiosyncrasies and family
traditions, there is no way to unambiguously determine (from the
spelling of the name) that pronunciation.

      Existing art which provides for phonetic searches is wholly
inadequate to address this problem, as those processes pre-suppose
that a user already knows the "correct" (or extremely close)
pronunciation.  In fact, the existing art addresses a quite different
problem: "Given that I know how to pronounce this, how is it
spelled?"  That approach presents to the user a candidate list of
often many entries which match given phonetic criteria.

      Provided is a new phonetic representation/determination
technique which provides to the user the individual or
"idiosyncratic" pronunciation associated with an entry in an
electronic directory.  This approach further enables the person
referenced by the directory entry to supply the "correct"
pronunciation.

      The simplest implementation of this approach would allow each
person in a directory to manually build a phonetic entry for each of
his/her names.  Using this approach, the owner of each individual
entry i.e., the person referenced would access a phonetic alphabet to
input the appropriate symbols.

      A more sophisticated and user-friendly implementation,
employing speech-to-text processes, would allow users to speak into a
voice-recognition input device.  Then, sophisticated artificial
intelligence software would produce candidate phonetic
representations for the user to choose from or modify as appropriate.

      Alternately, users could record an audio segment containing
that person's own pronunciation of his/her name(s).  Subsequent users
of the directory entry could then choose an option allowing that user
to hear the audio segment.  This approach eliminates any ambiguity
regarding pronunciation.

      Regardless of which of the above implementations is chosen, a
subtle but important change to the underlying data structure is
called for.  To understand that change, it is critical to recognize
the relational data scheme implicit in this approach, in contrast to
conventional phonetic fields.  A depiction of the two approaches is
shown here.

      Fig. 1 depicts the relationship between conventional phonetic
fields (termed Phonetic Representations in this figure) and their
associated Spellings in a logical data structure.  Importan...