|
Automated Translation
INTRODUCTION
Television networks broadcast both live captioned ("real-time")
and pre-captioned programs in dozens of languages across the
world. Caption encoding systems are capable of drawing from
diverse linguistic libraries and transcribing multiple languages.
At present, these captions can only be generated by human
caption-stenographers. Worldwide research into voice recognition
technologies presently focuses on the recognition of English.
There are no hardware or software constraints that make English
easier to transcribe; many other languages have much simpler
phonetic systems. However, there has not yet been market
pressure to focus on transcribing other languages. In the
interim, we are able to automatically translate English captions
into Spanish. We present this model to highlight Spanish as an
example of a language where automated speech recognition would
have a huge potential market. It is our contention that simply
having automated captioning in multiple languages, even without
translation capability, would be in and of itself a powerful
tool. Two-way automatic translation could easily evolve from
this robust environment.
Diagram 4: Automated
Translation System
|
Click on the yellow components for further information.
|
|
ASSUMPTIONS
There are no new assumptions relevant to this diagram.
Back to Diagram
SYSTEM TOUR
We have added a Teletranslator on the
RF signal path before it reaches the TV. This component decodes the English
captions and translates them into Spanish. The system can then
transmit Spanish captions, English captions, or both simultaneously - the user
controls which captions are shown. It should be noted that
this is simply one way to address the translation issue. It is
conceivable that the captions could be translated by the software on the computer
that recognizes speech initially.
COMPONENT DETAILS
Teletranslator
The Teletranslator is a
caption decoder with added features. It utilizes a specially-designed proprietary
database of idiomatic expressions culled from television programs.
The company claims 70% accuracy for translation - a first step on the path to
automated translation. This device exists today and is marketed largely to Hispanic
ESL (English as a Second Language) communities.
Back to Top
"InterCaption: Facilitating Communication Through Captioning"
© K. Acker, T. Lytle, J. Porvin (1998)
Back to VisCom 98
Projects Homepage
|