Return
to Story Page
To print: Select File and then Print from your
browser's menu
K Sreedevi
Chennai, Aug 24: Linguistic differences, blurring speeches of people living in various parts of the country may soon become a thing of the past if research by IIT, Madras succeeds.
For, the phonetic engine proposed by IIT professor B Yegnanarayana will ensure automatic translation of one Indian language into another. The research, which has now achieved only a conceptual breakthrough, is soon expected to take a concrete shape and be presented as a product.
A phonetic engine is one that converts speech from one language into that of another language. Engines of about 13 different Indian languages are proposed to be developed, which would ensure better interaction among all types of users of the engines.
The basic technology involved in a phonetic engine is matching of speech signals of various languages. The speech in one language is fed into the computer which is converted into a `text' format. The text format refers to a sequence of binary numbers corresponding to the speech signal. This signal isthen matched with a similar word of another language that is already stored in the computer. In the final step, the matched text is re-converted into a speech signal of the second language thus enabling communication.
This identification and conversion of the speech signal is made possible by the `signature' of each word. i.e. each word of a particular language is identified by a few key strokes involved in its pronunciation. This is taken into account while sequencing the words as binary digits in the computer.
This kind of word sequencing, says Yegnanarayana, has been made possible mainly because of the Sanskrit origin of Indian languages. This will be helpful in making the phonetic engines speaker-independent and hence, will improve the chances of access by more number of people than by a set of individuals whose voice format is already stored in the computer system.
Development in this technology can find greater applications in areas such as space technology, medicine, banking, e-commerce andtelecommunications.
However, this technology, which tries to establish human connectivity, faces some major challenges. The main bone of contention is the speed of speech delivery of individuals speaking the same language. Each individual speaks a particular word at a different pace than the other. This has to be matched properly.
Also, the factor of language accent, too has to be reckoned with. Individuals living in various regions of the country can utter the same word in different styles. This calls for further research in these areas to develop systems independent of these factors.
Copyright © 1999 Indian Express Newspapers (Bombay) Ltd.
------------------------------------------------------------
This story was printed from Net Express located at http://www.expressindia.com. Net Express provides a portal to India, with news from The Indian Express and The Financial Express along with sites on travel and tourism, the entertainment industry, the power sector, the environment and much more.
------------------------------------------------------------