BANGALORE, August 24: IBM
Software Group has launched the Indian English version of Via Voice for
the OEM market. IBM's Via Voice has been developed to suit the Indian
English accent with the help of Speech and Software (India), a Tata
company.
In order to make the software recognize Indian accent, an Indian
acoustic model was built which contains frequently used Indian words. The
software was developed using 40,000 sentences spoken by geographically
distributed native Indian speakers. The starter gets vocabulary that
consists of 62,000 words occurring frequently in the 16 million sentences
collected from various Indian publications, thereby ensuring the presence
of commonly used Indian nouns and vocabulary.
The words can be segregated to base vocabulary, personal vocabulary and
active or selected topics. Via Voice works by looking for the words in all
the three vocabulary data banks and if it does not find it there, then it
looks it up in the backup vocabulary and adds it to the personal
vocabulary. It also lets one select topics-mini-vocabularies that can be
turned on and off while dictating. IBM Software Group is targeting to sell
this product through PC manufacturers, peripheral distributors, large
publishers and OEM distributors across the country by bundling it with
their offerings.