Text to Speech

Though we have set high goals we are going to start on a moderate scale and continue building on that. Even our moderate program will have many new features such as reading abbreviations and acronyms, rendering numbers and fractions, vocalizing dates and many more.
We would like to mention just one significant aspect of this program. We need to reiterate point that our programs are based on sound theoretical foundations and are developed after intense research. An interesting example is the word ‘aye’ (they came). The pronunciation of this word is different from the from ‘aye’ in the word ‘ayega’ (he will come) or the word ‘aye’ (let him/her/it come). We cannot make the machine produce the same pronunciation in all these cases, as it would sound unnatural. How do we make the machine recognize the difference and behave accordingly? We have found out that the differences can be explained by syntactic analysis and production of sound can be monitored by spelling conventions. Our research not only helps us produce a very effective program, it also helps us reform the spelling accordingly. For a detailed analysis of this aspect of standardization of Hindi please look for the post our website.
The program will be available free to individual users limited to a few hundred words. A professional version with the facility of creating one’s own dictionary will be available for institutions or bulk users.

Projects