Description:
This is an effort to design and develop a continuous speech recognition system from the scratch and optionally looking at speech synthesis. Free thinking! I am not going to look at other implementations.It is interesting to understand how that comparison is going on until the correct words are identified in the correct context. Mapping sounds to characters is not sufficient. Joining such characters to form words is also not sufficient. Sometimes words are to be identified with sound patterns rather than with the collection of individual sounds. Strange feedback mechanisms! I am not going to loose!