Ella and Jake, this is for you:
Although I am sure I would not understand any of the speech analysis that goes into it, the implications of the recent research by the Institute of Radio Engineering and Electronics, at the Russian Academy of Sciences are quite clear: speech recognition is advancing nicely.
The article on Science Blog says that developers have found a way to get their software to filter out parts of the sound that are irrelevant to the meaning of the words, such as fluctuations from emotion or distortion from whispering. What this means is that the program can maintain a much smaller database, since it doesn’t have to allow for so many variations per word.
