Speech recognition is a very difficult task and it took a lot of years to develop successful implementations. It requires looking at many signatures of the voice, including spectral content and time-domain, and performing advanced statistical analysis on the data.