摘要 |
A method for real-time recognition of speech and call-progression tone energy includes the steps of digitizing an analog signal present on a signal line, correlating N neighboring portions of the digitized signal, performing FFT analysis on the digitized signal, identifying the three largest frequency-domain maxima, determining whether the two largest maxima are above a threshold frequency, and determining whether the ratio of the largest to the third largest maxima exceeds a predetermined value. According to the present invention, the steps of the method may be performed in real time using fixed-point hardware by approximating the correlation and FFT functions.
|