Speech amplitude and zero crossings for automated identification of human speakers
D. Wasson, Ross J. Donaldson · IEEE Transactions on Acoustics Speech and Signal Processing · 1975
Speech amplitude, low-pass (below 1 kHz) and high-pass (above 1 kHz) zero-crossing rates were represented as weighted sums of orthonormal functions. The weighting coefficients were used to identify unknown speakers. For a ten speaker population recognition rates as high as 96.6 percent were obtained using the same data for training and testing.