Cepstrum may be computed either by fft (real(ifft(rcunwrap(log(fft(x))))), or by a recursive approach [Oppenheim & Schaffer, ...) cep(0) = log(x(0)) cep(n) = f(x(0..n), cep(0..n-1)) With the obvious restriction of the first coefficient <> 0.0. Does someone know about similar formulations, but which could handle a zero first point while keeping the minimum-phase property of the input sequence ? (That is, adding something to the input sequence may break this property) Thanks in advance Pascal A. Dupuis
-- Q: How many existentialists does it take to screw in a lightbulb? A: Two. One to screw it in and one to observe how the lightbulb itself symbolizes a single incandescent beacon of subjective reality in a netherworld of endless absurdity reaching out toward a maudlin cosmos of nothingness.