Letter to the Editor

In the Gitlin and Weinstein paper, Fig. 3 and Formula (15) give the following result: the equalized signal, called u(t), is analytic, i.e., its real and imaginary parts are a Hilbert transform pair. This result is contradictory with the fact that in QAM, Re{w(£)} contains in-phase data a,,, and Im{w(£)} contains independent quadrature data bn. It is worthwhile to make the three following points: (i) q(t) in Formula (13), is not analytic since coc. (iii) On page 280, it should be noted that fsit) should be given by the convolution of V2XB{t)e~j6 with p(t), where 6 = The factor V2 was indeed mentioned in the Appendix, Formula (62). G. Kawas-Kaleh Ecole Nationale Superieure de Telecommunication Department Systemes et Communications 46 Rue Barrault, 75634 Paris, CEDEX 13 Author Response Mr. Kawas-Kaleh is correct in pointing out that q{t) is not an analytic signal. Since r(t) = r(t) -I- jr{t) is an analytic signal centered about wc, the demodulated signal q{t) will be a complex signal having frequency components centered about the origin, and hence cannot be an analytical signal.