A question has arisen concerning the similarity of a paper published by M.I. Celebiler and G.M. Coupe (ibid., vol. COM- 26, pp. 257-267, Feb. 1978.) and the above-named one published by the authors (ibid., vol. COM-29, pp. 285-297, Mar. 1981). The first paper considers the problems of cochannel and intersymbol interferences in CPSK systems. In 1981 paper the results of Celebiler and Coupe were extended to the case of quadrature modulated systems. Because the 1981 paper is an extension of Celebiler and Coupe, it naturally makes heavy use of their results. It was an unfortunate oversight on our part that the appropriate amount of credit was not given to Celebiler and Coupe. We did reference the workbut because the method used was very similar, it would have been appropriate to cite it at several points in our paper and perhaps draw more from their results, rather than repeating their procedure. We would like to apologize to Celebiler and Coupe for this oversight and acknowledge the usefulness of their results in interference analyses.

A method is presented for determining the performance of a quadrature-carrier modulation system in terms of probability of error in the presence of additive white Gaussian noise, intersymbol interference, and cochannel interference. This method has been applied to determine the error rate of several quadrature-carrier modulation systems using Butterworth receiving filters of different orders. It has been foand that among the different quadrature-carrier modulation systems studied, sinusoidal frequency-shift keying seems to exhibit the best overall performance in a cochannel and intersymbol interference environment. These results are shown through a number of performance curves that provide useful data for the systems designer.

The performance of a minimum-shift-keying (MSK) system has been studied in which the carrier and clock references are noisy and derived from the received waveform by means of narrowband tracking loops. It has been found that the probability of error is greatly affected by the characteristics of the synchronization structures and that the system is more sensitive to demodulation phase errors than the antipodal PSK system.

The observation by several investigators that the spectrum of a digital FM signal is concentrated near the carrier frequency for small frequency deviation ratios (h) led Tjhung and Wittke [1] to investigate the probability of error to determine the optimum bandwidth-bit period (BT) product and optimumh. Their results using discriminator demodulation showed that for a probability of error of 10-4the optimum values areh = 0.7andBT = 1.0. Further, their results showed that in some narrow-band cases digital FM is superior to PSK. The purpose of this correspondence item is to extend Tjhung and Wittke's results by applying their method to a biphase [2] baseband signal. Such a signal has the well-known advantage of facilitating synchronization since a zero-crossing occurs during each bit.

A linear algorithm for the merli scan is described which exhibits a better quantization SNR than a nonlinear algorithm proposed by Limb and is more easily implemented. The error propagation characteristics for the linear algorithm are similar to those for the Limb algorithm.

The probability of error in a binary communication system using pseudonoise (PN) modulation to spread the transmitted spectrum is influenced by noise, bandlimiting, and synchronization errors. In this paper the average probability of error is determined, using a series expansion of the characteristic function of the intersymbol interference. It is shown that the effects of intersymbol interference are reduced by spreading and that local code synchronization errors become the dominant factor in the probability of error when the system bandwidth is large.

The optimum deviation ratio (h =1/2) and observation interval (2-b periods) for a continuous-phase binary frequency-shift keying (CPBFSK) signal have been determined analytically for particular length sequences based on a squared-distance criterion.

The effect of the integration interval in an integrate and dump detector in a band-limited system is analyzed. It is found for optimum performance that the interval must be delayed by τ seconds, which can be approximated by finding the interval giving maximum area under the system pulse response.

The use of feedback with concatenation is proposed as a technique for achieving the desired performance with less system complexity. Feedback can be used with the inner code, outer code, or both. The concatenation error exponents for several feedback schemes employing the Gaussian channel are determined analytically and found to be significantly better than those of the channel with no feedback.

The probability ofKor more zeros following a one and the covariance function of the output noise bits from the generalized Gilbert channel are determined. These probabilities are used in developing a method for calculating the probability distribution of burst lengths in a sequence ofnoutput bits from the channel. Several applications of the distribution are described.