NAFI TR-li+^0 



the noise and "bandlimiting the input. 



Note that within this constraint there are still considerable 

 gains to be achieved by niakin;T intelligent use of sigiial spectral infor- 

 mation. The numerical results contained in this report (Figures ^4 and 5) 

 quite possibly are not applicable to the realistic sonar situation, but 

 the technique of analysis is. Thus, one could apply equations 1. 2, and 

 10 to more realistic estimates of the various sonar parameters (T(f) , 



H(R,f), W^(f), etc.), examining the influence of perturbations in these 

 quantities upon system performance (the output 3I\[R) with different operat- 

 ing bands. Such a study would be more profitable than that carried out 

 here since, first, the spectra would be more representative of the real 

 world, and second, the study would determine both fi and W, i.e., both the 

 location and width of the operating band. Such a comprehensive study may 

 appear formidable and lengthy, but the utilization of digital computer 

 techniques should reduce the effort required to an acceptable level. 



Thus we return once more: an intelligent choice of operating 

 band is sufficient. Such a conclusion has undoubtedly been bantered about 

 intuitively for some time, so that some would question the value of the 

 material contained in this report; however, the quantitative results pre- 

 sented here for simple spectra should provide a firmer basis for such an 

 intuitive judgement, and as such, provide firmer ground for the conventional 

 system designer to stand on. 



3h 



