i 



THE L3 SYSTEM — EQUALIZATION AND REGULATION 837 



The technical problem of determining the equalization states of the 

 system is normally solved by sending some kind of signals over the sys- 

 tem and observing the effect of the system on those signals. The raw 

 data are usually in the form of loss and delay as a function of frequency. 

 On the basis of these data, the equalization operator desires to correct 

 the system by means of some equalizers which have adjustable trans- 

 missions and delays as a function of frequency. Thus the operator has a 

 group of controls to be operated plus some data which has encoded in it 

 the information as to the proper adjustment of each control. From these 

 data, and a knowledge of the effect of each control, the operator must 

 suitably compute the proper adjustments. As this may be too compli- 

 cated a process to attempt on a trial and error basis, (or by numerical 

 methods), it is quite an obvious advantage to the operator to receive 

 the data as to the state of the system, not in its original form, but in the 

 form of the necessary adjustments to his equalizer controls. This new 

 form of the data simply represents a decoding process based on the 

 available controls. An operator with the same data, but different and 

 perhaps more complicated equalizers, would need the data in a form 

 suited to his different equalizers. 



Consequently : 



Rule Ilia 



The information as to the state of the system may best be presented to the 

 equalization operator in the form of the necessary adjustments of the avail- 

 able equalization controls. 



This rule has a closely related corollary which is based on the fact 

 that the available equalization controls determine the amount of in- 

 formation that is needed. For example, if the independent gain equaliza- 

 tion controls are "n" in number, measurement of the gain of the system 

 at "n" suitably chosen frequencies is sufficient to determine the settings. 

 (If the controls are not independent, fewer than *'n" frequencies need be 

 measured.) This is, of course, a restatement of the fact that ''n" un- 

 knowns may be determined by solution of "n" independent simultaneous 

 equations. The unknowns are the equahzer settings and the simultaneous 

 equations are the relationships of the shapes controlled by each equahzer 

 to the total system error. 



Thus: 



Rule 1 1 lb 



In general, the necessary and sufficient condition for the determination of 

 'n" independent equalization control settings is the knowledge of the system's 



