BACKGROUND NOISE IN BROADCASTING 347 



The action of an automatic volume control, in keeping constant the 

 level of the total carrier delivered to the detector, should become less 

 pronounced as the beat frequency rises and should fail altogether when 

 this frequency reaches the audible range. This reduction in efiticiency 

 of control may either increase, leave unaltered, or decrease the magni- 

 tude of the flutter, depending upon the amount of time delay involved 

 in feeding back the controlling voltage. In the receiver used, the re- 

 duction in efficiency of the gain control occurred between 20 and 40 

 cycles. A comparison of Fig. 9 with Figs. 7 and 8 indicates that in this 

 receiver the gain control tends to increase the flutter somewhat when 

 the heterodyne frequency is within this range. 



Interference of Undesired Program 



When the interfering station transmits a program which is different 

 from that of the desired station, serious interference may occur which 

 is due primarily to the beats between the undesired sidebands and the 

 desired carrier. If the carrier beat frequency is subaudible and there 

 is little or no noise background, this will be the predominant form of 

 interference. Its magnitude will depend upon the degree of modula- 

 tion of the undesired signal, but is practically independent of the type 

 of detector and gain control which are used. In the presence of con- 

 siderable noise background it may or may not be more important than 

 flutter effect. 



In order to get some data on this point, observations were made 

 with a square law detector and manual gain control. This represents 

 about the worst condition, as far as flutter effect goes, but will be ap- 

 proximated by AVC receivers. At a fixed noise level the carrier ratio 

 was determined at which the flutter could be noted, and also the ratio 

 at which the program interference was detectable. This was done for 

 receiver band widths of 7000 and 3500 cycles. The band width had 

 no appreciable effect upon the program interference but exercised a 

 very definite effect upon the flutter. Fig. 10 shows the results of the 

 observations which were taken. The solid sloping curve represents 

 the average of the observations on program interference, while the two 

 horizontal curves show the carrier ratio at which the flutter was just 

 detectable for the two bands widths used. The program interference 

 was classed as audible when it could just be heard on the peaks of modu- 

 lation. However, for considerable intervals of time it was entirely 

 inaudible. Consequently, when the same carrier ratio was recorded 

 for the flutter and for the program interference the former was actually 

 the more annoying. In order to take account of this difference of 

 character between the two types of interference it is necessary to 



