102 BELL SYSTEM TECHNICAL JOURNAL 



(b) Checking the transmission losses or gains iij an extended program 



network by simultaneous measurements at a number of points 

 on particular peaks or impulses of the program wave which is 

 being transmitted. 



(c) The indication of the comparative loudness with which programs 



will be heard when finally converted to sound. 



(d) The indication of a satisfactory level to avoid interruption of 



service due to instantaneous overloads tripping protective 

 devices in a radio transmitter, damage to sound recording 

 systems, etc. 



(e) Sine-wave transmission measurements. 



These services are different in nature and the ideal requirements for 

 an instrument for each may not necessarily be the same. One instru- 

 ment to serve them all must, therefore, be a compromise. From the 

 standpoint of the companies engaged in this development, items (a), 

 {b) and (c) in the above list were considered to be the most important 

 and therefore attention was first directed to the relative merits of the 

 two types of volume indicators with respect to them. 



Aural Distortion Due to Overload 



Tests of volume indicators as overload indicators with aural distor- 

 tion as the criterion [item (a)] had previously been made on a number 

 of occasions and more tests were undertaken during the present de- 

 velopment. The general procedure in such tests is to determine for 

 some particular amplifier the volume level at its output at which dis- 

 tortion due to overloading can just be heard by a number of observers 

 on each of a variety of programs. The volume levels thus determined 

 are read on the various volume indicators which are being compared. 

 The best instrument is considered to be the one whose readings are 

 most nearly alike for all the programs when overloading can just be 

 detected. 



The sole criterion of distortion due to overloading is the judgment of 

 observers, since it is the final reaction on listeners which is of impor- 

 tance. This judgment is not subject to exactness of measurement, but 

 is in fact somewhat of a variable, even with conditions unchanged and 

 with the most experienced observers. For significant results to be 

 obtained, therefore, a careful technique of conducting the tests is 

 required, many observations must be made, and statistical methods of 

 analyzing the resultant data must be employed. 



The arrangement of equipment and circuits used in these tests is 

 shown in simplified form in Fig. 3. A source of program, which may 



