344 



BELL SYSTEM TECHNICAL JOURNAL 



Room noise losses are difficult to determine with any degree of 

 accuracy by observing on working circuits since it is impractical to 

 introduce artificial room noise, and natural variations in room noise 



-8-6-4-2 2 4 6 8 10 



HIGH SIDETONE LOW SIDETONE 



SIDETONE IN DECIBELS BELOW REFERENCE CONDITION 



Fig. 6 — Effective loss due to sidetone relative to working reference condition. 



are, in general, accompanied by other variations affecting repetitions. 

 Laboratory or theoretical methods of determining these losses are not 

 satisfactory since, as room noise changes, the subscriber consciously 

 or unconsciously tries to counteract the effect and such reactions can- 

 not be studied satisfactorily under controlled laboratory conditions. 

 Each method, however, helps to determine the general magnitude of 

 these effects. 



In practice it is seldom possible to vary one parameter over a wide 

 range without causing some variation in the other parameters, but 

 this does not add any serious complication to the derivation of the 

 relations. It is merely necessary to make successive, approximate 

 corrections when deriving each relation. 



After these relations have been established, it is possible to predict 

 the performance of any ordinary circuit by computing the five param- 

 eters from the physical measurements. The magnitude of each pa- 

 rameter may then be compared with the magnitude of the same 

 parameter in the reference system, and the difference between the 

 two magnitudes may be converted into equivalent distortionless 

 changes in volume loss by means of the curves. These ratings may 

 then be combined to give a single effective loss rating of the circuit. 



