332 BELL SYSTEM TECHNICAL JOURNAL 



tests, volume tests and judgment tests and their relation to the repe- 

 tition method. 



The present paper describes the development of this rating method 

 into a system of data for exchange area circuits which gives a con- 

 venient means of determining from the physical makeup of a complete 

 telephone circuit a rating of the effectiveness of the transmission be- 

 tween normal subscribers using the circuit. Such data are called 

 effective transmission data to distinguish them from previous trans- 

 mission data which were based largely on volume losses. The effective 

 data are expressed in terms of the db of effective loss relative to a 

 reference circuit. An effective loss of 1 db is introduced into the 

 reference circuit when the loss of the trunk is increased by 1 db at all 

 frequencies without other change. Any other change in the circuit 

 which has the same effect on its transmission performance as this dis- 

 tortionless change in volume loss also causes an effective loss of 1 db. 

 The equality of performance is judged by the equality of repetition 

 rates. 



The problem of converting repetition data obtained from a rela- 

 tively small number of circuits into usable transmission ratings for the 

 very large number of practical circuit combinations resolves itself into 

 two major parts: first, a choice of the form in which the data should 

 be presented for use in laying out the telephone plant, and second, the 

 actual preparation of the numerical data. 



The preferable form for presenting the data is fixed by the nature 

 of telephone exchange service. The general transmission problem is 

 not to design a complete telephone circuit from one particular station 

 to another particular station, but rather to design each circuit element 

 separately in such a way that any complete circuit made up of these 

 elements will give satisfactory transmission. Thus, each element, such 

 as a subscriber loop or an interoffice trunk, must be designed to work 

 as a part of any one of a large number of different connections. The 

 technique for solving this problem on the basis of volume losses was 

 worked out long ago in a satisfactory way and has been in use for 

 many years. Volume loss data were prepared in convenient form for 

 all available types of circuit elements, with the losses of the elements 

 defined in such a way that when all the component losses were added 

 the loss of the complete connection was obtained. These component 

 volume losses were based, in general, on voice-ear tests or compu- 

 tations showing the effect on the volume of the received sound, of 

 inserting the element to be rated in a reference system. With data 

 set up in this way, it was possible to apportion the permissible overall 

 rating between the different types of circuit elements and then to 



