122 BELL SYSTEM TECHNICAL JOURNAL 



of the speech soundvS to be investigated and as observers to give indica- 

 tions of loudness and distortion and of their effects on the recognition of 

 the reproduced sounds. Two outstanding kinds of relations of this 

 type are those given by volume tests and articulation tests, which will 

 be discussed later. It has therefore been of great convenience to take 

 a further step and to study and specify the performance of telephone 

 circuits and their parts in terms of their functioning for single-frequency 

 sounds and currents. In this procedure, this functioning is investi- 

 gated for a number of different single-frequency sounds and currents, 

 so taken as to cover the range of frequencies transmitted by the circuit. 

 In the single-frequency transmission characteristics, the personal 

 element is eliminated and the measurements are made entirely on a 

 physical basis. 



A great deal of attention has been given to the correlation of speech 

 sound and single-frequency transmission characteristics so as to enable 

 the former to be derived from the latter and so extend the application 

 of the type which is more readily susceptible to quantitative determina- 

 tion. Also, use has been made of easily specifiable multi-frequency 

 sounds and currents to permit the physical measurement to approach 

 more nearly speech sound conditions, of phonograph ic reproduction to 

 reduce the personal factor in the generation of speech sounds for meas- 

 urement purposes and of meter arrangements to simulate the ear rat- 

 ings of sounds, particularly from the standpoint of relative loudness. 



As a result of the correlation of speech and single frequency charac- 

 teristics, extensive use has been made of determinations at selected 

 typical single frequencies to check the design, installation and main- 

 tenance of lines and other associated circuit elements. 



The widely used volume test is essentially a means of specifying the 

 action of a telephone circuit or its parts, on the relation between the 

 reproduced and impressed sounds from the standpoint of their relative 

 loudness. In this test use has been made for many years of the 

 Standard Cable Reference System and recently of the Master Reference 

 System for Telephone Transmission ^ as references for comparison. 

 These reference circuits with their adjustable trunks provide a means 

 of obtaining different loudness ratios between input and output 

 sounds. By talking alternately over the reference circuit and the 

 one being investigated and adjusting the trunk of the reference 

 system until the output sounds of the two circuits are judged to be 

 equally loud, a specification of the loudness reproduction ratio is 

 obtained of the circuit under investigation in terms of the length of 

 the trunk in the reference system. The effect of a change in the 



'^ See Reference (1). 



