REFERENCE SYSTEM FOR TELEPHONE TRANSMISSION 557 



element. An observer, located in a quiet room removed from the 

 systems or instruments under test, determined when the two systems 

 under comparison gave output sounds of equal loudness. Each 

 testing team, consisting of a speaker and observer, made 12 balances. 

 In order to attain a suitable precision in the final results of these tests, 

 a large number of testing teams were used, the number being largest 

 for the tests where the difference in the quality of the output sounds 

 of the two systems under comparison was greatest. For example, for 

 the first series of tests, six teams were used, for the fifth 25 teams, and 

 for the eighth series, where the master reference system without 

 distortion was compared with the standard cable system, balances 

 were made with 37 teams. In all, over 2,000 individual balances 

 were made. The standard deviation of the determination for each 

 of the first five series is of the order of 0.5 db and for each of the last 

 three series about 1 db. 



The response characteristics of the transmitter and receiver elements 

 of the master reference system, when adjusted on the basis of the 

 results of the voice tests, to be equivalent on a volume or loudness 

 basis to the corresponding parts of the standard cable reference 

 system, are shown in Fig. 10. The mean values weighted from the 

 standpoint of importance for volume are 0.027 volt per bar for the 

 reference transmitter, 16 bars per volt for the reference receiver and 

 0.43 bar per bar for the complete master reference system with db 

 in the reference line. Further consideration is being given to the 

 values of these response characteristics of the reference transmitter 

 and receiver to be adopted as standards. 



The response characteristics of the reference transmitter with trans- 

 mitter distortion network No. 1 and of the reference receiver with 

 receiver distortion network No. 1, when these elements are adjusted 

 on the basis of the above voice tests to be equivalent on a volume 

 or loudness basis to the corresponding parts of the standard cable 

 reference system, are shown in Fig. 11. 



Application of the System 



The results of articulation tests over the master reference system 

 when adjusted for optimum volume are practically equivalent to 

 those obtained in direct air transmission in a quiet room. This 

 system and replicas of it, are particularly adapted for use in making 

 articulation studies, since they provide an approximately ideal system 

 with which the loudness of the output sounds can be varied distortion- 

 lessly over a wide range and in which distortion networks of various 

 types and controlled amounts of noise can be introduced. In this 



