152 BELL SYSTEM TECHNICAL JOURNAL 



ducibility of the results and the ease and speed of making measure- 

 ments. Its disadvantage lies in the difficulty in determining the 

 complex nature of the human hearing mechanism and simulating its 

 characteristics sufficiently well in objective apparatus. 



One of the important steps In the development of the objective noise 

 meters has been the determination of the relative interfering efifects 

 of different single frequency tones. Two types of tests have been used 

 for this purpose, (a) judgment tests and {h) articulation tests. 



Judgment tests usually are set up so that the observer may compare 

 directly two noises in the presence of speech heard over a representative 

 telephone circuit. The magnitude of one of the noises is adjusted until 

 it is judged to be as disturbing as the second noise. The magnitudes 

 which the observer judges to be equally disturbing can be measured 

 and in the case of single frequency tones, the relative weighting which 

 should be applied to the two frequencies may thus be determined. 



An articulation test consists essentially in calling a number of 

 meaningless monosyllables over a circuit to a group of observers, each 

 of whom records the sounds that he hears. The percentage of sounds 

 correctly received is termed the "per cent articulation" for the 

 particular condition tested. On a given circuit, two different noises 

 which produce the same loss in articulation would usually be con- 

 sidered as equally interfering. As before, in the case of two different 

 single frequency noises, measurements may be used to determine the 

 relative weightings to be applied to the two frequencies. 



In 1919, the results of judgment and articulation tests on the 

 relative interfering effects of different single frequency tones were 

 published in a paper by H. S. Osborne.^ Since that time, several 

 other sets of tests of this character have been made in order to check 

 the values previously obtained and to extend the frequency range 

 covered. From the results of all these tests ^ and a recognition of the 

 trend toward more uniform frequency response in telephone message 

 channels, a single curve of relative interfering effects of different single- 

 frequency tones in a telephone receiver has been derived. This is 

 shown in Fig. 1, the curve being labeled " receiver currents." By 

 combining with this curve the frequency characteristic of a repre- 

 sentative transmission path between the toll circuit terminals and the 



^ "Review of Work of the Subcommittee on Wave Shape Standard of the Stand- 

 ards Committee," H. S. Osborne, A. I. E. E. Transactions, Vol. 38, Part 1, 1919. 



^ The articulation and judgment tests mentioned here also contributed largely 

 to the selection, by the C. C. I. F. (international advisory committee on telephony), 

 of a curve of relative interfering effects of single-frequency tones expressed in terms 

 of voltage across the receiver, which it has recommended as a basis for noise measure- 

 ment on international circuits. The weighting given in curve A of Fig. 1, when 

 expressed in similar terms is in conformity with the weighting recommended by the 

 C. C. I. F. 



