826 BELL SYSTEM TECHNICAL JOURNAL 



It was found, however, that the type of relation shown in Eq. 1 

 also relates the data obtained with the two techniques. In this case 

 the relation may be expressed as follows: 



(1 - S^) = (1 - S)-'-, (2) 



where Sz = decimal value of syllable articulation when called con- 

 nectedly, 

 6" = decimal value of syllable articulation when called singly. 



The curve of Fig. 5 is a graph of the above equation. 



In this case uniformity of experience with the various types of 

 distortion does not enter, as the tests with the two techniques 

 were made simultaneously. The only difference in the techniques 

 was that in the three-syllable case the observer listened to three 

 syllables before writing them down. It seems reasonable to conclude, 

 therefore, that when a crew has the same experience with different 

 types of distortion, then the results obtained by it at one time may 

 be compared with the results obtained by it at some other time by 

 using such a relation. No doubt other types of functions could be 

 found which would also fit the above data. The relation shown here 

 was chosen because it fit both the practice data and the data that 

 were obtained with the two different techniques and is very convenient 

 to use in making such corrections. 



It is evident that in order to use the practice curves it is necessary 

 to set up a reference circuit in order to obtain an appropriate value 

 of X. Theoretically, one reference condition should be sufficient, 

 provided that the practice of the crew had the same relative distribu- 

 tion over various types of speech distortion. Since this is usually not 

 the case, it is necessary to use several reference circuits representing 

 various types of speech distortion. When it is desired to correct 

 data for practice effects, the appropriate value of x is determined by 

 making tests upon the reference circuits having types of distortion 

 similar to the circuits for which the corrections are desired. A 

 description of several reference or control circuits which have been 

 found useful with the values of sound and syllable articulation as 

 obtained with the testing crew of five men and five women as previously 

 described, is given below. 



(a) Air Transmission. Master Reference System for Telephone 

 Transmission. — The air transmission tests were made in a quiet, well 

 damped room having a volume of approximately 1,000 cubic feet. 

 The observers faced away and were located at an average distance of 

 30 inches from the speaker. Sound articulation "L" 99.1 per cent. 



