ARTICULATION TESTING METHODS 



825 



standing of some other type of distorted speech. With this in mind 

 it will be seen that a constant value of x fits the data for the various 

 types of distortion reasonably well. In the case of changing speech 

 levels with a constant type of speech distortion, where the question 

 of uniformity of experience is not so important, the fit is even better. 

 It is reasonable to suppose that an inexperienced observer must 

 make a greater mental effort than an experienced observer to obtain 

 the same articulation values. In other words the element reflected 



80 



uj 60 



_i 

 o 



z 



JO 



I 



z 



9. 40 



20 



'0 20 40 60 80 100 



ARTICULATION -THREE SYLLABLES IN SUCCESSION- S3 



Fig. 5 — Relation between techniques 



by these curves is closely associated with the burden or strain upon 

 the observer. A somewhat analogous situation obtains when tests are 

 made with two different techniques which differ primarily in the 

 burden imposed on the observer. In making the filter tests described 

 above, two techniques which differed in this respect were used. One 

 was the standard technique, in which one syllable was called with 

 the introductory sentences. In the other, the syllables were called 

 in groups of three (three in succession) with the sentences. The 

 syllables were uttered as nearly in the manner of a three syllable 

 word as was possible. 



The results that were obtained with the two techniques are shown 

 in Fig. 5. When the syllables are called in groups of three, the 

 articulation values are smaller than when they are called singly. 



