TRANSMISSION FEATURES OF NEW TELEPHONE SETS 375 



characteristic of the mouth, cavity resonance of the ear, and disturb- 

 ance of the sound field by the head. The individual contribution of 

 some of these factors is not as yet definitely determined. 



Furthermore, for reasons mentioned, it is not self-evident that a 

 practical telephone system of limited frequency range should be "flat" 

 with respect to the corresponding frequency band in this more or less 

 basic orthotelephonic system which is not limited in frequency range. 

 Having decided on the band width that is desirable and justifiable, it 

 must still be determined, therefore, what particular frequency char- 

 acteristics are preferable in this band. 



In selecting from the many possible choices, the particular frequency 

 response that seems best, several factors must be taken into account. 

 This has been done by a study (under the conditions of actual service) 

 of the relative results of several different experimental instrument 

 designs, varying in frequency characteristics. The overall frequency 

 characteristics of the resultant choice are indicated for two typical 

 circuit conditions in Fig. 7. These measurements were made with 

 the artificial mouth and ear ^^ and are plotted with reference to corre- 

 sponding measurements on an orthotelephonic reference telephone 

 system. For comparison, the results of similar tests of the earlier 

 Bell System handset ^^ are shown also. 



In considering these overall telephone system frequency response 

 characteristics in the light of previous discussion, there are several 

 points of interest: 



1. The large increase in response at both higher and lower frequencies 



with respect to the older handset, which in itself was a notable 

 advance in this respect over previous types. This increase 

 amounts to 10 db or more from about 200 to 500 cycles and 

 from about 1,700 to 3,000 cycles. This wider frequency range 

 gives better naturalness of reproduction. 



2. The type of the response. The general uniformity and absence of 



any marked resonance or irregularity is obvious. For either 

 average or long loops the entire band from about 300 to 

 over 3,000 cycles lies within a range of 15 db. It will be noted, 

 however, that, for the average condition, the response at the 

 higher frequencies (1,500-3,000 cycles) is distinctly above that 

 for the frequencies below 1,500 cycles. This characteristic aids 

 materially in the understanding of the low intensity consonant 

 sounds. The response on the longer loops would undoubtedly 

 be correspondingly better if the high frequencies were raised 

 so that the overall characteristic more nearly resembled that 

 for the average condition shown. It should be remembered, 



