Now you can see what is necessary for telephonic transmission. Each and every component note which enters into human speech must be transmitted and accurately reproduced by the receiver. More than that, all the proportions must be kept just the same as in the original spoken sound. We usually say that there must be reproduced in the air at the receiver exactly the same “wave form” as is present in the air at the transmitter. If that isn’t done the speech won’t be natural and one cannot recognize voices although he may understand pretty well. If 165there is too much “distortion” of the wave form, that is if the relative intensities of the component notes of the voice are too much altered, then there may even be a loss of intelligibility so that the listener cannot understand what is being said.
What particular notes are in the human voice depends partly on the person who is speaking. You know that the fundamental of a bass voice is lower than that of a soprano. Besides the fundamental, however, there are a lot of higher notes always present. This is particularly true when the spoken sound is a consonant, like “s” or “f” or “v.” The particular notes, which are present and are important, depend upon what sound one is saying.
Usually, however, we find that if we can transmit and reproduce exactly all the notes which lie between a frequency of about 200 cycles a second and one of about 2000 cycles a second the reproduced speech will be quite natural and very intelligible. For singing and for transmitting instrumental music it is necessary to transmit and reproduce still higher notes.
What you will have to look out for, therefore, in a receiving set is that it does not cut out some of the high notes which are necessary to give the sound its naturalness. You will also have to make sure that your apparatus does not distort, that is, does not receive and reproduce some notes or “voice frequencies” more efficiently than it does some others which are equally necessary. For that reason when you buy a transformer or a telephone receiver it is 166well to ask for a characteristic curve of the apparatus which will show how the action varies as the frequency of the current is varied. The action or response should, of course, be practically the same at all the frequencies within the necessary part of the voice range.
Cf. Chap. VI of “The Realities of Modern Science.”
My thanks are due to Professor D. C. Miller and to the Macmillan Company for permission to reproduce Figs. 79 to 83 inclusive from that interesting book, “The Science of Musical Sounds.”