576 BELL SYSTEM TECHNICAL JOURNAL 



constant it simply increases the direct current component of the signal. 

 Similarly if the scene itself contains a source of steady light, this will 

 be visible only in so far as it reflects the scanning beam. 



Turning now to the upper part of the frequency range, experi- 

 mental data on the highest necessary components were obtained by 

 the use of circuits with low pass instead of high pass characteristics. 

 With the television terminal apparatus operating at 17.7 pictures per 

 second, it was found that a filter whose phase distortion had been 

 corrected over practically all of its pass band of 15,000 cycles produced 

 a degradation in image quality which was just detectable when the 

 human countenance was being transmitted. Since the electrical 

 terminal apparatus without the filters would efficiently transmit 

 frequencies higher than this, the experiment showed either that 

 frequencies higher than this were not present in the generated signal, 

 that they were not effectively reproduced, or that they contribute 

 little to the appearance of the image. This upper limit to the useful 

 frequency range for this apparatus is rather lower than was anticipated 

 from the initial survey, but because of psychological factors (decreased 

 discrimination of tone values for fine details, apparent improved 

 resolution when the subject is moving, etc.) it proves satisfactory for 

 television purposes. 



It is of importance, however, to know where the limitation in 

 frequency range occurs in the apparatus and how it might be modified. 

 Considerable information on this point is obtained by studying the 

 nature of the distortion introduced by the aperture in the optical 



system and that introduced by fre- 



quency limitation in the electrical 



part of the system. It is conve- 

 nient to consider them together as 

 the type of distortion turns out to 

 be similar for the two cases. This 

 distortion may be considered most 

 simply in relation to the type of 



fiME signal corresponding to a sudden 



Fig. 13— Elementary signal change unit change in tone value at some 



point in the subject. With an ideal 

 television system in which the instantaneous values of signal current 

 are at all times proportional to the tone values of the points being 

 scanned, the resulting signal would be represented by the graph of Fig. 

 13. Such a consideration involves no real loss in generality as any 

 signal shape may be considered as the result of infinitesimal abrupt 

 changes in intensity. 



« 



I 



