Abstracts of Technical Articles by Bell System Authors 



Two papers by Reverend Thomas Bayes — A facsimile publication from the 

 Philosophical Transactions, Vol. LI II, for the year 1763. This facsimile has 

 been prepared under the direction of W. Edwards Deming, Senior Mathe- 

 matician of the Bureau of the Census, Washington, from a copy of the 

 Philosophical Transactions in the possession of the Naval Observatory in 

 Washington. An interesting foreword to the volume has been supplied 

 by Edward C. Molina of the Bell Telephone Laboratories. The volume is 

 available at the Department of Agriculture, Washington, D. C, price $1.00. 



The Subjective Sharpness of Simulated Television Images."^ M. W. Bald- 

 win, Jr. Small-sized motion pictures, projected out of focus in simula- 

 tion of the images reproduced by home television receivers, are used in a 

 statistical study of the appreciation of sharpness. Sharpness, in the 

 subjective sense, is found to increase more and more slowly as the physical 

 resolution of the image is increased. Images of present television grade 

 are shown to be within a region of diminishing return with respect to 

 resolution. Equality of horizontal and vertical resolutions is found to be 

 a very uncritical requirement on the sharpness of an image, especially of a 

 fairly sharp one. 



Synchronized Frequency Modulation."^ W. H. Doherty. Probably the 

 foremost practical problem in FM transmitter design is that of stabiliza- 

 tion of the mean or carrier frequency. Crystal stability is required, but the 

 direct use of a crystal would necessarily give rise to a conflict between the 

 factors which stabilize the frequency and those which are to produce the 

 desired variation. 



In Synchronized Frequency Modulation, which makes its first appear- 

 ance in the 1000- Watt Western Electric 503A-1 Radio Transmitting Equip- 

 ment, this problem is solved by associating the crystal indirectly with the 

 system in a monitoring role which ignores the rapid frequency variations 

 due to modulation and responds only to variations in the mean frequency. 

 This is done by taking a sample of the output of the frequency-modulated 

 electric oscillator and shrinking the spectrum down through a succession 

 of frequency dividers to about l/8,000th of the transmitted carrier fre- 

 quency. It then consists of a strong central carrier (about 5,000 cycles) 

 with a few degrees of phase modulation. This is then compared with a 



^Proc. I.R.E., October 1940. 

 » Pick-ups, August 1940. 



125 



