786 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1953 



which were made to determine the maximum permissible magnitudes 

 consistent with other important message circuit objectives. 



2.12 Television Channel Interferen ce Objectives 



The amount of noise and single-frequency interference that can be 

 tolerated in a commercial grade television channel again depends on 

 judgements involving the subjective reactions of observers and the cost 

 of providing a given grade of service. The broad objectives are based 

 on subjective measurements which have been reported on by Messrs. 

 Mertz and Baldwin.^' ^' ^ From this work it has been determined that 

 95 per cent of the observers consider a signal-to-noise ratio of 40 db 

 (composite signal to rms noise) tolerable, providing the noise has a 

 frequency characteristic that rises about 11 db across the video band. 

 Likewise the tolerable single frequency interference can be set at —70 

 db (peak sine wave below composite signal) if the interference falls 

 below about one megacycle. The requirement becomes more lenient for 

 interferences falling in the upper part of the band. 



Again, for a refined system design, more detailed account must again 

 be taken of the requirements on short duration interferences, the prob- 

 ability of interference occurring, and the exact frequency in the tele- 

 vision spectrum that an interference occurs. 



In the L3 signal-to-noise design the broad television channel objec- 

 tives outlined above were used except when a specific complex inter- 

 ference was found to be limiting. For complex interferences, three ad- 

 ditional types of requirement data were used; (1) tests were made to 

 determine visual thresholds relative to steady tones of short bursts 

 of energy such as occur in the television channel due to switchhook 

 •*bang-up" and multifrequency key pulsing signals in the message chan- 

 nels. Fig. 3 shows the relation between the steady state and transient 

 requirement: (2) advantage was taken from the fact that interferences 

 falling between the 15.75-kc line scan multiples of the television sig- 

 nal would be less interfering than unwanted energy falling directly at 

 the line scan multiples; and (3) a judgement was used that the toler- 

 abihty of an interference depends on its probability of occurrence. The 

 judgement was not made on a quantitative basis but when an inter- 

 ference was found to exceed its requirement by a few db two or three 

 times a day it was ignored in the signal-to-noise design. 



2.13 Frequency Allocations 



The final frequency allocations shown in Fig. 4 are a result of the 

 signal-to-noise design. The principal features were determined on rather 



