696 



THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1957 



40 



38 



36 



34 



32 



30 



28 



26 



24 



22 



20 



18 



16 

 14 



SIGNAL POWER IN DECIBELS BELOW FULL LOAD SINUSOID 

 5 10 15 20 25 30 35 40 45 50 



< 



CC 

 LU 



5 



o 



Q. 



tr 

 O 

 tr 

 tr 



< 



z 



// = 100 



200 



n = 7 



0=6 



B=oo 

 A=l/Y2 



— V 



SIGNAL TO MAXIMUM 

 '\ NOISE POWER RATIO 



Fig. 20 — Curves illustrating the comparison of signal to quantizing error 

 power ratios with the ratio of signal to background noise. The line representing 

 the signal to maximum noise ratios corresponds to the hypothetical case where 

 the maximum background noise is determined bj- the requirement that the signal 

 to noise ratio be 20 db for a signal 50 db below full sinusoidal modulation. 



sidering a value providing a signal to noise ratio of 20 db for the weakest 

 signals in our hypothetical system. A signal to maximum noise power 

 curve may then be dra^Mi as a function of signal power for this constant 

 value of noise power. Such a graph has been combined, in Fig. 20, with 

 curves such as those which have previously appeared in Figs. 15 to 19. 

 These curves have been terminated at their intersections with the line 

 representing the signal to maximum noise power ratio since we are 

 assuming that little benefit will be derived from a signal to quantizing 

 error power ratio in excess of the signal to ma.ximum noise power ratio. 

 From Fig. 20 it is apparent that the previous conclusions that six 

 and seven digits are worthy of consideration are unaffected by the stipu- 

 lation that the signal to cjuantizing error power ratio should not greatly 

 exceed the signal to maximum noise power ratio. Similarly, the con- 

 clusions based on Fig. 19 (for B = 100) remain unchanged since the 

 curves therein fall below the maximum noise curve of Fig. 20 for all 

 values of the abscissa. 



