430 



BELL SYSTEM TECHNICAL JOURNAL 



system, it is merely necessary to displace the experimental curves to 

 the right by 11 decibels. 



Figure 16 shows such a comparison between the feedback system ad- 

 justed to give 25 decibels of feedback, and an ideal amplitude modula- 



60 



5^° 



z 



30 



20 



10 20 30 40 50 



AMP. MOD. SYSTEM R.M S. SIG.-NOISE RATIO IN DB 



Fig. 16 — Theoretical comparison of signal-to-noise ratios obtained at 25-decibel 

 feedback (curve B) with amplitude-modulated system (curve A) and ideal limiter 

 system with deviation ratio = 31.1 (curve C). Point X = threshold of noise im- 

 provement for limiter system. Point Y — point jvhere "crackling" first became 

 evident in the presence of ± 124.5-kilocycle modulation. 



tion system. There is also included a curve showing the theoretical 

 performance which would be approached by a frequency-modulation 

 system using amplitude limitation. Transmitted band width and 

 audio-frequency response equal to that used in the experimental feed- 



