428 



BELL SYSTEM TECHNICAL JOURNAL 



Kingsbury an amount corresponding to the differences shown by the 

 curve of Fig. 20, then the resulting values should be directly com- 

 parable to our data on the basis of decibels above threshold. Com- 

 parisons of his data on this basis with those reported in this paper are 

 shown in Fig. 21. The solid contour lines are drawn through points 

 taken from Table I and the dotted contour lines taken from Kings- 

 bury's data. It will be seen that the two sets of data are in good 

 agreement between 100 and 2000 cycles but diverge somewhat above 

 and below these points. The discrepancies are slightly greater than 

 would be expected from experimental errors, but might be explained 



500 1000 

 FREQUENCY IN CYCLES PER SECOND 



Fig. 21 — Loudness levels of pure tones — A comparison with Kingsbury's data. 



by the presence of a slight amount of noise during threshold de- 

 terminations. 



Appendix C. Optical Tone Generator of Complex Wave Forms 



For the loudness tests in which the reference tone was compared 

 with a complex tone having components of specified loudness levels 

 and frequencies, the tones were listened to by means of head receivers 

 as before; the circuit shown in Fig. 15 remaining the same excepting 

 for the vacuum tube oscillator marked "x Frequency." This was 

 replaced by a complex tone generator devised by E. C. Wente of the 

 Bell Telephone Laboratories. The generator is shown schematically 

 in Fig. 22. 



