152 



BELL SYSTEM TECHNICAL JOURNAL 



ference. Note that a curve of almost any desired trend may be 

 obtained by locating the condenser transmitter at the proper position. 



On Fig. 5 the data of Fig. 4 are plotted on polar coordinate paper to 

 show more clearly the approximate manner in which the sound field 

 varies. On these curves the magnitude of the response is expressed 

 relative to that on the axis and the approximate distribution at each of 

 four frequencies is shown. At the larger angles if a sufficiently large 

 number of measurements are made an irregular interference pattern is 

 obtained like that shown on Fig. 6 for 1,000 cycles. 



The second loud speaker measured consisted of a 3|" piston dia- 

 phragm (inertia control) mounted in one side of a cubical box approxi- 



20 



(rt 



50 



500 1000 



FREQUENCY 



Fig. 9 — Response-frequency characteristic of 3}4" piston diapliragm loud speaker. 

 Measured in highly absorbing room 12 feet from and on a line perpendicular to the 

 center of the diaphragm. 



mately 12" on a side and filled with wool. This loud speaker was thus 

 of a radically different type from the first and was chosen because the 

 size and nature of the diaphragm was such that any reaction of the 

 medium on its vibration amplitude would be unlikely. A direct 

 comparison of its performance under one medium condition with 

 that under another would therefore be justifiable. The response- 

 frequency characteristic of this loud speaker as measured outdoors 

 with the condenser transmitter at a distance of 12' on the diaphragm 

 center perpendicular is shown on Fig. 7. The irregularity in this curve 

 at 600 cycles has been shown by other tests to be due to poor dynamical 

 characteristics of the loud speaker itself. 



On Fig. 8 are polar coordinate curves for the piston diaphragm loud 



