260 MEASUREMENTS 



cabinet. The second class are complete units and require no additional 

 structures. Examples of this type are horn loud speakers or loud speakers 

 already mounted in housings. 



The loud speaker unit should be mounted one foot off center in a direc- 

 tion parallel to one side and six inches off center in a right angle direc- 

 tion in a twelve foot square baffle of sufficient thickness so that no radiation 

 results from vibration of the baffle. The microphone should be located 

 on the axis of the radiator at a distance of five feet from the surface of the 

 baffle when the maximum transverse dimension of the radiator is not 

 greater than two and one half feet. For larger radiators, the distance 

 should be the smallest integral multiple by five feet, which is greater than 

 twice the maximum transverse dimension of the radiator and should be 

 specified with the test results. The response frequency characteristic of 

 the loud speaker can then be obtained by one of the methods described in 

 the section on apparatus and plotted in decibels, equation 11.10 (on an 

 ordinate scale of 30 db or less per cycle of logarithmic scale of frequency). 

 This gives the response frequency characteristic of the loud speaker on the 

 axis. 



Complete speakers should be tested in the same manner as speaker units 

 alone, but without the use of additional baffles. 



Routine tests and development work upon acoustical instruments are 

 usually made indoors reserving the outdoor measurements for a standard. 

 The room should be made as large as possible in order to obtain a max- 

 imum ratio of direct to reflected sound. Standing wave phenomena will 

 be minimized if the dimensions of the room are made 3:4:5. Several 

 layers ^^ of absorbing material should be used with a separation of from 

 one to six inches between layers. Reflections will be further minimized 

 if strips of absorbing material twelve inches in width are placed normal 

 to the walls and spaced twelve inches apart. 



When only a small deadened room is available a " close up " curve may 

 be the only one possible. Such a curve may be useful in determining 

 system resonance and the general smoothness of the output. 



The rotating microphone "^ has been found to be very useful for reducing 

 reflection errors. The microphone is revolved in a circle about five feet 

 in diameter. The plane of the circle is inclined at an angle of 30° towards 

 the horizontal. The microphone is arranged so that it always iS directed 

 towards the source of sound. Any of the systems described in Sec. W.ZAJ. 

 may be used for recording the response frequency characteristic. 



23 Bedell, E. H., Jour. Acous. Soc. Amer., Vol. 8, No. 2, p. 118, 1937. 



24 Bostwick, L. G., Bell Syst. Tech. Jour., Vol. 8, No. 1, p. 135, 1929. 



