144 BELL SYSTEM TECHNICAL JOURNAL 



any distance large compared to a wave-length since this latter average 

 approaches the half wave-length mean. The mean square pressure is 

 therefore just as suitable as the maximum value as a measure of the 

 energy density and further, it lends itself more readily to the determi- 

 nation of the energy density in the case where there are standing waves 

 in several directions. For this latter case the energy density within a 

 specified region is proportional to the mean square pressure in all 

 directions or at all points within the volume of a sphere having a 

 diameter large compared to the wave-length. The response of the 

 loud speaker at any frequency can accordingly be measured in a room 

 with reflections by averaging the squares of the pressures throughout a 

 suitable volume. 



The above method of response measuring indoors however, is not 

 entirely independent of the measuring room. If the reflected energy 

 is large the response measurements will be affected by the variation in 

 the absorption power of the room with frequency so that a large room 

 with absorbing material having as uniform and large absorption 

 characteristics as possible over the measuring frequency range is still 

 desirable. Some sound absorbing materials have uniform absorption 

 characteristics but when this is the case the absorbing power is apt to 

 be very low. The use of such materials results in extremely large 

 pressure variations within the room so that a measuring device having 

 a sufficient amplitude range to average the squares of the pressures is 

 difficult to obtain. For this reason and because of the fact that the 

 region through which it is necessary to average the squares of the 

 pressures at low frequencies becomes prohibitive, a large room is most 

 desirable so that the difference between the direct and reflected energies 

 at the transmitter position will be as large as possible. Indoor 

 measurements under these conditions approach infinite medium 

 measurements. 



The use of a large room also results in less reaction of the room 

 inclosure on the loud speaker itself. While under most conditions 

 such reactions have little effect on the acoustic output power of the 

 loud speaker, in small measuring rooms at v'ery low frequencies where 

 the absorption is low and the radiator of the loud speaker (perhaps 

 designed for a large auditorium) is large, the phase and magnitude of 

 the reflected energy at the radiating surface of the loud speaker may be 

 such as to cause large variations in the acoustic impedance of the 

 medium on the area adjacent to this surface. This variation in the 

 acoustic impedance of the loud speaker load will cause variations in the 

 acoustic power density at the transmitter position. Response measure- 

 ments on loud speakers of large dimensions and particularly measure- 



