392 



BELL SYSTEM TECHNICAL JOURNAL 



According to Sabine's well known formula the reverberation time is 

 inversely proportional to the number of absorption units in the room so 

 that, if we assume this, we may immediately infer the shape of the 

 curve which represents the number of absorption units necessary at 

 any frequency, referred to the amount required at 1000 cycles, to 

 obtain our required condition. These values are plotted in Fig. 3. If 

 it should happen that the greater part of the sound absorption in a 

 room is caused by one particular kind of surface, then the curve in 

 Fig. 3 is the shape of the absorption curve that this material should 

 have. 



A pertinent observation on which every one seems to agree is that if 



2.4 



2 34568 2 34568 



10 100 1,000 10,000 



FREQUENCY IN CYCLES PER SECOND 



Fig. 2. — \ 'allies of Af vs. frequency. 



an auditorium has an unusually long reverberation time and conse- 

 quently is of little use, when empty, it attains excellent acoustic 

 conditions when filled with a large audience. In these cases a very 

 large part of the absorption is caused by the audience. The absorp- 

 tion of an average audience has been measured by W. C. Sabine ^ and 

 his results are also plotted in Fig. 3. The close agreement between this 

 curve and the one we have obtained from our hypothesis gives con- 

 siderable confidence in our general viewpoint. 



II. Reverberation Time vs. Volume 



It is generally accepted that the best acoustical conditions in a room 

 are obtained when the reverberation time is adjusted to a definite value 

 known as the optimum reverberation time. Observations reported in 

 literature agree that the value of the optimum reverberation time in- 



^ "Collected Papers on Acoustics," page 86. 



