ARCHITECTURAL ACOUSTICS 



117 



such an arrangement was of no use when the origin of the sound 

 was an orchestra, and it was finally abandoned. The effect 

 of such a reflector has been discussed in detail by Tallant (i i), 

 and by him adopted as the most suitable shape for the rear 

 wall of a large auditorium in the University of Michigan. 



The offending surfaces in the Ilhnois auditorium were known 

 from the diagnosis referred to, and very heavy materials were 

 hung in such positions that they prevented reflection at these 

 surfaces. In this way 

 the hall was made 

 suitable for speech, 

 although at the sacri- 

 fice of features of the 

 architectural design. 



Another way in 

 which diagnosis of the 

 echo conditions of an 

 auditorium can be 

 effected in advance 

 has been suggested 

 and carried out by 

 Sabine (12). A small 

 model is made in 

 accordance with the 

 proposed plans, and 

 actual sound waves 

 inside the model are 

 photographed. The 

 form of the waves 

 suggests necessary 

 alterations in design, 

 and this method has 

 been used with suc- 

 cess. This method 

 has confirmed the view 

 that a curved wall is 

 almost always a men- 

 ace to good acoustics. 



The full nature of the effect of such a wall is best illustrated by 

 Fig. 8, a diagram of Sabine's, showing the variation of intensity 

 at head level in a room with a barrel-shaped ceiling, the radius 

 of curvature of which was on the floor, the source of sound 

 being in the middle of the room. Such an interference pattern, 

 comprising regions of intensification and comparative silence, 

 always occurs with a curved surface, a point verified by recent 

 observations in St. Paul's Cathedral (13). A possible way of 

 regarding these regions of intensification is as a series of main 



Fig. 8. — Distribution of intensity at the head level in a 

 room with a barrel-shaped ceiling, with centre of 

 curvature on the floor level. 



