ANALYTICAL PROCEDURES 



279 



Figure 7. Oscillograph records of reverberation from three successive pings. 



laboratory to be analyzed and averaged. These rec- 

 ords are divided into sets, each consisting of records 

 taken within a short space of time under similar 

 conditions. The reverberation measurements making 

 up a set are then averaged, and the resultant averages 

 are supposed to represent the expected reverberation 

 under the known external conditions for the set. 

 Obviously the averages cannot be computed for every 

 time instant after midsignal. Times are chosen which 

 are spaced closely enough so that the major system- 

 atic changes in reverberation level will be evident." 

 At UCDWR, two methods of averaging have been 

 used : "point method" and "band method." 



The point method of averaging is to select a set of 

 points, such as 1, 2, and 3 in Figure 7; measure the 

 amplitude at these points; repeat for all records in the 

 set (usually from 5 to 10) ; make proper allowance for 

 gain changes and projector-receiver calibration; con- 



» In selecting and manipulating the data, places on the 

 records where obviously extraneous noise showed up have 

 customarily been rejected. Examples of extraneous noises are 

 pings from destroyers, echoes from porpoises, and bursts of 

 ship noises. 



vert the amplitudes to decibels above the chosen ref- 

 erence level; and finally, plot the resulting average 

 reverberation levels as a function of time or range. 

 Outstanding features, such as reverberation from the 

 bottom or from a suspected deep scattering layer, can 

 be emphasized by choosing many points in their 

 vicinity on the records; and uneventful portions of 

 the record can be passed over with but one or two 

 points to set the general level. Usually the points 

 chosen were spaced so as to give equal intervals on 

 logarithmic coordinate paper. 



The alternative method, the band method, was 

 introduced because of the considerable difficulty in- 

 volved in computing an accurate average with the 

 point method. On some records, the amplitudes are 

 changing very rapidly close to the predetermined 

 point where the amplitude is to be read; and to 

 measure these amplitudes accurately it is necessary 

 to look at the records very closely with appropriate 

 viewing devices. This procedure is both time-con- 

 suming and hard on the eyes, especially when the 

 amplitude at the predetermined time is small. 



In the band method, a set of points is chosen as in 



