Breaker and O'Neil 



The projector was a USRD Type G31 transducer produced by 

 the Underwater Sound Reference Division of the Naval Research 

 Laboratory. It was mounted in an Ocean Research Equipment, Inc. 

 towed body. The transmitted frequencies were 1.3, 3.0, 5.3, and 

 9.0 kHz (determined by the resonant peaks in transducer output). 

 USRD Calibrated the transducer as mounted in the body and obtained 

 the following sound pressure levels at 1 meter from the transducer 

 face, measured in dB//l^ bar: 



Frequency Level 



Transmitting beamwidths were found to be very broad, even at high 

 frequencies. 



A NUS Model LM-2 Deep Sea Hydrophone was used on the 

 receiving end, suspended at a depth of approximately 3000 feet. The 

 signal was passed through a filter, set to a one octave band centered 

 at the transmitter frequency, amplified, and recorded on magnetic 

 tape at 7-1/2 i.p.s. 



It can be seen from Eq. (6) and Figure 1 that the effective 

 width of P(f) and hence the amount of data which must be collected to 

 obtain satisfactory resolution of the shape of the spectrum is 

 proportional to Vj^. It was decided that a v^ = 1.3 m/sec, corre- 

 sponding to a ship speed of 5 knots, represented a satisfactory 

 compromise between ship and towed body capabilities on the one hand 

 and data-taking time on the other. 



Ten different "events" or subexperiments were undertaken. 

 Table 1 displays the important fea tures of these events arranged in 

 order of increasing values of^/XTT and decreasing acoustic frequency. 

 The durations were chosen according to the methods outlined in 

 Ref. 6 to give equally good resolution and stability for each event's 

 sample estimate of the spectrum P(f). 



After consideration of the propagation-velocity data 

 available for the TOTO area, it was decided that a 2 -second keying 

 interval would provide adequate insurance against overlap of bottom 

 reflections with the succeeding direct pulse. This, of course, 

 fixed the effective sampling rate and thus the Nyquist (or folding) 

 frequency for the spectral estimates. In the worst case (Event 10), 



^^ ^ J^|jnAec_ ^ Q26 Hz 

 V27r-20m 



374 



