of calibration, using two easily inserted calibration signals, 

 allows the entire electrical system to be calibrated very 

 simply and as often as desired. The calibrate signals are 

 recorded on tape along with the ambient -noise signal, and 

 so also serve to calibrate any subsequent analysis equipment 

 as well. 



In practice the calibration signals are recorded just 

 prior to the ambient-noise measurement. The level of the 

 tone is adjusted precisely to the same known level each 

 time by monitoring the current in the calibration circuit. 

 Since the ambient-noise signal is present when the calibrate 

 signals are applied, it is necessary to apply the calibrate 

 signals at a much higher level than the ambient -noise 

 voltage level. 



To use this method of calibration it is only necessary 

 to know the sensitivity of the hydrophone as a function of 

 frequency, the insert level of the tone (as monitored) and the 

 relative spectrum shape of the random noise as applied at 

 the 10-ohm resistor. The sensitivity, S(f), of the hydro- 

 phone is measured periodically at an acoustic calibration 

 facility. The spectrum shape of the random noise is mea- 

 sured periodically using a one-third octave band analysis, 

 and the relative spectrum level, N'(f), is obtained by sub- 

 tracting 10 log 10 (bandwidth) from the band level. The spec- 

 trum level of the random noise is determined by the differ- 

 ence, d, between the known tone level, T, and the one-third 

 octave band level centered at the tone frequency. 

 N(IOOO) = T - D - 10 log (bandwidth) 



where n(1000) is the absolute spectrum level of the random 

 noise at 1000 Hz. The absolute spectrum level at other 

 frequencies is determined by 



Nif) = /V(IOOO) - [N'(IOOO) - N'(f)l 

 Conversion to an equivalent SPSL is made using the appropri- 

 ate hydrophone sensitivity curve (fig. 4 or 5) as measured at 

 an acoustic calibration facility. 



Equivalent SPSL = Nif) - Sif) + 74 

 where the 74 is a conversion factor to a reference pressure 

 of 0. 0002 dyn-cm -2 . Figure 10 shows the equivalent SPSL 

 of the tone and random noise of the two calibrate boxes as 

 normally used. If the broadband level of the random noise 

 should change, this will show up as a change in the difference, 



15 



