Noise Level Calculation
Many types of noise can be eliminated during data collection or postprocessing. Noise removal during data collection is preferable. Noise that cannot be removed will increase S_{v} measurements and bias in situ TS measurements, leading to errors in fish density or biomass estimates.
Some instruments output noise levels at 1 m depth and this should be recorded. We recommend that authors report the noise levels or the SNR at the deepest depth included in the analyses. Remember that noise levels at 1 m are different if expressed as TSu or S_{v} data. This is because S_{v }values include a term for the equivalent beam angle and sampling volume whereas uncompensated target strength (TSu) values do not (compare Equations 3 and 6). Therefore, the difference between TSu and S_{v} data is depth dependent. Note that TS refer to the target strength of the fish and TSu refer to the measure of echo level with a 40 log R TVG function, also sometimes called nonadjusted TS. This distinction is important and not clear in some existing software. From combining equation 6 and 12, we get:


[24] 
A noise level measured in the S_{v} domain as 120 dB at 1 m depth (with a 120 kHz, 7.8º transducer and therefore a Ψ of –20.4 dB, a sound speed of 1450 m•s^{1}, and a pulse duration of 0.4 ms) is 146 dB in the TS domain at 1m (20 log (1) = 0)