Section 2 Noise

2.1  Sound pressure level, L_p

Sound pressure levels (SPL) are measured in decibels (dB) using a logarithmic scale where the sound level Lp is given by:

where p is the measured root mean square sound pressure level in Pascals and p_o is the reference sound pressure level, 20 x 10^-6 Pa. Note: 1 Pascal = 1 N/m². The relationship between dB and the pressure ratio p/p_o is given in Table 2.2.1 Conversion of dB and pressure ratio .

A subjective assessment of changes in sound pressure levels is given in Table 2.2.2 Subjective changes in pressure levels . and typical sound pressure levels in dBA (see para. Ch 2, 2 Noise 2.3 ) are given in Table 2.2.3 Typical sound pressure levels .

2.2  Frequencies of interest

The normal frequency range of hearing of young adults is approximately 20 to 18,000 Hz. The frequency range of human speech is principally between 350 and 3,500 Hz.

2.3  A-weighted sound pressure level, L_pA

Sound pressure level measurements relating to effect of noise on humans should be made in decibels using an A-weighting filter, dBA. The A-weighted frequency filter is used to reproduce the frequency response of the human ear. The A-weighting corrections are given in Table 2.2.4 Octave band frequencies and A-weighting .

2.4  Equivalent continuous sound level, L_Aeq,T

This is the notional A-weighted, continuous, steady sound pressure level that, within a specified time interval T, has the same mean square sound pressure level as a sound whose level varies with time. It is defined by:

For a continuous, unvarying noise, L_Aeq is numerically the same as L_pA. It is therefore possible to use a conventional sound level meter to determine L_Aeq if the noise levels for the whole period vary by less than 5 dB with a slow meter response.

2.5  Measurement technique

Measurements can be made using L_Aeq (or L_eq as necessary) if the sound level meter has the capability. This is often easier and less subjective than using L_pA or L_p. L_Aeq should be used whenever the L_pA fluctuates by more than 5 dB, or the sound is cyclic, irregular, or intermittent.

Readings should made to the nearest decibel. The microphone should be at head height, 1.2-1.6 metres above deck, and pointed towards the dominant noise source if any. Measurements should normally be taken in the middle of spaces and no closer than 1 metre from bulkheads and major reflecting sources.

If measurements are made using L_pA or L_p, the meter should be set to “slow” response except where noise levels approach the overriding limits. A measuring time of at least 5 seconds should be allowed. If the level fluctuates by no more than 5 dB, an average of the maximum and minimum excursions can be made by eye.

2.6  Octave band filters

If the maximum noise level for a space is exceeded or if subjectively annoying tones are present, then the unweighted noise level in each of the octave bands should be determined. The levels in each band are used to determine the Noise Rating number, Ch 3, 4 Noise .

The octave band frequencies cover the normal range of human hearing frequencies and are listed in Table 2.2.4 Octave band frequencies and A-weighting . The bands have a ratio of upper to lower frequencies of 2 and are centred on preferred frequencies given in ISO 266. Octave bands are comparatively coarse and use is sometimes made of third-octave bands.

2.7  Equipment standards

Sound level meters can be either precision or industrial grade. Meters should conform to international standards, for example IEC 651 type or better and IEC 804 type 2 or better for integrating-averaging meters. Precision meters have an accuracy of about ±1 dB; industrial grade meters have an accuracy of about ±3 dB. A factor of 3 dB should be added to industrial grade meter readings to cater for the reduced accuracy. The use of a precision grade meter is, therefore, recommended where the noise levels are likely to be close to the recommended levels or in cases of dispute.

Octave filter sets should conform to IEC 225 or equivalent.

Microphones should be of the random incidence type and should conform to IEC 179 and IEC 651 (Types 1 and 2) or equivalent.

2.8  Calibration

A suitable calibrator, approved by the manufacturer of the particular sound level meter, should be used. Calibrators for precision meters should be accurate to ± 0.3 dB and for industrial meters to ± 0.5 dB.

The sound level meter, octave filter set and calibrator should be returned to the manufacturer or other organisations which provide a calibration traceable to national standards at intervals not exceeding two years.

2.9  Wind screens

A microphone windscreen should be used when taking readings on bridge wings, on open deck, or below decks where there is any substantial air movement.

2.10  Recordings

A tape recorder having a linear response in the frequency range 20 to 20,000 Hz should be used in cases of dispute or investigation work.

2.11  Speech intelligibility

Speech intelligibility over public address and similar systems may be determined using techniques such as the Speech Interference Level (SIL) or Rapid Speech Transmission Index (RASTI). See also Ch 9 Audible alarms and public address systems.