Roof terrace noise: assessing people rather than plant

Roof terrace noise planning objections are decided on a different basis from plant noise, because the standard councils reach for first does not apply. BS 4142:2014+A1:2019 assesses sound of an industrial or commercial nature and expressly excludes noise made by people. A communal roof terrace is people. So there is no rating level to calculate, no 5 dB correction to argue about, and no single number an officer can hold you to.

What fills the gap is a mixture of prediction, baseline measurement, design mitigation and conditions. Get the design right and the conditions follow easily. Rely on conditions alone and you will be back at the council within a year of occupation.

Why people noise sits outside BS 4142

The exclusion is deliberate. Speech and raised voices are intermittent, highly variable in level, and carry information, which makes them more noticeable at a given decibel level than steady plant noise. A rating level built on a measured specific sound level over a reference period does not describe that well.

The consequence for an assessment is that the acoustician has to construct a method and justify it, rather than apply one. Councils accept that. What they do not accept is a report that measures the background, says the terrace is unlikely to cause harm, and stops.

Mechanical plant serving the building is a separate exercise and does fall under BS 4142, which is why most schemes need both parts in the same Noise Impact Assessment. Keep the two assessments visibly separate in the report so the officer and the environmental health officer can each find their part.

Establishing the baseline and the character of the area

Baseline measurement does more work here than in a plant assessment, because the question is whether terrace noise will be audible above what is already there and whether it will sound out of place.

Measure at the facade of the nearest noise-sensitive receptor, at the relevant height, over a period that captures evening and late evening conditions on both a weekday and a weekend. A single 24-hour unattended measurement in the wrong week tells you very little, and the duration and position questions are covered in unattended noise monitoring: duration, position, validity.

Character matters as much as level. A terrace above a busy high street with a measured evening ambient in the mid fifties is a different proposition from the same terrace facing a quiet mews, even where the arithmetic looks similar. Record what the dominant sources are, not only how loud they are, and say whether the area already has evening activity of the kind the terrace would add to.

BS 8233:2014 gives guideline internal levels for dwellings and a guideline range for external amenity areas, and those figures are the usual reference points for the receptor. Use them as context rather than as a pass or fail test, because they were not written for this source.

Predicting the terrace as a source

The common approach treats the terrace as an area source. You assume a sound power level per person based on published data for speech at normal and raised vocal effort, assume an occupancy density and a proportion of people speaking at once, distribute that over the terrace area, then propagate to the receptor using a standard outdoor propagation method with screening and reflection accounted for.

Every one of those assumptions is arguable, so state each one explicitly and run a sensitivity case. An officer who can see your occupancy assumption can disagree with it. An officer who cannot see it will assume you picked the convenient one.

Three assumptions deserve particular care.

1. Occupancy. Use the physical capacity of the terrace at a realistic density, not the number the management plan says will be allowed up there.
2. Vocal effort. Raised voices are the design case for an evening terrace, not normal conversation.
3. Simultaneity. The proportion of occupants speaking at any moment is low, but the peak condition matters more than the average for a receptor trying to sleep.

Amplified music should be excluded by design rather than modelled. If the scheme needs music on the terrace, that is a licensing question as well, and the interaction is covered in planning control and licensing control of venue noise.

Parapets, screens and absorptive surfaces

Mitigation on a terrace works in the order you would expect, and the first move costs the least.

A solid parapet of adequate mass, with no gaps at the base and no open balustrade, screens a seated occupant effectively. Glass works acoustically provided it is laminated, sealed and tall enough. The common failure is a 1.1 metre balustrade designed for falls protection only, which does nothing for a standing adult's mouth height at 1.5 metres or more.

Setting the usable area back from the parapet edge increases the screening angle for little cost. Hard reflective soffits above seating send sound back out over the parapet, so an absorptive soffit treatment under any canopy is worth specifying. Avoid narrow courtyards with reflective facades on three sides, where sound builds rather than disperses.

Planting screens views and does almost nothing for sound. Say so in the report rather than letting a landscape drawing imply otherwise.

Hours, management plans and enforceable conditions

Conditions have to meet the six tests in the Planning Practice Guidance on the use of planning conditions, and the one most often failed on terrace schemes is enforceability. A condition that no noise shall be audible at the boundary is unenforceable and should be resisted, politely, at the draft condition stage.

Conditions that do work on this kind of scheme include a restriction on hours of use expressed as clock times, a prohibition on amplified sound, a requirement that access doors to the terrace be self-closing and that the terrace not be used for deliveries or servicing, approval of a management plan before first use, and a limit on the number of people permitted on the terrace at any one time where that number is physically controllable.

The management plan is where the detail belongs. Name who controls access, how, and at what hours. A fob-controlled door that locks at 22:00 is worth more than a page of behavioural expectations. Signage and a residents' code of conduct are useful additions and poor substitutes.

An illustrative case: a 48-flat building with a shared terrace on the fifth floor, 20 metres from the rear windows of a terrace of houses. The acoustic report modelled 30 standing occupants at raised vocal effort. Raising the parapet to 1.8 metres in laminated glass and setting the seating back 2 metres brought the predicted level close to the measured evening ambient, and the council conditioned a 22:00 curfew with the door fob system rather than refusing. Figures are illustrative.

Where a terrace should not be provided

Some locations cannot be mitigated, and it is better to say so at design stage than to defend them at committee.

Directly above or immediately opposite bedroom windows in an adjoining building. In a narrow enclosed courtyard with hard surfaces on three or more sides. On a boundary where the terrace edge is the party line. Where the only access is through a stair core that would need to stay open for ventilation. And above a single-aspect flat in the same building whose only openable window faces the terrace, which is the version that generates complaints from inside the development rather than outside it.

Internal receptors within your own scheme are the ones applicants forget. The council will not, and nor will the management company.

Evidence from comparable terraces

The most persuasive material in a terrace assessment is a measurement at an occupied terrace of similar size, occupancy and construction, taken at an equivalent distance. It converts an argument about assumptions into an argument about comparability, which is easier to win.

Where you have access to such a site, measure it, describe the occupancy at the time, and present it alongside the prediction. Where you do not, say which published source levels you used and why. A road traffic noise baseline at the same receptor is often already available from earlier work on the site, and the approach to that is set out in road traffic noise assessment for a new housing site.

A Noise Impact Assessment of this kind should be prepared or reviewed by a suitably qualified acoustician, usually a member of the Institute of Acoustics, and professional review is strongly advised.

Browse all 66 PlanScope report types to see which environmental documents a scheme with communal amenity space is likely to need.

PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.