What is the best fence or wall to block road noise?

What is the best fence or wall to block road noise?

The best barrier is a solid one with no gaps, tall enough to break the line of sight between you and the road. Material matters less than height and gaps. The US Federal Highway Administration (FHWA) handbook says to expect about 5 dB when the line of sight is only just blocked, and that a well-designed barrier can approach 10 dB. The government's Noise Action Plan for roads says up to 10 dB(A), and a paper by Hothersall for the Institute of Acoustics (2000) calls about 20 dB probably the most achievable in practice.

A garden boundary is shorter and lower than a motorway barrier, so in our view the lower end of that range is the realistic expectation. Almost all the independent measurements are for highway barriers, not garden fences.

How much do height and line of sight matter?

They matter most. A barrier works by making sound travel further over its top edge, so a fence that does not hide the road from your ear does very little, however solid it is. FHWA says each extra metre above that point adds about 1.5 dB.

Our guide to fence height explains why the right answer differs from house to house.

Why do gaps and open ends matter so much?

Sound takes the easiest route, so a hole can undo much of what the solid part achieves. Hong Kong's government barrier guidelines (2022) treat a gap as having no sound insulation. In their example, a leak covering 1.5% of the area costs about 2 dB of insulation when 10 dB is needed, but about 13 dB when 25 dB is needed.

Ends work the same way. A worked example in Hothersall's paper has a 3 m barrier cut to 100 m length losing about 7 dB, and a garden boundary is much shorter than that. Sound arriving around the ends can matter as much as sound over the top. Our guide on whether acoustic fencing works covers closing the ends.

Does the panel material matter?

Less than most sales material suggests, once the panel is solid and gap-free. Sound passing through should be well below sound bending over the top. The Hong Kong guidelines want the material's own sound loss at least 10 dB(A) above the reduction you want, Hothersall says transmitted energy should be at least 10 dB below the diffracted energy, and the UK highway standard LD 119 (2020) now asks for sound insulation 15 dB(A) above the barrier's design insertion loss. So the sources ask for panel insulation to exceed the target gain by about 10 to 15 dB, which for a 5 to 10 dB gain means roughly 15 to 25 dB. That step is our own arithmetic, not a figure from a source.

The most complete independent table we found is in the Hong Kong guidelines. It is non-UK government guidance for road barriers, and it describes the panel alone, not the benefit at a house. It gives about 21 dB for 25 mm timber, about 18 to 25 dB for thin steel or aluminium, depending on gauge, about 32 dB for 15 mm acrylic, and about 40 dB for 150 mm brick or 100 mm dense concrete.

The sources disagree on minimum weight. Hong Kong and ISO 9613-2 (quoted by Hothersall) treat about 10 kg per square metre as typically enough, FHWA uses 20 kg as a rule of thumb, and an author from a barrier company in Acoustics Bulletin (2023) calls 10 to 12 kg "inadequately low". Hong Kong's own 25 mm timber, at about 18 kg, gives 21 dB, below the roughly 33 dB it calls adequate anywhere. So we give no single minimum.

We do not rank timber against brick for a garden, because independent garden-scale data is thin. We have not repeated supplier dB claims, which describe the panel rather than the benefit at a house.

Does a timber fence lose performance as it ages?

A barrier-industry article reports that it can. A Transport Research Laboratory study (Morgan, 2010) looked at timber barriers on the strategic road network, but its abstract gives no dB figure. The number comes through an article by an author from a barrier company in Acoustics Bulletin (2023), which says single-leaf timber barriers typically lose up to about 7 dB of insulation within their first five years. These are highway-grade panels, and whether a garden fence ages the same way has not been shown.

Is an earth bund better than a wall?

The sources point in opposite directions, and neither gap is large. FHWA says a berm typically gives an extra 1 to 3 dB(A) over a wall of the same height. A University of British Columbia scale-model study (38 configurations) found walls beat berms by 2 dBA. HA 66/95 says an earth mound is less effective than a thin screen of the same height, and says only that a short screen on a mound "might be beneficial".

We do not rank a bund against a wall. The usual case for a bund is cost, as HA 66/95 rates earth mounds cheaper to build than timber screens. Bunds need width, though: with sides sloping one in two, a 2 m bund is at least 8 m across at the base (our arithmetic).

Can a wall facing my fence make the noise worse?

Possibly, but the figures we have are not for a garden. FHWA says practical measurements commonly show not more than 1 to 2 dB(A) from a reflective barrier on one side of a road reflecting sound to the far side. For facing parallel reflective barriers, FHWA's handbook allows for a loss of 2 to as much as 6 dB(A). A garden fence facing a wall is not covered by either set of figures. Sound-absorbing faces mostly help people on the other side: Watts and Godfrey (1999), quoted by Hothersall, found under 0.5 dB difference on the protected side of a 3 m barrier.

What will a fence not do?

It will not help what it cannot hide. An upstairs window that looks over the top gets little, and we found no UK figure for upper floors; see acoustic fencing versus secondary glazing for rooms. It is also weaker against low-frequency sound, such as heavy-vehicle rumble. A normal hedge does far less (up to about 3.6 dBA measured, typically about 1 to 3 dBA); see hedges and trees.

Limits of this evidence

Garden-scale evidence is thin. The Hong Kong table is non-UK guidance for road barriers. The UK ageing figure reaches us second-hand, through a barrier-industry article rather than the TRL study itself, and this page does not cover standards such as CRTN. Two figures come from a low-quality scan of Hothersall's paper, so treat them as approximate.

Every figure here is a published value, not a measurement or guarantee for your garden.

What is a sensible next step?

A modelled estimate for your own boundary is more useful than any general figure. The Fence Feasibility Check and barrier design pages show modelled options with indicative costs to discuss with a builder. They are estimates, not measurements or quotes.