How Acoustic Barriers Reduce Urban Noise Pollution
They work by breaking the line of sight, which is why upper floors get nothing at all. Here is the real 5 to 10 dB figure and where barriers fail.
An acoustic barrier works by breaking the line of sight between the noise source and the listener, forcing sound to bend over the top. If you can see the road, the barrier is not working for you. A well placed one delivers 5 to 10 dB, occasionally 15 when it is tall and close to the source, and every doubling of height after that buys only a few decibels more.
That single rule explains most of what barriers do and almost all of what people expect from them and do not get.
The mechanism
Sound travels in more or less straight paths at the frequencies that matter for traffic. Put a solid wall in the way and the direct path is blocked, so the only route to the listener is diffraction over the top edge.
Diffraction costs energy. How much depends on the path difference, meaning how much further the sound has to travel going over the barrier compared with going straight through it. A larger path difference means more attenuation.
Two things follow. Height matters, because it increases the path difference. And position matters more than most people expect, because a barrier close to the source or close to the receiver creates a bigger path difference than the same barrier halfway between them.
The mass of the barrier is a secondary consideration. It needs to be heavy enough that sound does not simply pass through, which in practice means around 20 kg per square meter, and beyond that extra mass does very little because diffraction over the top dominates.
Why frequency decides the result
Barriers work better at high frequencies than low ones, and traffic noise is unhelpfully distributed.
Short wavelengths behave like light and are shadowed effectively. Long wavelengths bend around obstacles readily. A 1 kHz wave is 13 inches long and is well controlled by a 4 meter wall. A 63 Hz wave is 18 feet long and treats the same wall as a minor inconvenience.
Tyre noise peaks around 800 Hz to 1 kHz, which barriers handle reasonably. Engine and exhaust noise sits well below that, and heavy vehicles produce substantial energy under 200 Hz.
The practical consequence is the complaint acoustic consultants hear constantly: the barrier removed the hiss and left the rumble. That is not a defect in the installation. It is the physics working as expected. Our explainer on how sound wave propagation works covers why wavelength governs everything.
Where barriers fail
Upper floors. The most serious limitation in a city. A barrier tall enough to shield a fourth floor apartment is rarely feasible, so the residents most exposed to traffic are the ones a barrier cannot help. Ground floor and first floor benefit, and above that the line of sight is clear.
Gaps. A barrier with an access opening, a drainage gap or a poorly sealed joint performs far below its rating. Sound leaks through openings the way light does, and a small gap in a long barrier can cost several decibels.
Reflection. A hard barrier on one side of a road reflects sound across to the other side, which can raise levels there by 1 to 3 dB. Absorptive facing on the source side addresses it, and parallel barriers on both sides of a road create repeated reflections that reduce the performance of both.
Ends. Sound diffracts around the ends as well as over the top, so a barrier needs to extend well past the receiver in both directions, typically four times the distance from the receiver to the barrier.
Distance. Barriers work best close to the source or close to the receiver. One in the middle of a wide verge is the least effective position.
What actually gets built
Concrete and masonry walls, durable and reflective unless faced with an absorptive layer.
Timber and composite panels, lighter and cheaper, needing enough mass per square meter to avoid transmission through the panel.
Earth bunds, which are the best performing option where land is available. They are massive, they absorb rather than reflect, they can be planted, and they avoid the maintenance and graffiti problems walls have.
Transparent acrylic panels, used where a solid wall would be unacceptable visually or would block light. They perform acoustically like any other barrier of similar mass and cost considerably more.
Absorptive faced barriers, with a mineral wool or porous concrete layer behind a perforated skin on the source side, which stops the barrier reflecting noise back across the road.
Vegetation, which is mostly a myth
A row of trees or a hedge is often proposed as a noise barrier and it is worth being honest about what it delivers.
Dense vegetation of significant depth, meaning tens of meters rather than a screen of trees, provides a few decibels of attenuation. A single row of trees provides almost none.
What vegetation reliably does is psychological. Screening the source visually reduces reported annoyance at the same measured level, and that effect is real and well documented. It is also why a planted bund outperforms a bare wall in resident satisfaction surveys even where the measured difference is small.
Presenting planting as noise mitigation without stating the measured contribution is where this crosses from useful to misleading.
The measures that outperform barriers
Barriers treat the path. Treating the source is almost always cheaper and more effective across a whole city.
Low noise road surfaces deliver 3 to 5 dB, and they apply to every receiver on that road simultaneously including upper floors, which no barrier does.
Speed reduction delivers 2 to 3 dB for a 10 km/h reduction, and it costs a sign.
Facade treatment at the receiver, meaning acoustic glazing and treated ventilation, protects upper floors that barriers cannot reach.
Our piece on the main challenges in urban acoustic design covers how these fit together, and urban noise mitigation in city planning covers the policy side.
Frequently asked questions
How much noise do acoustic barriers reduce?
Typically 5 to 10 dB, and up to about 15 dB for a tall barrier placed close to the source. A 10 dB reduction is perceived as roughly half as loud, so the effect is genuine. Performance depends entirely on breaking the line of sight, so a barrier you can see over is providing very little regardless of its specification.
Why do noise barriers not help upper floors?
Because attenuation comes from forcing sound to diffract over the top edge, and an apartment on an upper floor usually has a clear view of the road over any barrier that could realistically be built. Height gains also diminish quickly, with each doubling adding only a few decibels. Facade treatment at the building is the answer for those residents.
Do trees and hedges block traffic noise?
Barely. Dense vegetation tens of meters deep provides a few decibels, and a single row of trees provides almost nothing measurable. What planting reliably does is reduce reported annoyance by screening the source visually, which is a real and well documented effect. Presenting it as acoustic mitigation without stating the measured contribution is misleading.
What is the best position for a noise barrier?
Close to the source or close to the receiver, since both maximize the path difference that sound must travel to diffract over the top. A barrier halfway between the two is the least effective position. It also needs to extend well past the receiver at each end, typically four times the receiver to barrier distance, or sound diffracts around the ends.