One Dry Day

How loud is rain on a conservatory roof? The lab figures, then British rain

How loud is rain on a conservatory roof? In the standard lab test, heavy rain on 25 mm polycarbonate measures about 64.6 dB and on double glazed laminated roof glass about 52.0 dB, a gap of 12.6 dB. Those are sound intensity levels taken under a rig that drops 40 mm of water an hour, which is far harder than most rain that falls on an English garden.

So the useful answer comes in two steps: what the roofs do under the same test, and what happens when you swap the lab downpour for the drizzle and steady rain we actually get.

White panel lean-to roof across the back of a terraced house
A terrace lean-to with its solid roof.

What the lab measures, and how

The test behind these figures follows ISO 140-18. Artificial rain falls at a fixed rate on a sample of roof laid over a test room, and a probe reads the sound coming off the underside as an A-weighted intensity level, written LIA. A-weighting tilts the reading towards the pitches the human ear is most bothered by.

BRE ran the comparison for conservatory and lightweight roofs in 2006 (Information Paper IP 2/06). Polycarbonate came out loudest of the common conservatory roofs, and it also gave off the most energy above 1.6 kHz. That upper band is the hard, ticking patter people describe as a biscuit tin lid under a tap. ETFE pillows, the inflated plastic cushions used on atriums, were louder still at 68.7 dB.

How loud is rain on a roof? Measured and estimated levels in decibels
  • ETFE pillow roof, lab heavy rainBRE68.7 dB
  • 25 mm polycarbonate, lab heavy rainBRE64.6 dB
  • Polycarbonate, natural heavy rain (estimate)BRE, Liverpool61.9 dB
  • Roof window, laminated double glazing, lab heavy rainForum Acusticum52.5 dB
  • Double glazed roof glass, lab heavy rainBRE52.0 dB
  • Polycarbonate, natural moderate rain (estimate)BRE, Liverpool45.7 dB
  • Glass, natural moderate rain (estimate)BRE, Liverpool33.1 dB
  • 75 mm aluminium panel roofNo published testNo published figuren/a
  • Vacuum cleanerNIDCD70 dB
  • Normal conversationNIDCD50 to 65 dB
  • Fridge hummingNIDCD40 dB
  • WhisperNIDCD30 dB
  • WHO: single noises in a bedroom at night, not aboveWHO45 dB
  • WHO: bedroom average through the night, not aboveWHO30 dB

Roof figures are A-weighted sound intensity levels from laboratory rain at 40 mm an hour (BRE IP 2/06 and a 2026 Forum Acusticum study). Natural rain estimates apply the University of Liverpool offsets (natural heavy rain 2.7 dB lower, moderate 18.9 dB lower). Everyday sounds and WHO guidance are sound pressure levels, shown for scale, not as a like for like comparison. Nobody has published a rain test for 75 mm aluminium skinned conservatory panels, and we have not measured ours.

From a lab downpour to an English Tuesday

Forty millimetres an hour is cloudburst rain. Our Rain Ledger, built from the Met Office 2 km model data we pulled for 100 towns, counts only 26 hours a year above 4 mm in the average town, against 480 hours of moderate rain. Across two years and every town, only two hours reached the rig rate of 40 mm: 42 mm in Hastings on 18 July 2025, and 40.9 mm in Melksham on 1 September 2024.

Researchers at the University of Liverpool (Hopkins and Yu) measured how real rain compares with the rig. Natural heavy rain sat about 2.7 dB under the lab level and natural moderate rain about 18.9 dB under it. Apply those corrections and the picture changes a good deal, though the order of the roofs does not.

Lab intensity level, adjusted for natural rain
  • Polycarbonate, natural heavy rain61.9 dB
  • Glass, natural heavy rain49.3 dB
  • Polycarbonate, natural moderate rain45.7 dB
  • Glass, natural moderate rain33.1 dB

BRE IP 2/06 lab figures with the Hopkins and Yu corrections. Lab measure, not the level in your room.

What a 10 dB gap feels like

Decibels run on a log scale, so they do not add up like centimetres. A rise of about 10 dB is commonly heard as roughly double the loudness. The 12.6 dB between polycarbonate and glass is therefore more than a small tweak: under the same shower, the plastic roof sounds something like twice as loud, and sharper in tone.

For scale, the NIDCD chart puts a whisper at about 30 dB, a fridge at 40, ordinary conversation at 50 to 65 and a vacuum cleaner at 70. The World Health Organization suggests bedrooms stay near 30 dB through the night, with single events under 45 dB. These are room levels and our roof figures are lab intensity levels, so treat the comparison as a sense of size, never a like for like reading.

12.6 dB
between polycarbonate and glass in the same lab rain
52.5 dB
a glazed roof window, measured in 2026
480
hours of moderate rain a year, average town

Roof windows and part glass roofs

A 2026 study by Nusser and colleagues measured a glazed roof window at 52.5 dB, close to the BRE glass result. That matters if you are weighing up a panel roof with a roof light set into it: the light is glass, so the rain on that one pane sounds like rain on glass.

Part glass designs follow the same logic, each section behaving like its own material. Our page on hybrid roofs covers how the split works.

Where 75 mm insulated panels sit

Here is the plain truth. No one has published a rain noise test for 75 mm aluminium skinned foam conservatory panels, and we have not measured ours. We will not put a decibel figure on our own roof, or claim it beats another roof by a number.

What we can describe is the build. There are no hollow flutes to ring like an organ pipe. The foam core is bonded to a metal skin on top and another underneath, and you look up at a solid white ceiling instead of a translucent sheet. Whether that adds up to a quieter room in your house is something to judge by ear, ideally in a finished conservatory on a wet day.

Putting the numbers to use

If the roof over you is polycarbonate and rain is what drives you out of the room, the lab gap to glass is real and well measured. Look up how many nights of moderate rain your town gets in the Noise Nights league table, then read why polycarbonate drums the way it does before choosing between a cheap fix and a new roof.

Panel roof lean-to on a bungalow beside a patio garden
A bungalow lean-to with a white panel roof.

Sources

  1. Rain noise from glazed and lightweight roofing, BRE Information Paper IP 2/06 (Hopkins, 2006)
  2. Hopkins and Yu, rain noise in natural rainfall, Acoustics 2020 (University of Liverpool)
  3. Nusser and others, Rain sound insulation of timber roofs, Forum Acusticum
  4. Marshall Day Acoustics letter on rain noise of roof systems (Ballagh, 2015)
  5. World Health Organization, Guidelines for Community Noise (1999)
  6. World Health Organization, Night Noise Guidelines for Europe (2009)
  7. Everyday sound levels, NIDCD chart reproduced by the University of Colorado
  8. Met Office National Meteorological Library, Fact sheet 3: Water in the atmosphere

Questions we get asked

Is a decibel figure in a lab the same as the noise in my room?

No. The lab reads sound intensity coming off the underside of a sample, while your room level depends on its size, furnishings and the walls. The lab figures are best used to rank roofs against each other.

Why is drizzle so much quieter than the test?

Small drops falling slowly carry far less energy than the big, fast drops in a 40 mm an hour rig. Hopkins and Yu found natural moderate rain about 18.9 dB below the lab heavy rain level.

Do you have a rain noise figure for your panel roof?

No. Nobody has published one for 75 mm foam conservatory panels and we have not tested ours, so we do not quote one.

Does wind make rain louder on a roof?

Wind can drive drops in harder and at an angle, and rattle anything loose, but the published lab figures are for still, vertical rain, so we have no number to give for windy rain.

Where to next

Find out what your roof would cost

Seven taps and a postcode give you a guide range straight away. A free survey turns it into a written price, and the forecast chooses the day.