The air coolers here from Noctua, Thermalright, Scythe, ID-COOLING and Cooler Master all publish maximum figures between 22.6 and 28.6 dB(A).
Noctua's NH-U12A publishes 22.6, Thermalright's Peerless Assassin 25.6, Cooler Master's Hyper 212 Halo 27 and ID-COOLING's Frozn A620 27.2. Those are different coolers with different fans and different fin stacks, and by the box they are almost the same machine. They do not sound alike in a case. A maximum speed figure just measures the least interesting state a cooler ever reaches, and at that speed one 120 mm or 140 mm fan sounds much like another.
be quiet! is the one maker here that prints whole cooler figures below full speed, at 50, 75 and 100% PWM. Scythe prints a per fan range whose bottom end is the fan's minimum speed.
The two Dark Rock Pro models swap places across those steps. The Pro 6 reads "7.5 / 20.2 / 32.4" and the Pro 5 "8.9 / 16.3 / 23.3": the newer cooler is quieter at half speed, the older one at 75 and 100%.
A 50% figure set beside another brand's maximum tells you nothing. Inside be quiet!'s range, though, the rows line up, and the Dark Rock Elite at 11.0 / 19.5 / 25.8 is louder than the cheaper Pro 5 at every point. Our be quiet! fan and cooler page puts every current be quiet! cooler on that same three step row.
Both Dark Rock Pro models carry a Quiet and a Performance switch, and the Pro 6's quiet mode keeps its fans "off below 40% PWM". Still, each cooler gets one set of figures, with no word on which switch position they describe.
DeepCool did print a figure for each switch position on its ASSASSIN IV, 22.6 dB(A) in Quiet Mode and 29.3 in Performance Mode. Both carry a less than or equal to sign, which makes them ceilings. The US Treasury put Beijing DeepCool Industries on its sanctions list in June 2024, and on November 8, 2026 no DeepCool-brand listing turned up on Amazon US, so neither the ASSASSIN IV nor the AK620 is a pick.
Can you compare CPU cooler dB ratings across brands?
Not reliably. Of the ten brands checked, only be quiet! names a measuring distance, not one cites ISO 7779 or any other standard, and none describes the test room.
Even be quiet!'s distance sits on a separate noise page, not the spec tables: "all measurements are made with a measuring distance of one meter". Noctua, Thermalright and the rest say nothing about distance at all.
Within one brand's range the figures are useful, because the test was presumably the same. Across brands, a 2 dB difference is noise in the data rather than a gap you would hear. Appliances have the same problem. A cooler maker that names no microphone distance gives you a figure that cannot be lined up against another brand's, and the explainer shows why.
Whether a figure covers one fan or the whole cooler varies too. be quiet! labels its rows "Overall noise level", Noctua gives one figure per cooler on its infosheets, and Thermalright's number sits in a fan block on a cooler that ships with two fans.
Among these sheets were a copied block, the same pump rated twice and coolers with no figure at all.
A copied block
Thermalright's Peerless Assassin 120 SE and Phantom Spirit 120 SE use different fans at different rated speeds, 1550 and 1500 RPM. Both pages publish "25.6 dBA", "66.17 CFM" and "1.53mm H2O". One block was copied from the other, and our Thermalright cooler page shows how often that 25.6 recurs across the range.
The same pump, rated twice
Cooler Master publishes "Pump Noise Level 12 dBA (Max)" for the MasterLiquid 360 ATMOS White and 25 dBA for the 240 Atmos, with the same pump dimensions, the same MTTF and the same power draw. The rest of the range is on our Cooler Master cooler page.
ARCTIC gives no acoustic figure for the Freezer 36 or the Liquid Freezer III Pro 360, on the product page or in the downloadable spec sheet. Only the older Freezer 50 carries one, in sones. The first Liquid Freezer III 360 printed none either, and ARCTIC's US site now lists it under End-of-Life; our ARCTIC page covers the sone ratings it does print. ENDORFY fills a full table for the Fortis 5 and leaves noise out.
How can you make a CPU cooler quieter?
On a Noctua, use the Low-Noise Adaptor in the box. It is a small cable that caps fan speed, and it costs nothing.
The figures move a long way. The NH-L12S drops from 1850 RPM and 23.9 dB(A) to 1400 RPM and 16.8 dB(A); the NH-U12A from 22.6 to 18.8. Our page of every Noctua fan and cooler figure lists Noctua's cooler figures with and without the adaptor, grouped by the fan each cooler ships with, and the same adaptor works on case fans: our PC fan guide shows Noctua's NF-S12A dropping 7 dB with it.
The rule holds beyond Noctua. Buy more cooler than you need and run it slower.
That beats buying the lowest number on the box and running it hard. A tall tower needs room, so check its height against the cooler clearances in our PC case guide, or see the low profile cooler guide for small cases.
