A decibel reading only means something when you know the range, the weighting and the accuracy behind it, and those are exactly the specifications that budget meters leave off the box. This hub compares ten current sound level meters on the figures their manufacturers actually publish, including which ones claim a standard and which ones can log. It links down to the sound measurement guides on this site.
Class Before Price
A or C Weighting
Calibrate on Site
As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Meters were compared on what each listing states: the decibel range covered, the accuracy figure, any standard or class claimed, logging capacity and whether a calibrator or certificate is supplied. We did not run laboratory tests, did not compare any meter against a reference source and publish no sound measurements of our own.
The only meter here that names a standard and ships with a calibration certificate.
Records up to 43,000 readings and exports them straight to a spreadsheet.
The tightest stated accuracy in the budget tier, at plus or minus 1.5 dB.
No pick matches that combination yet. Try a wider budget or a different measurement.
| Product | Stated range | Stated accuracy | Standard stated | Logging | Notes | |
|---|---|---|---|---|---|---|
| CRYSOUND CRY2831 | Not stated | Not stated | IEC 61672-1:2013 Class 2 | Not stated | Calibration certificate supplied | View |
| DANOPLUS data logger | Not stated | Not stated | Not stated | Up to 43,000 readings, Excel export | Viewable on meter and PC | View |
| Goldeep SL750A | 30 to 130 dB | Plus or minus 1.5 dB | Not stated | Not stated | Frequency response 31.5 Hz to 8 kHz | View |
| Tadeto decibel meter | 30 to 130 dB | Plus or minus 2.0 dB | Not stated | Not stated | Max hold and data hold | View |
| RISEPRO decibel meter | 30 to 130 dB | Not stated | Not stated | Not stated | Backlit display | View |
| TopTes TS-501B | Not stated | Not stated | Not stated | Not stated | 2.25 inch backlit LCD | View |
| Extech 407730 | 40 to 130 dB | Not stated | Not stated | Not stated | A and C weighting | View |
| Extech 407730-NIST | 40 to 130 dB | 2 dB, 0.1 dB resolution | Not stated | Not stated | Supplied with NIST traceable certificate | View |
| Extech 407732 | 35 to 130 dB | Not stated | Type 2 | Not stated | Not stated | View |
| Extech 407732-KIT | 35 to 130 dB | Not stated | Type 2 | 16,000 readings with USB download | Supplied with a calibrator | View |
Values come from each manufacturer own listing text or, where the bullet text describes a different instrument, from the product title only. Decibel figures are only comparable when the weighting and response setting match. Where a listing publishes no value the cell reads Not stated.
Work down these four in order:
If a reading will be quoted to an employer, a council or a court, the instrument needs a stated class and a calibration record. Without those, the number is an observation rather than evidence.
A weighting follows human hearing and under-weights low frequencies; C weighting keeps them. A compressor rumble that reads acceptable on A weighting can be the whole complaint on C.
Intermittent noise is invisible to a handheld. If the question involves nights, weekends or cycles, buy a logger with export and leave it running.
A 30 to 130 dB range covers domestic and most plant noise. A published frequency band tells you whether low frequency sources will be measured properly at all.
Class 1 and Class 2 are the tolerance grades defined in IEC 61672, with Class 1 holding tighter limits across a wider frequency range. Type 1 and Type 2 are the older designations from the superseded ANSI and IEC standards and are broadly equivalent grades. A meter claiming Class 2 or Type 2 is stating that it was designed to a defined tolerance band; a meter claiming neither is stating nothing.
A weighting mimics the sensitivity of human hearing and deliberately under-weights low frequencies, which is why most occupational and environmental noise limits are written in dB(A). C weighting is much flatter at low frequencies and is used for peak measurements and for characterising rumble from plant, traffic and bass. Measuring a compressor only in dB(A) can hide the very thing people are complaining about.
For any reading that will be relied on, yes. An acoustic calibrator fits over the microphone and produces a known level, usually 94 dB at 1 kHz, so the meter can be checked before and after a measurement session. The Extech 407732 kit in this list is supplied with one. Without a field check you have no way of showing the meter was reading correctly at the time.
Phone microphones are designed for speech, are compressed and limited by the operating system, are affected by the case and cannot be calibrated in any meaningful way. Apps are useful for a rough sense of whether something is loud. They are not measurements, and no jurisdiction accepts them as evidence.
It is the time constant the meter averages over: fast responds in roughly an eighth of a second and follows short events, slow averages over about a second and gives a steadier reading. Noise limits normally specify which to use, and two readings of the same source on different settings will legitimately differ. Set it deliberately rather than leaving it wherever the meter powered up.
Everything else on this page publishes a decibel range. This is the one that publishes a standard, and for anyone whose readings will be written into a report, that difference outweighs every other specification in the table. IEC 61672-1:2013 is the international standard for sound level meters, and Class 2 defines the tolerances the instrument must hold across frequency, including how far the microphone response may deviate and how the weighting networks must behave. A meter that claims it has been designed to a tolerance band; a meter that does not claim it has been designed to nothing in particular.
The calibration certificate matters for the same reason. A sound reading is only evidence if the instrument can be shown to have been within tolerance at a known date, and a certificate in the box starts that chain. CRYSOUND also states a self-developed condenser microphone, which is the component that actually decides Class 2 performance; electret capsules in commodity meters are the usual reason a cheap instrument reads plausibly at 1 kHz and poorly everywhere else.
The listing does not publish a decibel range, a numeric accuracy figure or a logging capacity, so those cells read Not stated. For workplace assessments, check whether the standard your jurisdiction applies demands Class 1 rather than Class 2 before ordering.
Noise complaints are almost never about a single moment. They are about a compressor that cycles at night, a plant room that runs on a timer, a neighbour whose music starts at eleven, a machine that peaks for thirty seconds every few minutes. A meter you stand and hold cannot capture any of that, because you are not there when it happens, and if you are, your presence changes it.
DANOPLUS states 43,000 recorded readings, on-meter review and PC review, and Excel export over the connection. The export is the part that turns the instrument into evidence: a spreadsheet can be plotted against a timeline, matched to a plant schedule and handed to an environmental health officer or a landlord without them needing the manufacturer software. Forty three thousand points is a long unattended run at a sensible interval.
The listing publishes no decibel range, no accuracy figure and no standards claim, so this is a recording instrument rather than a certified one. The sensible pairing is obvious: log with this to establish when and how often, and take the defensible readings with a meter that states a class.
Among the budget handhelds, this is the one that publishes the most and claims the best. Goldeep states a 30 to 130 dB range, accuracy of plus or minus 1.5 dB and a frequency response from 31.5 Hz to 8 kHz. Compare that against the Tadeto meter in the same tier, which states the same range with plus or minus 2.0 dB, and the difference is meaningful: half a decibel of stated tolerance is a large share of the margin in a borderline noise complaint.
The frequency response figure deserves attention because most cheap meters omit it. A stated band from 31.5 Hz to 8 kHz tells you the instrument is designed to respond down into the low frequencies where plant, traffic and bass noise live, rather than being optimised around speech frequencies and quietly under-reading a rumble. If the complaint is about a compressor or a subwoofer, that band is the specification that decides whether the meter sees it at all.
There is no standards claim, no calibration certificate and no logging. This is a spot-reading instrument for understanding a situation rather than documenting it, and for that it is the best value here.
Short answer: the CRYSOUND CRY2831 is the meter to buy when the reading has to stand up, because it is the only one here stating IEC 61672-1:2013 Class 2 and shipping a calibration certificate; the DANOPLUS data logger records up to 43,000 readings and exports them to Excel, which is what intermittent noise complaints actually need; and the Goldeep SL750A is the best budget handheld, with a stated plus or minus 1.5 dB and a published frequency band.
Three settings change the reading on the same meter pointed at the same source: the weighting network, the response time constant, and whether you are reading an instantaneous level, a maximum or an average. Change the weighting from A to C and a plant room can gain several decibels. Change the response from slow to fast and a machine that cycles will show a higher peak. Report a maximum instead of an equivalent continuous level and the number roughly doubles in perceived severity. None of that is the instrument misbehaving; it is the instrument answering a different question. Before comparing any reading against a limit or against someone else number, establish which weighting, which response and which quantity were used. Our sound level meter explainer covers the vocabulary.
Only three entries in this group make a standards claim. The CRYSOUND CRY2831 states IEC 61672-1:2013 Class 2. The Extech 407732 and the 407732 kit state Type 2, the older equivalent designation. Everything else states nothing, which means the manufacturer has committed to no tolerance across frequency at all. That is fine for satisfying curiosity about a noisy road or comparing one room against another, and it is not fine for a workplace assessment, a planning condition or a complaint that may end up in front of an authority. If the reading will be quoted to someone who can act on it, buy a meter with a class, keep the calibration certificate, and check whether the regulation in question requires Class 1 rather than Class 2 before ordering.
Most of the handhelds here state a range: 30 to 130 dB on the Goldeep, Tadeto and RISEPRO meters, 35 to 130 dB on the Extech 407732 and its kit, 40 to 130 dB on the 407730 and the NIST version. The bottom of the range matters more than the top for domestic work, because a quiet bedroom at night can sit near 30 dB and a meter that starts at 40 dB simply cannot describe it. The less advertised figure is frequency response, and the Goldeep is the only listing in this group to publish one, at 31.5 Hz to 8 kHz. Low frequency noise is the hardest kind to deal with and the most common cause of persistent complaints, because it passes through structure, is poorly attenuated by walls and is deliberately discounted by A weighting. A meter that cannot respond below about 30 Hz will under-report exactly the sources people find intolerable.
Standing in a room holding a meter tells you about the two minutes you were standing there. The noise that generates complaints is usually intermittent, nocturnal, or stops when someone is visibly measuring it. Two meters here address that: the DANOPLUS logger with a stated 43,000 readings and Excel export, and the Extech 407732 kit with a stated 16,000 readings and USB download. Both let you leave the instrument in place, capture a full cycle of whatever is happening, and produce a timeline. A timeline is far more persuasive than a single figure, because it shows duration and repetition, which is what most noise regulation is really concerned with. If you can only own one meter and the problem is a neighbour or a plant item, buy the one that logs.
An acoustic calibrator is a small device that clamps over the microphone and produces a known sound pressure level, typically 94 dB at 1 kHz. Checking the meter against it immediately before and immediately after a measurement session is the single practice that separates casual readings from defensible ones, because it demonstrates the instrument had not drifted during the work. The Extech 407732 kit is the one package here supplied with a calibrator, and the Extech 407730-NIST comes with a traceable certificate instead, which addresses the laboratory half rather than the field half. If you buy a bare meter and expect to rely on it, budget for a calibrator separately; it is not an accessory, it is part of the method.
The microphone is the instrument. A measurement-grade condenser capsule, as CRYSOUND states for the CRY2831, is designed to hold a flat response across the audible band and to behave predictably with temperature and humidity. A commodity electret capsule can read convincingly at 1 kHz, where everyone tests it, and depart badly at the extremes, which is precisely where interesting noise lives. The same argument disposes of phone apps: phone microphones are tuned for speech, gain-limited by the operating system, obstructed by cases and impossible to calibrate. Use an app to decide whether to investigate. Use an instrument to decide anything else. For the decision framework, see how to choose a sound level meter.
Meters were compared on manufacturer listing text: stated decibel range, stated accuracy and resolution, frequency response where given, standards or class claims, logging capacity and export, and supplied accessories such as calibrators and certificates. Where a listing bullet plainly described a different instrument, only the product title was used and the remaining cells were left as Not stated. We did not run laboratory tests and publish no sound measurements of our own. Amazon data supplies images, availability and links only.
Twelve devices sorted by what each listing states: WiFi and USB temperature loggers, single-use shipment recorders, an NDIR CO2 meter and a particle counter.
Read Guide→
Nine airflow instruments sorted by what each listing states: a balancing flow hood, hot wire probes for in-duct velocity, and vane anemometers with CFM.
Read Guide→Ten sound level meters sorted by what each listing states: measuring range, accuracy, weighting, IEC and NIST claims, and whether readings can be logged.
Read Guide→