This page is the most important one on the site. If you only read one thing before trusting a number we publish, read this.
The short version
We do not have a listening room. We do not own the equipment we write about. We have never heard most of it.
So we do not review it.
What we do instead is narrower and, we think, more useful: we take manufacturer-published specifications, apply published acoustics and calibration standards, and compute answers you can check. Every formula is shown. Every source is linked. When a number is our estimate rather than a manufacturer’s, we say so on the page where it appears.
That is the whole method. The rest of this page is the detail.
What we publish, and why we are allowed to
There are claims whose truth depends on who is making them, and claims whose truth does not. “This speaker sounds warm in the midrange” is the first kind. You have to trust the ears attached to the sentence. “This speaker drops to 2.9 ohms and this receiver publishes no 4-ohm rating” is the second kind. You can go check both source pages in ninety seconds and find out whether we are lying.
We publish the second kind only.
Things we will tell you:
- What a piece of equipment’s manufacturer publishes about it, with the measurement conditions attached, because the conditions are what make a spec mean anything
- What a calculation says, with the arithmetic exposed
- What a published standard requires, quoted and cited to the standard itself rather than to somebody’s summary of it
- Where a manufacturer’s own numbers are internally inconsistent, ambiguous, or conflict between their regional sites
- Which products satisfy a constraint you have given us, and which do not
Things we will not tell you:
- How anything sounds
- Which of two products is “better” in any sense we cannot compute
- That we tested something we did not test
- Anything at all about a product’s build quality, reliability, or long-term ownership experience
If you want to know how a speaker sounds, we are the wrong site and we would rather say so than pretend. Go read the people who actually put it in a room: What Hi-Fi and Stereophile for listening impressions, Audio Science Review and Erin’s Audio Corner for measurements. We link to them constantly and we are not competing with them.
Where our numbers come from
Product specifications come from the manufacturer’s own product page, published spec sheet, or owner’s manual. Not from retailer listings, not from other publications, not from aggregator databases. Every specification in our tools and articles carries a link to the manufacturer source.
Standards are cited to the standards body. When we say Dolby reference level is 105 dB peak, we point at SMPTE RP 200 and the Dolby Atmos Specifications, and we show the two-step arithmetic that gets from the 85 dBC calibration level to the 105 dB figure, because that step is an addition we are performing rather than a sentence anyone published.
Third-party measurements, when we reference them, are always attributed and linked to whoever did the work. We do not re-host measurement data as though it were ours. The independent measurement community in audio is small, unpaid, and doing the most valuable work in the field. Taking credit for it would be both dishonest and stupid.
When we cannot verify something
We leave it blank and say why.
This sounds obvious and almost nobody does it. A specification we could not find on a manufacturer source is recorded as unavailable, not filled in from a plausible-looking third-party listing. Our database currently has null values for, among other things: Klipsch’s US pricing, Yamaha’s US pricing, several manufacturers’ 4-ohm ratings, and Revel’s frequency response tolerance. Those gaps are real and we would rather show you a gap than a guess.
We dropped Anthem’s MRX receivers from our first database release entirely because Anthem’s own specification pages returned errors and we would not publish their power ratings on the strength of a retailer listing.
Where our model is weak
Our headroom calculator combines a manufacturer’s 1-metre sensitivity figure, the standard power relationship, the free-field inverse-square distance term, and an additive room-gain correction. It is an estimator, not a measurement, and it is wrong in specific ways we want you to know about:
The room-gain figure is the weakest number we publish. There is no standards body value for it, because it depends on room volume, absorption, speaker placement, and whether you are sitting inside or outside the reverberation radius. We use 3 dB as a convention and let you change it. Treat it as carrying roughly 3 dB of genuine uncertainty in either direction.
The inverse-square term is an approximation indoors. It holds strictly only in the direct field. In a typical domestic room the critical distance is often one to two metres, so a listener at a normal three-metre seat is already in a mixed direct-and-reverberant field where level falls off more slowly than the formula predicts. Applying full inverse-square loss and then adding a room-gain bonus is a crude two-term stand-in for a transition that deserves better physics.
Below the room’s Schroeder frequency, none of this holds. Boundary reinforcement in a small room can add 6 to 12 dB in the bottom octaves while modal nulls subtract just as much at a different seat. No single scalar captures that. Our number is a broadband estimate and says nothing useful about your bass response.
Sensitivity specifications are not comparable across brands, and we correct for this. Klipsch publishes half-space measurements that read about 3 dB higher than the full-space figures most brands use, so we subtract 3 dB before comparing. Nearly every manufacturer specifies at 2.83 volts rather than 1 watt, which are the same thing into 8 ohms but not into 4, so we normalise per model. Magnepan specifies at a single 500 Hz tone. Q Acoustics prints a row label and a unit that contradict each other. We flag each of these on the page rather than quietly averaging them into a false precision.
Our all-channels-driven derate is our estimate. Almost every receiver manufacturer publishes only a 2-channels-driven power figure, which overstates what the amplifier delivers with seven or eleven channels running. Only two brands in our database publish both numbers: NAD’s T 778 drops 2.2 dB from its 2-channel to its all-channels rating, and Arcam’s AVR11 drops 1.2 dB. We apply a 2 dB derate as a conservative midpoint of those two. Where a manufacturer publishes their own all-channels figure, we use theirs and ignore ours.
Money
We earn affiliate commission when you buy through some of our links. That is how the site is funded and we would rather be blunt about it than bury it.
Three commitments that follow from that:
No calculator output is affected by any commercial relationship. The tools run on published specifications and fixed formulas. There is no branch in the code that knows or cares whether we have an affiliate relationship with a manufacturer. You can read the arithmetic on every result.
Recommendations are derived and stated as derived. When we say a product fits, the sentence looks like “these are the models whose published minimum impedance and sensitivity satisfy the constraint you entered.” It never looks like “we loved these.” We do not have an opinion about how they sound and saying we did would be the exact lie this whole page exists to prevent.
We will name products we cannot earn on. If the answer to a question is a used unit, a brand with no affiliate program, or not buying anything, that is what the page will say.
Disclosure appears before the first affiliate link on every page that carries one, per FTC guidance.
Corrections
We will get things wrong. Specifications change, manufacturers update pages, and we will misread something.
If you find an error, tell us at corrections@hifisystems.net. We will fix it, and we will note on the page that it was corrected and when. Corrections are not embarrassing. Quiet edits are.
Our gear database carries a “last verified” date. If that date is old, the data may be stale, and you should treat it accordingly.
Who is responsible for this
Emilio Goris operates HiFiSystems.net and is responsible for everything published here.
To be clear about credentials, because this is where sites in this category usually start inventing things: no formal training in acoustics or audio engineering, and no professional history in the audio industry. What is here instead is a documented method, primary sources for every claim, arithmetic you can check, and a stated willingness to publish the gaps.
If that is not enough for a given question, it is not enough, and the right answer is to go to someone with a measurement rig. We will tell you who.
Contact: corrections@hifisystems.net · Last updated: 8 August 2026