Test Equipment and Measurement Standards

What this page is, and what it is not

Many review sites show a photo of a lab full of machines and let you assume that every number on the page came from it. We do not want to do that. So the first thing to say is plain: Ranking Hours is an editorial site. We do not claim to conduct in-house testing, and we do not claim that every product we write about was measured by us.

The picture on this page is an illustration of the kind of tools that measurement work uses. It is not a photo of our own workshop.

So what is this page for? It explains how measurements work, which tools produce them, what each number really tells you, and how to judge whether a number you see in a review, on a box or in our guides is worth trusting. Reading it should make you a sharper buyer, whichever website you use.

It also explains how we treat numbers in our own articles. When we repeat a measurement made by someone else, we say that it was published by them. When we only have a manufacturer's claim, we say that too. And when there is no good number at all, we explain the factor in words and tell you what to check, instead of inventing a figure to look impressive.

Three kinds of information

Almost everything you read about a product falls into one of three groups, and mixing them up is the biggest source of confusion.

1. Specifications from the maker

These are the numbers on the box and the product page: screen size, battery capacity, stated weight, claimed battery hours. They are useful and usually accurate about things that are simply facts, like dimensions or ports. They are least reliable where they describe performance, because the maker chooses the test conditions and has a reason to choose flattering ones.

2. Independent measurements

These come from a person or lab that measures the product with instruments, using a described method. A good measurement tells you the tool used, the settings, the software version and how many units were tested. This is the most valuable kind of evidence, but it is only as good as the method behind it.

3. Direct impressions

This is what a person feels from using a product: whether the keyboard is comfortable, whether the speaker sounds warm, whether the menus are confusing. Impressions are honest and important, but they are personal. Two reviewers can disagree without either being wrong.

A trustworthy article keeps these three apart. If a page says "the battery lasts twelve hours" you should be able to tell whether that is the maker's claim, a test result or a guess.

The main measurements and the tools behind them

The rest of this page looks at the most important of these in more detail, including the traps that make a number less useful than it looks.

Screens: brightness, colour and speed

Brightness

Brightness is measured in nits, which is the same as candelas per square metre. A higher number means a screen that is easier to read in light. Be careful about what is being quoted. Many phones advertise a very high peak brightness that only appears in a small area for a short time, for example when showing a bright highlight in a video. The brightness you get for ordinary use, with a full white screen, is lower. A good test states which one it measured.

Colour

Colour accuracy is usually shown as Delta E, where a smaller number means colours are closer to the target. A value under about two is generally considered hard for the eye to notice. Coverage of colour spaces, like sRGB and DCI-P3, tells you how wide a range of colours the screen can show. A screen can cover a wide gamut and still be inaccurate, so both numbers matter if you edit photos or video.

Response and refresh

Refresh rate is how many times per second the screen updates. Response time is how fast a pixel changes colour. Manufacturers often quote the best case for response time, measured between two specific greys. Real blur depends on the whole picture. If you care about fast games, look for a test that shows motion, not only the figure.

The trap

Screens are measured with the device set up in a specific way, such as brightness at maximum or in a certain picture mode. A review that does not state its settings gives you a number you cannot compare.

Battery and charging

Battery life is the number people care about most and trust least. The reason is simple: it depends heavily on what you do.

A fair battery test uses a repeatable job, such as looping web pages or playing a video, with the screen at a fixed brightness. The device runs until it turns off, and the time is recorded. Because the job is the same every time, results from the same tester can be compared between devices. Results from different testers usually cannot, since the job and the brightness differ.

Maker claims often come from a light, easy workload, and your own use will probably be heavier. A rule of thumb is to treat a claimed figure as a best case and a good independent test as a more realistic guide.

Charging is measured with a power meter placed between the charger and the device. It shows the watts going in and how they change as the battery fills. This matters because many devices charge fast only for the first part, then slow down. A claim like "half full in thirty minutes" can be true and still say nothing about a full charge.

Capacity is shown in milliamp-hours for phones and watt-hours for laptops. You cannot compare milliamp-hours between devices with different voltages, so look for watt-hours when you can. Airlines also use watt-hours for their limits on batteries.

Heat, noise and sustained speed

A device can be very fast for ten seconds and slow after ten minutes. This is called thermal throttling, and it is a major difference between a thin machine and a thicker one with the same chip.

To check it, testers run a heavy task for a long time and record the speed, the temperature and the fan noise. A thermal camera shows where the heat collects. Contact probes give exact temperatures at certain points. A device that feels hot on the keyboard or the back is not necessarily unsafe, but it may be uncomfortable to use.

Noise is measured with a sound level meter at a stated distance, usually a short distance from the device, in a quiet room. The unit is the decibel, and it is a logarithmic scale, which means that a small number difference can sound like a big change. An increase of about ten decibels sounds roughly twice as loud. A figure without a distance or a background noise level is almost meaningless.

The room matters too. A test in a noisy room hides the quiet fans. Any serious noise measurement should state the background level.

Sound: speakers, headphones and microphones

Audio is the area with the most marketing and the least agreement, so it needs extra care.

A measurement microphone and a test rig can record the frequency response, which shows how loud each pitch is. A flat line is not always the best sound, since people enjoy a little extra bass or warmth, but big dips and peaks usually mean a coloured, uneven sound. The measurement depends on the rig, the distance and the room, so numbers from different labs are rarely directly comparable.

For headphones, results depend on the head and ear used in the test, and on how well the pads seal. For active noise cancelling, labs measure how much outside noise is reduced at different pitches. Cancelling is strongest on steady low sounds like engines and weaker on voices.

For microphones, a quiet room and a standard test phrase let you compare clarity and background noise. For earbuds, a call test in a noisy place tells you more than a spec sheet.

Most importantly, listening still matters. Numbers cannot tell you which sound you like. Use them to spot problems and to compare, then trust your own ears. Our guides on headphones and soundbars explain what to listen for.

Computers: performance scores and what they hide

performance score programs give a score for the processor, graphics, memory and storage. They are handy, and also easy to misuse.

A score depends on the whole setup: the cooling, the power limit, the driver version, the system update, even the room temperature. Two identical laptops can score differently if one has newer firmware. So a result without that context is a rough hint.

A synthetic score is not the same as real work. A chip can win a performance score and still feel slower in a program that depends on something else, like fast storage or memory. When possible, look for tests that use real tasks, like exporting a video, compiling code or running a game at stated settings and measuring frames per second.

Frame rate averages can also hide stutter. Good testers show the slowest one percent of frames as well, since that is what you feel as hiccups.

Finally, remember that new software updates change results. A performance score from a year ago may no longer describe the product you can buy today. See also our laptop guide for what to check beyond the score.

Cameras

Camera quality is difficult to reduce to a number. Labs use test charts under fixed lighting to measure sharpness, noise and colour, and they take pictures in different conditions: bright light, dim light, and scenes with both bright and dark areas, to judge dynamic range. They also compare video stabilisation and autofocus.

The trouble is that the phone or camera processes the image heavily, so the result depends on software, which changes with updates. A camera test from launch week may not describe the same phone after a few updates.

Sample pictures also deserve care. A photo shared online has often been resized or edited. Look at the original files if they are offered, and compare them in the same conditions. Our camera guide explains what matters for different kinds of photographers.

Cars, appliances and other products

Different products need different tools, but the same principles apply.

For cars, the useful measurements include acceleration timings, braking distance, interior noise, fuel or electricity use and real driving range. Official range and consumption figures come from standard test cycles, and different regions use different cycles, so figures are not directly comparable between standards. Real use, with weather, speed and load, will differ.

For home appliances, relevant measurements include energy use, water use, noise levels and capacity. Energy labels give a standard comparison, and it is worth reading what the label stands for. For vacuum cleaners, tests look at suction on different floors, battery runtime at each power mode and dust capacity. For air conditioners, cooling output, energy efficiency and noise indoors and outdoors are the key points.

In every case, a measurement is helpful when it is repeatable, described and relevant to how people actually use the product.

Why numbers differ between sources

You may see two reviews of the same product with very different results. That does not always mean that one is lying. Common reasons include:

• Different settings. Brightness, power mode, volume or resolution changed the result.
• Different software. A firmware or system update changed performance or battery use.
• Different units. Individual products vary a little, and some vary a lot. This is called sample variation.
• Different methods. A different test loop or a different microphone position gives a different number.
• Different conditions. Room temperature, background noise and lighting all matter.
• Different goals. One tester measures a best case, another a typical day.

When two sources disagree, the sensible reaction is to find out why, not to average them. If the reason cannot be found, trust the one that explains its method best.

Red flags when you read a review

• A number with no unit, no method and no source.
• Extremely precise results, like three decimals, for a measurement that is naturally uncertain.
• Test photos that look like stock images or product renders.
• Every product scoring almost the same high mark, which makes comparison useless.
• No weaknesses mentioned for any product.
• Claims of "lab tested" without saying which lab or which tests.
• Results copied from a maker's press release and described as the reviewer's own.
• No date or software version, for a product that gets frequent updates.

One or two of these may be an accident. A pattern is a reason to be careful. We hold ourselves to the same standard, and if you see us breaking it, tell us on the contact page.

How we use numbers in our guides

This is our practical rule set, and it matches what we describe on the methodology page and in our testing process page.

• We label the source of each important figure: the maker, a named independent test, or general knowledge about the product class.
• We do not present a number as our own measurement unless we measured it.
• When a number is time-sensitive, like a price or a battery estimate that changes with updates, we say so and avoid freezing it in a table.
• When two sources conflict, we point it out instead of choosing the nicer figure.
• When no reliable figure exists, we describe the factor in words and tell you what to check.
• We correct mistakes openly. See our corrections page.

We also tell you when a page contains affiliate links, and we explain how they work on the disclosure page. Links do not change the way we describe a product.

A simple checklist for you

You can use this on any review, ours included.

• Who made the measurement, and can I tell how?
• Is it a claim, a test result or an opinion?
• Are the settings and the software version stated?
• Is the number relevant to how I will use the product?
• Do other independent sources agree?
• Is it recent enough for the product I can buy now?

If most answers are yes, the number is a good guide. If not, treat it as a hint and look for more evidence before you spend money.

If you want to measure at home

You do not need a lab to learn useful things about a product you own. A few cheap tools go a long way.

A USB power meter shows how much power a charger and a device really use. A free sound-level app on a phone is not accurate in absolute terms, but it can show you relative differences between two fans in the same room. A simple battery test, like looping a video at a fixed brightness until the device turns off, tells you how your device behaves in your own life. A speed test between two devices on your own network shows what your router really delivers.

Keep notes. Write the date, the settings and the software version, so that you can compare again after an update. Repeat each test at least twice, and repeat it under the same conditions. If two runs give very different numbers, look for the reason before you trust either.

And if you test something and find a result that contradicts a guide on this site, tell us. Reader reports have helped us fix mistakes before.

Common questions

This page may be updated as our standards develop. If you find an unclear or wrong statement, please tell us on the contact page.