As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
The native array size turns up in exactly four places: a bullet that names two numbers in one sentence, the technical details block beneath the bullets, a photograph of a specification label or an on screen menu, and the downloadable manual. Search those four in that order and stop at the first one that gives a flat figure with no verb of improvement attached to it. If none of them does, the honest conclusion is that the decisive hardware number has not been published, and the right response is to treat the instrument as undocumented rather than to assume the flattering reading.
The four places a native array size actually appears
Start with the bullets, but read them for structure rather than for claims. The useful bullet is the one that contains two different grid figures in the same sentence joined by a word such as enhanced, upscaled or effective. That sentence has told you the whole story: the smaller figure is the hardware and the larger one is the output. A single figure standing alone in a bullet tells you nothing about which of the two it is.
Second, drop below the bullets to the technical details block. On a retail page this is usually populated from a structured product record rather than written by hand, and it frequently carries a sensor or detector field that the bullets omit. It also carries the field names the manufacturer uses internally, which is useful on its own: a record that distinguishes an infrared resolution field from an image resolution field is telling you that the two are different quantities even when the values look similar.
Third, read the photographs. Manufacturers who print a specification panel on the box, on the back of the instrument or on an about screen in the menu system often photograph it, and that panel is written by engineering. The same applies to screenshots of the settings menu, where a resolution option listing the available output sizes implicitly brackets the array.
Fourth, go to the manual. It is slower and it is the only one of the four that is written as a specification rather than as a sales document.
| Where to look | What a useful entry looks like | How much weight to give it |
|---|---|---|
| Feature bullets | Two grid figures in one sentence, joined by a word describing a transformation | High when both numbers appear, none when only one does |
| Technical details block | A separate field for the sensor or detector, distinct from the image field | Moderate, usually closer to the engineering record than the bullets |
| Product photographs | A specification panel on the housing or packaging, or an on screen menu | High when legible, because it is printed by the maker |
| Downloadable manual | A specification table with the array, the field of view and the spot ratio together | Highest available, and the only source that also gives measurement conditions |
Language that signals a processed figure
Marketing copy leaves fingerprints, and the most reliable one is a verb. A hardware figure is stated flatly because there is nothing to explain. A processed figure needs a sentence around it, and that sentence almost always contains a transformation: improves to, boosts to, upgrades to, enhanced by, delivers an effective, achieves, reaches, equivalent to. Each of those presupposes a starting point that the sentence has declined to name, and the presupposition is the clue.
A second fingerprint is the adjective super, along with its relatives advanced, intelligent and enhanced, attached to a word that is otherwise a plain technical term. A third is a figure quoted with no unit of hardware context: a grid size with no field of view, no detector pitch and no optics description is a figure about a file rather than about a sensor. A fourth is internal inconsistency across the listing itself, where the title, a bullet and the details block give three different numbers. Inconsistency is not necessarily dishonesty, but it does mean nobody in the chain was treating the figure as a specification.
Cross checks that pin the number down without the manual
When the four places fail, a few indirect checks narrow the range. Processing that doubles each linear dimension is by far the most common, so a stated figure is often exactly twice the array in each direction. Quadrupling appears occasionally and halving never does, which means the array is almost never larger than the stated figure.
Frame rate is another lever. Genuine micro scanning combines several captures into one output frame, so it cannot produce composite frames at the same rate the detector reads out. A listing that claims an enlarged grid at a full video refresh rate is describing an operation performed on single frames, which is interpolation.
Sensitivity cuts the other way and is worth understanding so you do not over read it. A noise equivalent temperature difference in millikelvin is a per element property, and larger elements collect more energy and generally post better figures. A very good sensitivity figure alongside a modest array is a coherent design rather than a contradiction, so it neither confirms nor denies a large array.
Finally, the physical evidence. Lens aperture is visible in photographs, and a larger array at the same field of view generally needs a longer focal length and a larger aperture for the same optical speed. An instrument whose window is the size of a shirt button is not hiding a large array behind it.
When the listing publishes nothing at all
The failure mode is not unique to thermal cameras, and the clearest examples of it come from instruments where the decisive figure is a detection threshold rather than a grid. The Testo gas leak detector carries no feature bullets whatsoever on its listing, so the sensor type, the gases covered and the sensitivity are all absent: Not stated in the listing bullets. The Inficon combustible gas detector does have bullets, and they consist of a product category, a country of manufacture, a manufacturer name and the package dimensions. That is catalogue metadata rather than specification, and the figure a buyer needs is nowhere in it.
Partial publication is the more common case. The plug in natural gas and carbon monoxide detector states a ten year sealed battery life, a loud alarm with a flashing light and continuous monitoring from a wall outlet, and describes its sensor only as highly responsive, with no alarm threshold and no sensor technology named. The Robinair LD7 refrigerant leak detector publishes a colour graphic display with bar graph and sweep modes, an audible alarm and ultraviolet lighting for dye work, and gives no detection sensitivity figure, which is the number that separates one leak detector from another. If you want to see how much that single figure governs the choice, the explanations of how a refrigerant leak detector works and what a combustible gas detector responds to both turn on it.
What to do when the array size stays hidden
Three options remain, and all three are better than guessing. Ask the manufacturer directly, using their own support channel rather than the sales listing, and ask for the figure in the form you need it: the number of elements across each dimension and the field of view in degrees. Second, buy from a source that publishes a manual before purchase, because the existence of a downloadable specification document is itself a quality signal about the product record. Third, widen the comparison to products that do publish, and let the absence count against the ones that do not.
That last step is the one people skip, and it is the most effective. When two instruments look similar and only one has published the figure that determines what it can resolve, the published one has given you a way to be wrong on purpose rather than by accident. The same reasoning underlies any careful instrument comparison, including the way a multi gas instrument is assessed on what it states per channel, how a single gas detector review is forced to say which figures are missing, and why a head to head such as this carbon dioxide monitor comparison spends most of its length on published numbers rather than on impressions.
Frequently Asked Questions
Where is the array size most likely to be hidden?
In the technical details block that sits below the bullets rather than in the bullets themselves. The bullets are written by a marketing team and the details block is often populated from a product database, so the two disagree surprisingly often. When they disagree, the details block is usually closer to the engineering record, although neither is authoritative in the way a manual is.
Can I work the array size out from a sample image?
Sometimes, and it is worth trying. If a listing shows a screenshot or a saved file and you can establish the pixel dimensions, you have the output grid. If the picture shows visible square blocks of uniform colour, counting how many blocks span the frame gives the array directly, because interpolation smooths edges but a nearest neighbour or lightly processed view does not. Treat it as a strong hint rather than a specification.
Does the field of view help pin the number down?
It helps you use the number rather than find it, and the two together are what you actually want. A field of view in degrees divided by the element count across that dimension gives the angle each element covers, which converts directly into a target size at a working distance. A listing that gives a field of view but no array size has given you half of a calculation, and a listing that gives neither has given you nothing you can act on.
Is the manual always more reliable than the listing?
Generally yes, because a manual is written by the people who built the instrument and is reused across sales channels, while a listing is rewritten by whoever is selling it today. Manuals also carry the figures that marketing finds inconvenient: the measurement spot size, the accuracy conditions, the warm up time and the calibration interval. The listing is the advertisement and the manual is the closest thing to a specification most consumer instruments publish.
What if the seller answers a question with a number?
Treat it as better than nothing and worse than a document. Answers in a question and answer section are often written by sellers rather than manufacturers and are frequently wrong about exactly this kind of figure, because the person answering is reading the same bullets you are. A number that also appears in the manual or on a specification page is confirmed. A number that exists only in a reply is a lead to check.
Does an absent figure mean the hardware is poor?
It means it is undocumented, which for a measuring instrument is a defect in itself. Some genuinely capable instruments have thin listings because they are sold mainly through distributors who supply a datasheet separately. But when the decisive figure is absent and a processed figure is prominent, the most economical explanation is that the processed figure is the flattering one. Buying on an undocumented hardware number means accepting that you cannot calculate what the instrument will resolve.