12µm vs 17µm Pixel Pitch: Which Thermal Sensor Is Actually Better?

Thermal Academy · Core Curriculum

Open any thermal scope spec sheet and, sooner or later, you hit a number written in microns: 12µm or 17µm. It sits there next to the resolution and the NETD, quietly implying that one of them is the “better” sensor — and every manufacturer, naturally, advertises whichever number their own product happens to use. For someone who has already decided to buy a thermal optic and is now comparing two units on a shelf, that raises a fair question:

Is a 12µm thermal sensor actually better than a 17µm one?

The honest answer is more nuanced than most marketing lets on. Pixel pitch is real, and it genuinely changes how a thermal system is built — it affects fine detail, how much magnification a given lens produces, how big and heavy the optics have to be, and how hard the sensor is to manufacture. But it is one ingredient in the recipe, not the whole dish. Sensitivity (NETD), image processing, lens quality and the display often shape the picture you actually see just as much as the pitch does — sometimes more.

This guide explains exactly what pixel pitch is, in plain language, and what it does and doesn’t change. No barracking for a number. By the end you’ll be able to read “12µm” or “17µm” on a spec sheet and know precisely how much weight to give it — and, just as importantly, when to ignore it and look at something else instead.

Where pixel pitch fits in the system

Before we zoom into the pixels, it helps to see the whole chain a thermal image travels through. Heat from your target is turned into an electrical signal by the sensor; the pixel pitch describes how that sensor is laid out; a lens focuses the scene onto it; image processing cleans and sharpens the raw data; and a display finally shows it to your eye. A weak link anywhere along that chain limits everything after it — which is why no single specification, pixel pitch included, should ever be judged on its own.

The thermal imaging chain. Pixel pitch is one stage — the way the detector is built — and it hands off to four more stages before anything reaches your eye. Keep this picture in mind: it’s the antidote to spec-sheet tunnel vision.

What Does “Pixel Pitch” Actually Mean?

A thermal sensor — properly, an uncooled microbolometer focal plane array — is a tiny grid of heat-sensitive squares called detector elements, or pixels. Each pixel measures the infrared energy landing on it and turns that into one dot of your thermal picture. A 640×512 sensor has 640 of these across and 512 down: about 328,000 individual heat detectors on a chip smaller than your fingernail.

Pixel pitch is simply how far it is from the centre of one of those pixels to the centre of the next one. A 17µm sensor spaces its pixels 17 microns apart; a 12µm sensor packs them 12 microns apart. For scale, a human hair is roughly 70 microns thick — so we’re talking about detector pixels spaced four to six across the width of a single hair. It is a measure of density: how tightly the heat detectors are crowded together, nothing more.

Pixel pitch is the distance between the centres of adjacent detector pixels (the orange line). Smaller pitch means pixels are packed more tightly — so more of them fit in the same physical space.

That’s the entire definition. Everything else in this article — detail, magnification, lens size, sensitivity, range — flows from that one geometric fact and, crucially, from the choices a manufacturer makes around it.

Why Smaller Pixels Matter

If you can pack pixels closer together, one of two useful things becomes possible — and understanding which one a given product chose is the key to reading its spec sheet honestly.

Option one: keep the sensor the same physical size and fit more pixels on it. Shrinking the pitch from 17µm to 12µm frees up room. On the same-sized chip you could step up from a 384×288 sensor to a 640×512 one — from about 110,000 heat detectors to about 328,000, roughly three times the pixels sampling the same scene. More pixels across your target means finer spatial resolution: smaller temperature differences resolved, cleaner edges, more of the shape that lets your eye (and your brain) tell a fox from a fence post.

Option two: keep the pixel count the same and make the sensor smaller. A 640×512 array at 17µm measures about 10.9mm × 8.7mm. Drop to 12µm and that exact same 640×512 array shrinks to roughly 7.7mm × 6.1mm — a little over half the surface area. Nothing about the resolution changed; the whole detector just got physically smaller. That sounds academic, but as we’ll see in the lens section it has real consequences for the size, weight and cost of the optic you end up carrying.

Two sensors of the identical physical size. Shrinking the pitch from 17µm to 12µm lets a manufacturer pack far more detector elements into the same footprint — the route from a 384 to a 640 sensor without growing the chip. (Grids are illustrative, not pixel-exact.)

So far, so good for “smaller.” More pixels or a smaller chip both sound like wins — and this is exactly the part of the story marketing loves to tell. But notice what we’ve quietly assumed: that everything else stayed equal. It rarely does. Making pixels smaller costs you something too, and that’s where the honest version of this discussion begins.

Does Smaller Always Mean Better?

Here’s the part the spec sheet won’t tell you. A detector pixel works by catching infrared energy and warming up ever so slightly; the electronics read that tiny temperature change and turn it into signal. The bigger the pixel, the more infrared energy it catches — like a wider bucket in the rain. Shrink the pixel and you shrink its collecting area, so each 12µm pixel gathers less infrared energy than a 17µm one. Less signal per pixel means the sensor’s own electronic noise is relatively louder, which — all else being equal — makes it harder to resolve the faintest temperature differences.

A smaller pixel has a smaller light-collecting area, so it gathers fewer infrared photons. That’s the built-in trade-off: more pixels, but a weaker raw signal from each one — a gap the manufacturer then has to close with better sensor design and processing.

That’s why smaller pixels are genuinely harder to build well. To stop a 12µm sensor looking noisier than a 17µm one, the maker has to improve the light-sensitive fraction of each pixel (its “fill factor”), use faster optics, and lean harder on image processing to clean up the result. When they get it right — and modern sensors increasingly do — a 12µm sensor can match or beat a 17µm one. When they cut corners, the extra pixels just show you a more detailed picture of the noise.

The Relationship Between Pixel Pitch and Lens Size

This is the section that pays for the price of admission, because it explains a real, physical advantage of 12µm that has nothing to do with image quality — and everything to do with the optic you actually carry.

A thermal scope’s magnification and field of view are set by two things together: the size of the sensor and the focal length of the lens. Because a 12µm sensor is physically smaller than a 17µm sensor of the same resolution, it needs a shorter focal-length lens to produce the same field of view. The relationship is almost exactly proportional — 12 ÷ 17 ≈ 0.7 — so a 12µm sensor reaches the same view with a lens roughly 70% of the focal length.

That isn’t theory. Pixfra’s own Pegasus 2 P635 pairs a 640×512, 12µm sensor with a 35mm lens and delivers a 12.5° horizontal field of view. To get that identical view from a 17µm 640 sensor, you’d need roughly a 50mm lens. Same picture width — a noticeably bigger, heavier piece of glass.

Same resolution, same field of view — but the 12µm sensor gets there on a 35mm lens instead of 50mm. Figures reflect Pixfra’s real Pegasus 2 optics.

Both units frame exactly the same scene. The 12µm sensor’s smaller size lets it use a smaller objective lens for the identical field of view — which flows straight through to a lighter, more compact, and often cheaper optic (thermal lenses are made of germanium, so less glass means real savings).

Read the other way, the same physics gives you a choice: on the same size lens, a 12µm sensor delivers more magnification than a 17µm one. So manufacturers use the smaller pitch two ways — to make a compact, lightweight unit at a given magnification, or to squeeze more reach out of a given lens. Either way, the win here is about the package, not the picture.

Field of view is set by the sensor’s physical width divided by the lens focal length; the detail each pixel resolves (its IFOV) is the pixel pitch divided by focal length. A smaller sensor (12µm) reaches the same view with a shorter f — the root of the smaller-optic advantage.

Pixel Pitch vs Resolution — Not the Same Thing

This is one of the most common mix-ups, and it’s worth being precise. Resolution is the total number of pixels — 384×288, 640×512, and so on. Pixel pitch is how far apart those pixels sit. They are two independent dials. You can build a 384 sensor at 12µm (small chip, modest pixel count) or a 640 sensor at 17µm (big chip, high pixel count). One number tells you how many detectors you have; the other tells you how tightly they’re packed. Neither, on its own, tells you how much detail you’ll see on an animal at 150 metres.

That last part surprises people. What actually lands on your target is decided by the whole optical chain — resolution and pitch, yes, but also the lens (which decides how much of the world each pixel covers), the NETD (whether the signal rises above the noise), the processing (how that raw data is cleaned and sharpened) and the display (whether your eye can even see the detail the sensor captured). A 640 sensor behind a wide-angle lens can put fewer pixels on a distant fox than a 384 sensor behind a narrow, high-magnification lens. Resolution is a headline number; on-target detail is a system result.

Why Two 640 Sensors Can Produce Completely Different Images

Put two thermal scopes side by side, both with a 640×512, 12µm sensor and the same NETD on paper, and they can still look nothing alike in the field. That isn’t marketing spin — it’s the reality of how many stages sit between the detector and your eye, and how much room each stage leaves for a manufacturer to do a good or a lazy job.

The raw output of any microbolometer is a noisy, low-contrast grey soup. Turning it into a crisp, readable picture depends on a stack of engineering the spec sheet never lists:

This is the single most important idea in the whole article: two identical sensors are only as good as the systems built around them. It’s why the same 640 sensor can look superb in one brand and mediocre in another — and why chasing a pixel-pitch number in isolation is the wrong way to shop. (Image processing is a big enough topic that it deserves its own guide, which is on the way.)

Pixel Pitch vs NETD — Two Different Measurements

Here’s a myth with a grain of truth in it, which is what makes it stubborn. Pixel pitch and NETD measure completely different things. Pitch is about detail — how finely the scene is sampled. NETD (measured in millikelvin, mK) is about sensitivity — the smallest temperature difference the sensor can pick out of the noise. A low NETD is what lets you see a faint, low-contrast animal against ground that’s almost the same temperature. Pitch does not measure that, and NETD does not measure detail.

The grain of truth: because larger pixels collect more energy, a 17µm sensor has a genuine theoretical sensitivity advantage, and for years that showed up as lower NETD numbers on 17µm parts. That history is where the claim “17µm is more sensitive, 12µm tops out around 40mK” comes from. The problem is that it’s now out of date.

Modern 12µm sensors have closed the gap through better fill factor, faster f/1.0 optics and improved read-out electronics. Pixfra’s own 12µm units are the plain proof: the Pegasus 2 and the Cetus 640 both post ≤15 mK — comfortably better than many older 17µm sensors rated at 25mK or more.

Smaller pixels, better sensitivity — because NETD is set by the sensor’s generation, optics and processing, not by the pitch label. The “bigger pixel = more sensitive” rule is a tendency, not a law, and modern design routinely overturns it.

Does 12µm Improve Detection Range?

Short answer: not by itself. To see why, it helps to split “range” into the three jobs a thermal optic actually does, which optical engineers call Detection, Recognition and Identification (DRI). Each needs more detail on the target than the last.

The framework behind DRI is Johnson’s criteria, a rule of thumb from the 1950s that’s still used today. It says a task needs a certain number of “cycles” (pairs of light-and-dark lines) resolved across the target’s smallest dimension:

Johnson’s criteria (50% probability). The exact numbers shift with the convention used, but the ladder is what matters: identifying a target needs far more resolved detail than simply detecting it.

Notice what those criteria depend on: how many pixels you can lay across the target. And that is governed by the target’s real size, the lens focal length and magnification, the sensor resolution, the atmosphere (fog, humidity, heat haze), the thermal contrast on the night, and the image processing — a whole list in which pixel pitch is a single, indirect contributor. Pitch influences the detail each pixel resolves (its IFOV) only through the lens; change the lens and you change the range far more than the pitch ever could. Two 12µm scopes — one on a 25mm lens, one on a 50mm — will have wildly different identification ranges. Pitch didn’t decide that; the optics did.

Why Most Premium Thermal Optics Have Moved to 12µm

If 12µm isn’t automatically better, why has so much of the premium market moved to it? Because the industry has simply gotten better at making small pixels — and once you can build a good one, the smaller pitch unlocks advantages that buyers genuinely value.

Uncooled thermal sensors have marched from 25µm to 17µm to 12µm as fabrication matured. Smaller isn’t a marketing gimmick here — it reflects real manufacturing progress. (Sub-12µm sensors exist, but the energy-per-pixel trade-off gets harder the smaller you go, and gains taper off.)

For a premium optic, 12µm buys a compelling package: the resolution to fit a 640 sensor without a bulky chip, the smaller optics that keep a scope light on the rifle, and — now that sensitivity has caught up — no meaningful sacrifice in NETD. That combination, not the number itself, is why flagship models have largely converged on it. Crucially, this is an evolution, not an extinction: 17µm is not obsolete. It remains a perfectly capable, widely used technology, and a well-built 17µm optic will out-shoot a poorly-built 12µm one every night of the week.

When a 17µm Thermal Sensor Still Makes Sense

It would be easy — and wrong — to end with “just buy 12µm.” There are sound, practical reasons a 17µm optic can be the smarter purchase:

The point isn’t that 17µm is better or worse. It’s that the right sensor is the one attached to the best overall system for your budget and your job — and sometimes that’s a 17µm optic.

Common Marketing Myths, Corrected

If you strip the spec-sheet spin away, most pixel-pitch marketing collapses into four claims — and none of them survives contact with the physics.

Taken one at a time: “12µm is always better” — no; it raises a sensor’s potential, which good engineering must then realise. “Smaller pixels increase detection range” — not on their own; range is set by the lens, resolution, target size and conditions. “17µm is outdated” — untrue; it’s a mature, capable technology still used in excellent optics. And “pixel pitch is the most important specification” — usually not; NETD, lens quality, processing and display routinely matter more to the image you actually see.

Real-World Buying Advice

So how much should pixel pitch weigh on your decision? Almost always less than the marketing implies — and exactly how much depends on what you’re doing. Here’s what actually matters, by the kind of shooter you are.

Read down that table and a pattern jumps out: for every buyer, the things that actually matter are sensitivity, the lens, the processing, the resolution, and the support behind the unit — with pixel pitch a minor or negligible factor in the final decision. That’s not an anti-12µm point or a pro-17µm one. It’s simply where pitch sits in the pecking order once you’re shopping for real.

12µm vs 17µm: The Honest Summary

Neither column is a “winner.” Each describes a set of tendencies a good manufacturer either exploits or squanders. The optic in your hands is what turns these traits into an image.

So, Which Is Actually Better?

After all of that, the honest answer is the one the marketing never gives you: the best thermal optic is never determined by one specification. A high-quality thermal system is the result of the sensor, the optics, the image processing, the display and the software all working together. Pixel pitch is a real and interesting part of that equation — it shapes detail potential, sensor size, lens size and manufacturing complexity — but it is only one part, and rarely the deciding one.

So when you next see “12µm” or “17µm” on a spec sheet, give it exactly the weight it deserves: a useful clue about how a sensor is built, read alongside the NETD, the resolution, the lens, and the reputation of the optic as a whole. Buy the system, not the micron. Do that, and you’ll never be talked into — or out of — the right thermal by a single number again.

From here, the natural next steps are What Is NETD? (the sensitivity number that usually matters more than pitch), 384 vs 640 Sensors (the resolution side of the story), and Detection vs Recognition vs Identification (what “range” really means). For the frame-rate side of the same spec sheet, see Thermal Refresh Rate: 30 Hz vs 50 Hz vs 60 Hz.

Frequently Asked Questions

Going deeper: for the full breakdown of what 256, 384, 640, 1024 and 1280 sensors actually change — and why the number alone never tells the whole story — read Thermal Resolution Explained.