Thermal vs. night vision: two different senses

The physics, in one paragraph each

Image intensification (I2 — the tube in a PVS-14) takes the tiny amount of light already present — starlight, moonlight, skyglow — and multiplies it tens of thousands of times, in real time. It sees the same world your eyes see, just brighter: you get texture, shadows, printed signs, terrain. No light at all, and a tube has nothing to amplify.

Thermal doesn't use light at all. Every object above absolute zero radiates heat, and a thermal sensor paints those temperature differences as an image. It needs zero illumination, works in total darkness, and cuts through light fog, dust, and vegetation shadow — but what it shows you is a heat map, not the visual world.

What each one wins at

• Thermal wins DETECTION: a warm animal or person against cool ground glows like a road flare. Nothing hides body heat behind a bush. This is why hunters and property observers reach for thermal first.
• I2 wins IDENTIFICATION and NAVIGATION: you can read terrain, recognize a face-shaped blur as your buddy, see the fence wire, and move through woods without walking into things. Thermal terrain reads flat and depthless — walking under thermal alone is genuinely hard.
• I2 wins SHOOTING with a helmet-mounted setup: your IR laser and illuminator are visible through a tube, so the aiming system works. Thermal cannot see near-IR laser beams at all — a PEQ box is invisible to it.
• Thermal wins through concealment and bad air; I2 wins through glass (thermal can't see through windows — glass is opaque to heat wavelengths).

Why the two ecosystems split the way they do

Hunters detect with thermal, then identify before anything else happens — because "warm blob" is not a species, and every state's rules (and basic ethics) require knowing exactly what you're looking at. Meanwhile the helmet-and-rifle world runs I2 because the whole aiming chain — lasers, illuminators, passive aiming through an optic — is built around light the tube can see. Neither camp is wrong; they're solving different problems.

What thermal sees — and does not

Thermal is the sensor people most often buy on a wrong mental model, so it is worth a section on what the picture is actually made of. A thermal imager reads long-wave infrared — roughly the 8 to 14 micron band — radiated by everything above absolute zero. It does not measure temperature directly. It measures radiated energy and converts that into an apparent temperature using an assumption about the surface it is pointed at.

That assumption is emissivity: how efficiently a surface radiates heat compared with an ideal emitter. Skin, cloth, painted wood, and vegetation are high-emissivity and read close to their real temperature. Polished metal and glass are low-emissivity — they radiate poorly and instead reflect the infrared around them, so a shiny surface on a thermal image is often showing you the temperature of the sky or the room rather than the object. It is why a vehicle reads as a hot engine bay and flat, uninformative windows.

And the whole image is built on contrast, not on absolute heat. A warm animal against cool ground is trivially easy; the same animal against ground at the same temperature is nearly invisible. The spec that describes this is NETD, quoted in millikelvin — the smallest temperature difference the sensor can resolve. Lower is better, and it is a more honest number to compare across units than resolution alone.

What changes what a thermal sees

Every one of these is a gear-evaluation fact: it is why two units behave differently on the same night, and why the same unit behaves differently on two nights.

• Distance. Detection is a pixel-count problem. Past a certain range a person is a handful of pixels, and then one. That is how the same unit can honestly "detect at 1,000 yards" and be useless for telling a deer from a dog at 300 — detection, recognition, and identification are three different ranges, and marketing quotes the first.
• Solid matter. Thermal does not pass through solid objects at all. A berm, a wall, or a parked vehicle is a complete block, not a reduction.
• Glass. Ordinary window glass is effectively opaque at thermal wavelengths. That is why a thermal scope cannot see into a lit building, and why thermal optics use germanium lenses instead of glass — and why those lenses cost what they do.
• Wet. Evaporation cools a surface, so a soaked jacket or wet vegetation reads colder than the same thing dry. It is the same mechanism that makes you cold in the rain, showing up on a sensor.
• Time. Thermal images are histories. An engine that ran an hour ago is still bright, a recently fired barrel glows, and warm footprints on cool ground fade over minutes.
• Cold and equilibrium. An object that has sat out in the cold for hours has equalized with its surroundings and has almost no contrast left to show. This is the single most common reason a scan comes up empty and the unit gets blamed.
• Low-emissivity materials, including mylar. An emergency blanket is metallized plastic: very low emissivity, very reflective at infrared wavelengths. That is exactly why it works as survival gear — it reflects your radiated heat back at you instead of letting it leave — and it is the same property that makes mylar-wrapped objects photograph oddly on thermal. One physical property, doing both jobs.
• Atmosphere. Heavy rain, snow, and dense fog scatter and absorb long-wave infrared and shorten effective range. The claim that thermal "sees through fog" holds for light haze and smoke, where it genuinely outperforms your eyes, and not for a downpour.

Myths worth naming, because they sell hardware

• "Thermal sees through walls." No. It sees the wall. A heat source close to a wall for long enough can warm a patch of it, and that is a heat stain — not an image of the thing.
• "Thermal sees through foliage." Partially, and the marketing overstates it badly. It penetrates shadow and darkness beautifully, which looks like seeing through cover. Dense leaves are a physical obstruction and reduce a target to broken fragments.
• "Camo does nothing on thermal, so camo is pointless." Two sensors, two problems. A pattern is doing its job against eyes and image intensifiers; nobody ever claimed it was a thermal feature.
• "More resolution is always better." Resolution without a good detector and a matching lens is a bigger picture of the same uncertainty. A 640 sensor with a poor NETD and a short lens can lose to a well-built 384.
• "Expensive means military grade." Physics prices this market. As the budget ladder below says, a cheap unit claiming flagship performance is usually a different technology wearing a costume.

The hard limits, stated once

Put all of it together and the honest summary is narrow: thermal is the best detector in the catalog and a poor identifier. It finds warm things fast, across distance, in total darkness, and through smoke and light haze, with no illumination of any kind.

It cannot see through anything solid. It cannot see through glass. It cannot read a face, a sign, or a fence wire. It goes flat the moment the temperature contrast does, and dense foliage or heavy weather will take most of its range away. Everything in the budget ladder below is really one question: how much detection do you need, and how much identification are you willing to give up to get it?

Running both: bridges, clip-ons, and fusion

The people who can afford it stop choosing. A common answer is a bridge: a PVS-14 over one eye and a compact thermal monocular (a SkeeTIR-class unit) over the other, so one eye navigates and one eye detects. Clip-on thermal overlays (COTI-style) project a heat outline INTO your I2 image — one eyepiece, both senses. And true fusion devices blend the two feeds in one image. All of it works; all of it costs like it sounds.

The honest budget ladder

• $300–700 digital NV: a low-light camera on your face or rifle. Real capability from a fixed position, fine for spotting and property watching. Latency and a weak low-light floor make it a poor tool for moving fast; it is the honest "try the hobby" rung.
• $600–1,500 entry thermal: compact monoculars that genuinely detect warm bodies hundreds of yards out. Low resolution and refresh — detection yes, identification barely.
• $2,000–4,000 thermal: the serious hunting tier. Resolution and refresh good enough to watch, track, and (in weapon-sight form) shoot with.
• $4,000+ I2: where a real PVS-14 with a good modern tube starts. Add the helmet, mount, and counterweight (our helmets guide covers that chain) before it's usable on your head.
• The trap tier: anything sold as "military grade" at a price from the wrong column. Physics prices this market; a $900 "Gen 3 equivalent" is a digital camera in a costume.

Digital vs. analog night vision

Digital night vision is a camera. A light-sensitive sensor reads the scene, a small screen shows it to your eye, and in true dark an infrared illuminator has to light the picture for you. It is cheap, it records what it sees, and it is genuinely good at watching one place from one spot. Analog night vision is an image-intensifier tube — the PVS-14 kind. It multiplies the light already in the sky and hands it straight to your eye: no screen, so no lag, and it works passively, with no IR lamp running. It costs several times more, and that gap is the whole trade.

Two practical consequences fall out of that. The first is lag: a sensor and a screen are always a beat behind the world, which you will not notice sitting still and will notice immediately walking over uneven ground. The second is the IR lamp — in real dark most digital units need one, and it is a lit bulb to anyone else out there on a tube.

KitID’s call: if you are buying night vision to actually move or shoot under it, wait and save for a budget analog unit. Digital is for watching, not for working.

That is a recommendation, not a rule, and it is ours — plenty of people buy digital on purpose and are right to. If the job is a fence line, a back yard, a feeder, or footage you want to keep, digital does that job well for a fraction of the money and nobody should feel talked out of it. The mistake we are trying to prevent is the other one: a few hundred dollars spent expecting a tube, on a tool that was never going to be one.

Which one first

Buy for the job, not the tier list. Watching a fence line or hunting hogs: thermal, and a mid-tier unit does the work. Moving, navigating, or shooting at night with a helmet: I2, saved for until you can do it once, correctly. And if the honest answer is "I mostly want to see what's in the back yard at 2 a.m.," a $500 digital unit answers that question for a tenth of the money — there is no shame in the bottom rung of the ladder; it's the only rung most people ever need.