What figure of merit does and does not tell you
Every night vision listing you read has a number attached to it. FOM 1400. FOM 1800. FOM 2000 plus. It is usually the largest text on the page after the price, and it is doing a lot of work in the buyer's head: bigger number, better device.
It is not that simple, and it is worth twenty minutes of your time to understand why. Not because the number is meaningless. Because it was never designed to answer the question you are asking it.
Where the number comes from
Figure of merit is a multiplication. Two numbers, multiplied together:
FOM = signal to noise ratio × limiting resolution in line pairs per millimetre
That is the whole formula. The US Bureau of Industry and Security wrote it out in exactly that form in the Federal Register, so there is no ambiguity about it.
Signal to noise ratio is how clean the image is against its own electronic noise, measured at a standardised and genuinely dark light level. Limiting resolution is the finest bar pattern a human tester can still just make out on a chart, read at the centre of the tube, in line pairs per millimetre.
You can check the arithmetic yourself against published tube specifications:
| Tube | Minimum SNR | Minimum resolution | Product | Published minimum FOM |
|---|---|---|---|---|
| Photonis XR5 | 23 | 64 lp/mm | 1472 | 1472 |
| Photonis 4G | 28 | 64 lp/mm | 1792 | 1800 |
| Omni VIII floor | 25 | 64 lp/mm | 1600 | 1600 |
Two things follow from that table, and both matter.
First, a quoted minimum FOM is usually just the minimum SNR multiplied by the minimum resolution. It is a floor derived from two other floors. Nobody measured 1472 on your specific tube. It is the worst the tube is permitted to be.
Second, because it is a product of two floors, a tube can sit exactly on its FOM floor while being well above the floor on one input and scraping it on the other. The single number hides which.
Why it exists at all
This is the part that reframes everything, and almost nobody selling night vision explains it.
Figure of merit was not invented to help you choose a device. It was invented to help an export licensing officer decide whether a tube was allowed to leave the United States, and to whom.
In 2001 the US Government concluded that the generation label, Gen 2, Gen 2+, Gen 3, was not a reliable indicator of how a tube actually performed. It needed a replacement, something a licensing officer could sort on. Figure of merit is that replacement. The US Night Vision Export Policy Implementation Guidance sets release parameters by FOM: for end users outside the privileged categories, the standard release is non-autogated devices with a figure of merit not exceeding 1400 for ground use and 1600 for aviation, with anything above requiring extraordinary justification, decided country by country.
So the number answers a bureaucratic question. May this device leave the country, and for whom. It was designed to compress a fourteen parameter acceptance test into one axis so that a decision could be made without reading the datasheet.
That is a perfectly reasonable thing for a government to want. It is a strange thing to buy a device on.
It also explains a market artefact you will notice once you know to look for it. Tube specifications cluster just above the policy thresholds. Not because 1472 is a naturally occurring quality level, but because products get engineered and sorted to land on the correct side of a regulatory line.
One more correction, because the internet gets this wrong constantly. The current US Munitions List Category XII text does not actually use figure of merit as its control criterion. It keys on autogated third generation tubes and design intent. FOM lives in licensing policy, not in the control text itself.
What the number cannot see
A tube goes through roughly fourteen measured parameters at acceptance. Figure of merit uses two of them. Here is some of what the other twelve contain.
Halo. The bloom around a bright point source. A streetlight, a headlight, a lit window. Physically it is about four times the gap between the photocathode and the microchannel plate, so it is a design property you cannot tune away afterwards. Photonis quotes roughly 1 mm for conventional Gen 3 and less for 4G. If you use a device anywhere near civilisation, halo will affect your night more often than resolution will. It is not in FOM.
EBI, equivalent background illumination. The faint glow the tube produces with no light at all reaching it. This sets the noise floor, and therefore sets the darkest scene the tube can still render usefully. It is strongly temperature dependent, so the same tube performs differently on a warm night. Not in FOM.
Photocathode sensitivity. Measured in microamps per lumen. Gen 3 typically 1350 to 2800, Gen 2 around 700 to 800. Higher is generally better, but it is not a proxy for the finished tube: the ion barrier film on a filmed Gen 3 tube kills photoelectrons, which is why a good Gen 2 type tube with lower photocathode sensitivity can match a Gen 3 on SNR. Not in FOM.
Blemishes, and where they are. Cosmetic defects are graded by count and by which of three concentric zones they sit in. A dark spot counts once it exceeds roughly 0.076 mm. Commercial grading typically permits four spots with one allowed in zone one, the centre. Mil-spec grading typically permits four with none in zone one. A tube can be figure of merit 2400 and have a spot in the middle of your field of view. Not in FOM.
Autogating. The power supply pulsing the voltage when a bright light enters the scene, so you are not blinded and the tube is not damaged. This is a property of the power supply, not the tube, so it can never appear in a tube figure of merit. Recovery speed varies enormously between systems: tens of milliseconds at the good end, around a second at the poor end. A second of blindness when a car turns the corner is a large practical difference that the number cannot express. Note that export licensing treats autogating as a separate criterion alongside FOM, which tells you how significant it is.
Distortion. Gross distortion, shear distortion, and the S shaped distortion that comes from fibre optic manufacture. Not in FOM.
Lifetime. Figure of merit is a beginning of life number. Gen 3 tubes are rated for 10,000 hours and up, some unfilmed types over 20,000. Tubes degrade, and they do not all degrade the same way. A documented Gen 2 example degrades from SNR 25 at beginning of life to SNR 11 at end of life. Two tubes with the same FOM on the day you buy them can be a long way apart three years later.
MTF, modulation transfer function. This is the most important omission and the least discussed. Limiting resolution records only the finest pattern still barely discernible: one point, at the edge of perception. MTF records how much contrast survives across the whole range of detail sizes. The practical consequence is that a tube with lower limiting resolution but better contrast at coarse and medium detail looks clear, while a tube with higher limiting resolution and poor low frequency contrast looks hazy. Both report the same resolution row on the spec sheet. Only one of them is pleasant to look through.
Why two identical numbers look different in the field
Put the measurement conditions next to the use conditions and the gap explains itself.
Figure of merit is measured in a laboratory, at one specific light level, at the centre of the tube, against a black and white bar chart, on a brand new tube, with no bright lights in the scene.
You will use the device at the edges of the field as well as the centre, in light that varies and is often far darker, with headlights and porch lights in frame, after two or three years of use, looking at low contrast natural texture rather than a test pattern.
None of the conditions match. That is not a scandal, it is what a standardised laboratory measurement is for. It is only a problem when the number is sold to you as a verdict.
Worth noting: the largest US tube manufacturer calls figure of merit the most often used metric, then immediately qualifies that it only roughly correlates to range in certain conditions. Their own buyer's guide does not print the formula.
Which of the two inputs actually matters
This is the practical heart of it.
There are two regimes. In low light, the image is limited by noise, and how good it looks depends on the light level. In better light, the noise disappears and the image is limited by sharpness and contrast instead.
On a genuinely dark night, overcast, no moon, under tree cover, you are in the noise limited regime. Signal to noise ratio is the input that matters, and resolution is close to irrelevant, because you cannot resolve a fine bar pattern you cannot see through the scintillation in the first place.
There is a measurement asymmetry that compounds this. SNR is measured in genuinely dark conditions. Resolution is the parameter that falls away fastest as light drops, which is exactly why manufacturers publish a separate high light level resolution figure. Yet resolution contributes half the FOM product.
A worked illustration, our own arithmetic rather than manufacturer data:
- Tube A: SNR 22 × 72 lp/mm = 1584
- Tube B: SNR 30 × 53 lp/mm = 1590
Those are the same figure of merit to within half a percent. In decent light Tube A looks sharper. On a dark overcast night Tube B is the better instrument by a clear margin, and Tube A will look noisy enough to be frustrating. The number cannot tell them apart.
For scale, one manufacturer measured that holding everything else constant and dropping SNR from 28 to 21 cost 8 percent of recognition distance under clear starlight.
So: of the two inputs, ask for SNR on its own. A tube that reached a high figure of merit through resolution is a tube optimised for conditions where you did not need night vision quite so badly.
One more trap: the numbers are not comparable across manufacturers
Test conditions differ by convention and by company. US and European practice measure gain at different input light levels, and because gain ratios rise as input light falls, European measured tubes report structurally higher gain for the same physical device. EBI is quoted in different units on each side of the Atlantic. Limiting resolution is read by a human observer off a chart, which is exactly as subjective as it sounds.
Comparing a US figure of merit to a European one is comparing two differently calibrated instruments. Nobody puts that on a product page.
What to ask for instead
Not instead of figure of merit. Alongside it, and weighted more heavily.
- The per unit data sheet tied to the serial number of the actual tube. This is the single highest value question you can ask. It carries measured resolution, SNR, gain, EBI and the blemish map for that specific tube, not for the model.
- SNR and resolution broken out separately. Never accept only the product.
- A range, not a floor. "FOM 1400 plus" is not a specification, it is a lower bound with the real value withheld. Within a 1400 plus listing, real tubes run from about 1400 to about 1950. The commercial incentive is obvious, and you should assume it is operating.
- Halo in millimetres, and what input spot size it was measured with.
- EBI, with its units and the measurement temperature stated.
- The spot chart with zone locations, not just a count.
- Autogating: present or absent, and the recovery time.
- Filmed or filmless, which moves halo, gain and SNR together.
- Photocathode sensitivity in microamps per lumen, understood as an input rather than a verdict.
- Look through it. At night. Next to another tube. No datasheet substitutes for this, which is most of why our showroom exists.
And phosphor, briefly
White phosphor does not change figure of merit. The manufacturer states plainly that there are no laboratory performance differences between P43 green and P45 white, and that the choice remains an operator preference. Trials found no difference in effectiveness. The clearest confirmation is that spec tables list identical FOM figures with either phosphor offered as an option on the same tube.
The honest case for white is reduced eye fatigue over long sessions and, for some people, better perception of edges. That is a real thing and worth paying for if it suits you. It is not a performance upgrade, and a price premium presented as one is a market premium, not a physics one.
The short version
Figure of merit is a sorting number that a government invented so it could rank tubes on one axis without reading the datasheet. It is a floor, not a measurement of your tube. It is blind to halo, EBI, blemishes, autogating, distortion, ageing and contrast. Half of it is a parameter that stops mattering exactly when you need the device most. And it is not comparable across manufacturers.
Use it to narrow a list. Do not use it to make the decision.
Ask us for the data sheet. Ask us to break the number into its two parts. And if you can get to Villmergen, come and look through two tubes in a dark room and decide with your own eyes, which is the only test that has never lied to anybody.
Questions at any level of detail are welcome. info@dtdsystems.ch, +41 76 225 50 04.