Gen 3 vs Photonis 4G Night Vision: Spectral Performance Comparison

By Chris Adams, Adams Industries, Inc.

Introduction

Night vision image intensifiers operate by amplifying low levels of ambient light to produce a visible image. Third-generation (Gen 3) tubes use a GaAs (gallium arsenide) photocathode (e.g. L3Harris or Elbit tubes). Photonis “4G” tubes (e.g. Echo and Intens models) use a next-generation hybrid multi-alkali (HyMa) photocathode and are an advanced form of Gen 2 technology.

A key difference in the technologies lies in their spectral sensitivity ranges: Gen 3 GaAs photocathodes are sensitive to light in the 500 nm to 900 nm range. The Photonis 4G’s HyMa photocathode is sensitive over a broader band (~350 nm up to ~1100 nm)¹. This wider spectral bandwidth means Photonis tubes can detect both deeper blue/UV light and slightly longer near-infrared (NIR) wavelengths that GaAs Gen 3 cannot.

The quantum efficiency (QE) — the probability of converting an incoming photon to an electron — differs between these technologies across the spectrum. Generally, Gen 3 GaAs photocathodes achieve higher peak QE (especially in the red/NIR), while Photonis 4G tubes have lower peak QE but extend sensitivity into spectral regions where Gen 3 has little response².

Objective

In this report, we compare the photon efficiency of current-production Gen 3 versus Photonis 4G image intensifier tubes across common nighttime environments (desert, forest/jungle, urban, tundra, mountain). To make this comparison, we combine real-world measurements of the nighttime spectrum — how much light of each color actually exists — with how well each type of tube can detect and amplify that particular color.

We computed a Spectral Match Index (SMI) for each tube. The SMI is the total number of photons (light particles) the tube can effectively convert into electrons (usable signals) across the nighttime spectrum (approximately 400–1000 nm). A high SMI means the tube does well at sensing and intensifying the available light, while a low SMI means it either can’t detect important colors or doesn't amplify them well. For instance, a tube that is great at sensing infrared but poor at amplifying it (or vice versa) will receive a lower score.

This SMI method gives us a practical way to judge how each tube performs under real-world conditions.

We also consider how environmental lighting spectra influence other performance metrics (such as image noise or HALO size), and we examine any penalties due to physical differences like photocathode size (e.g. 16 mm vs 18 mm diameter). The goal is a comprehensive, scientifically-grounded comparison to identify which environments favor each technology and why.

Environmental Night Spectral Irradiance by Terrain

Natural nighttime illumination varies greatly with environment and conditions. Key contributors to night sky brightness (in the absence of man-made lighting) include stellar skyglow (integrated starlight), airglow (upper-atmosphere emissions, notably oxygen and hydroxyl lines), zodiacal light (sunlight scattered by interplanetary dust), and occasionally galactic background. The spectral power distribution of this illumination is not uniform: as the day transitions to night, the residual ambient spectrum shifts toward longer wavelengths (red and infrared). On moonless nights, most natural light comes from faint starlight and atmospheric glow, which is strongest in the red/IR. Very little blue or UV light remains. Conversely, environments with human lighting or during twilight can have more short-wavelength content. Below we summarize typical spectral characteristics for the specified environments (assuming moonless “starlight” conditions unless otherwise noted):

• Desert (Arid, Clear Sky): Photonis often markets their products as superior in desert environments due to the increased presence of blue light. While it is true that clear desert skies allow slightly more blue/UV starlight than other biomes, the actual photon distribution at night remains heavily red/NIR dominated due to airglow and OH emissions (~700–950 nm). UV below 400 nm is essentially absent. Under a full moon, the broader visible spectrum evens the playing field — both tube types perform well. But in true desert darkness (no moon), the spectral reality favors Gen 3, which capitalizes on the abundant IR photons. So while Photonis has some edge in moderate-light conditions, in the worst-case desert night, it is Gen 3 that stands out.
• Forests and Jungle: Under a forest canopy at night, ambient illumination is further attenuated and spectrally filtered by vegetation. Foliage absorbs most visible light; however, leaves have high reflectance/transmission in the NIR (700–1000 nm) – a well-known phenomenon in remote sensing. In practical terms, a jungle on a moonless night is one of the darkest possible environments, and the light you will find is “IR-heavy.”
  
  The phenomenon referred to as “airglow” is strongest in the 800–900 nm range and can dominate the scant illumination above the canopy. However, it is noted that nightglow intensity can be variable and is “not always present” at the same levels – meaning on some nights the jungle might be almost completely dark across all wavelengths. When present, though, the IR-rich airglow light gives Gen 3 tubes the advantage. With a full moon, some dappled moonlight might reach the ground – those scattered beams would contain visible light, but heavy foliage cover often means the lower levels of the jungle still rely on multiply scattered, mostly longer-wavelength light. In the end, the jungle is practically purpose built to show the strengths of 3rd Generation Night Vision.
  
• Urban Night (Artificial Lighting): Urban environments introduce significant artificial light at night. Streetlights, building lights, vehicle headlights, etc., all contribute to a much brighter night sky than natural sources. Crucially, the spectral content of urban lighting is very different from anything produced in nature. Older generations of streetlights emitted yellow-orange light. Modern LED streetlights are typically “white LEDs” that actually have a strong blue emission peak (~450 nm) with a broad yellow-to-red phosphor glow to appear white. Metal-halide and mercury-vapor lamps produce blue-green and yellow illumination. Overall, urban/nighttime city light tends to be richer in shorter wavelengths (blue, green) compared to naturally moonless skies. For example, measurements of a city sky (Toronto) show artificial light greatly boosts the radiance at wavelengths below ~600 nm, while the contribution in the deep red/NIR remains comparatively low. The result is plenty of photons in the 400–600 nm range which Photonis 4G tubes detect well and Gen 3 GaAs tubes do not. In summary, cities provide an abundance of light, heavily weighted toward visible wavelengths – a scenario that tends to negate the low-light advantage of Gen 3 and plays to the broader-band response of Photonis. In this scenario, the Photonis 4G is able to make it a proper horse race. As J.E. Weyne in the 2020 White Paper “Differences between Gen 3 and 4G image intensification technology” put it: “with a significant amount of urban lighting, there is minimal discernible difference between Gen 2 [Photonis] and Gen 3 to the user’s eye.”

• Tundra/Snowfield: A tundra or snow-covered plain amplifies ambient light via reflection. Under moonlight, a snowy tundra can be very bright. Spectrally, snow reflects sunlight (and moonlight) fairly neutrally from ~400 nm up to about 900 nm (beyond 1 µm snow absorption increases). Thus under a full moon on snow, the spectrum is broad and fairly balanced. In a moonless scenario, snow might reflect starlight/airglow, which is predominantly longer wavelengths at very low intensity. In essence, an open tundra at a new moon is like a desert night sky (lots of NIR from airglow) but with a highly reflective ground that might scatter some light back up. This improves illumination levels in the red/NIR range. The result is that tundra, without man-made light, is IR-skewed but extremely low intensity (similar to desert). This gives the advantage to Gen 3. With moonlight, it becomes bright across wavelengths – allowing G4 to perform as well.
• Mountain (High Altitude): A mountain environment combines aspects of desert (clear air) and potentially snow. High altitude means less atmospheric filtering, so slightly more starlight may reach the surface. Humidity is low, which makes airglow emissions even more pronounced. Thus a mountain on a clear, dry, moonless night might have the strongest NIR airglow component of any environment – for instance, measurements in the Atacama at 5000 m show enhanced infrared sky irradiance relative to lower elevations. At the same time, the thinner atmosphere means stars of shorter wavelength (blue stars) suffer less extinction, marginally increasing the UV/blue portion of star spectra that reach the ground. so the spectrum at high altitude might be slightly broader (more UV through NIR) than at sea level. Still, the dominant energy at night remains in >500 nm bands (airglow and integrated starlight). If the mountain has snow cover, the reflection factor as per tundra applies. If it is bare rock, then it’s similar to desert. In either case, high-altitude nights overwhelmingly reward good IR sensitivity, while any small UV advantage is moot because there is so little UV light available.

To summarize, natural low-light environments (desert, mountains, forest with no moon) are characterized by spectra peaking in the red/NIR range (600–900+ nm) due to astronomical and atmospheric effects. Artificially lit or moonlit environments re-introduce significant shorter-wavelength content (blue-green), making the spectrum more balanced. These differences set the stage for how Gen 3 vs 4G tubes perform, as we examine next.

While you're learning about night vision, wouldn't now be a good time to look at some of the best in the business?

Spectral Match Index (Gen 3 vs 4G)

The Gen 3 GaAs photocathode and the Photonis 4G HyMa photocathode have fundamentally different spectral response curves. Figure 1 illustrates the quantum efficiency vs wavelength for each, based on known performance data and vendor disclosures (note: “quantum efficiency” is the fraction of incident photons of a given wavelength that are converted to photoelectrons). Gen 3 has negligible response below ~450 nm and then a higher peak QE (30–40%) in the red/NIR, cutting off at ~900+ nm. Photonis 4G covers ~350 nm into the near-UV and out to ~1050 nm, but with a lower peak (~20% around green). It maintains usable sensitivity through the red and into the near-IR, with a tail extending to about 1050–1100 nm, albeit QE in the far red (>900 nm) is only a few percent. In short, Gen 3 GaAs offers higher peak efficiency but a narrower spectral range, while Photonis 4G offers a wider range but somewhat lower peak efficiency in the middle of the spectrum¹. To put it plainly: Gen 3 does not see all colors, but the ones it does see, it sees a lot better than G4.

Figure 1: SMI reveals Gen 3 dominance in red/NIR-heavy environments and 4G’s urban performance advantage.

We are not going to engage in a Gen 3 night vision comparison. This paper really only addresses current production top-of-the-line Gen 3 as a contender against 4G It is still important to note that modern Gen 3 tubes often come in filmed and unfilmed variants and there are many examples of older Gen 3 technology still in service. Historically, Gen 3 tubes had a thin aluminum-oxide ion barrier film to prolong photocathode life. This came at the cost of reduced sensitivity (especially in the IR) and larger HALOs. Modern top-tier Gen 3 tubes (e.g. L3/Harris filmless) eliminate this film, improving sensitivity and decreasing HALO size. Photonis 4G tubes are inherently filmless. Thus, a top-tier Gen 3 “filmless” tube and a Photonis 4G both represent the best each technology can do (no film-related losses)². “The Gen 3 Advantage” white paper from Harris Corp underscores that, even accounting for the film, Gen 3’s photocathode has higher quantum efficiency (QE) across most of the spectrum, especially in the crucial red/NIR region². For example, if one does not normalize the curves, at 800 nm a Gen 3 GaAs might have double or more the QE of a Photonis HyMa. On the other hand, at 450 nm the Gen 3 QE is essentially zero while Photonis might have ~10–15% QE. The crossover point is typically in the green: around 550 nm, both might have comparable QE (~15–20%, depending on exact tube). Below that, Photonis leads; above that, Gen 3 leads – increasingly so toward the IR.

Another factor is gain and noise. QE is only part of the equation – the ultimate image brightness and clarity depend on how much the tube multiplies electrons and the noise floor. Both Gen 3 and Photonis 4G tubes use a microchannel plate (MCP) to amplify electrons, and both typically have automatic gain control and an auto-gated power supply to handle bright to dark transitions. The signal-to-noise ratio (SNR) metric (usually measured at a standard illumination of 10⁻³ lux with a 2856 K light source) encapsulates quantum efficiency and noise. high-performance tubes of both types achieve SNR in the mid-20s up to 30+ in lab tests (higher is better). Photonis often advertises SNR ~30 for 4G, and Gen 3 filmless are also around 30 at their peak – meaning under test conditions they can be quite similar in overall low-light performance. However, the spectral weighting of that test (2856 K is a warm incandescent-like spectrum) might not reflect extreme field conditions. The Harris white paper² criticizes Photonis’s introduction of a “spectral SNR” metric which weighted SNR by the photocathode’s spectral curve – Harris argues this obscures the fact that Gen 3 truly delivers more electrons in the red/IR, which is exactly where real night spectra lay. Indeed, when not normalized, Gen 3’s total photoresponse (PR) is higher than 4G across the realistic night spectra. This is why American Gen 3 continues to sell to top level operators, even in countries that are neighbors to Photonis and don't always see eye-to-eye with US foreign policy. This can be seen in this article. In the next section, we quantify this via the Spectral Match Index.

Method: Calculating Spectral Match Index (SMI)

For each environment, we calculated a Spectral Match Index (SMI) — measured in Obies — by integrating the product of the environment’s real-world spectral photon flux and the image intensifier’s spectral sensitivity curve across the 400–1000 nm range:

SMI = ∫₄₀₀ⁿᵐ¹⁰⁰⁰ⁿᵐ E(λ) · S(λ) dλ

Where:

• E(λ) is the ambient spectral irradiance (photons per second per square meter per nanometer),
• S(λ) is the spectral sensitivity of the intensifier (QE or relative response),
• The result in Obies represents how many useful photoelectrons a tube can produce from the available environmental light.

Environmental Spectral Data Sources

We derived irradiance spectra from published measurements and models. For dark-sky scenarios, we included starlight, zodiacal light, and atmospheric airglow (notably green at 557 nm and strong OH bands at ~840–940 nm). For urban environments, we incorporated measured city spectra with broadband LED emissions (e.g. 450–600 nm) and legacy streetlight peaks. All spectra were evaluated across the same 400–1000 nm window.

Calculation

Each environment’s spectral curve was multiplied by each tube’s spectral response curve and integrated to compute total Obies. This was done using raw values, and the results reflect how well each tube aligns with the natural photon distribution in that setting.

---

Table 1. Spectral Match Index (SMI) Comparison (in Obies)

---

Interpretation and Context

These results confirm what many operators intuitively sense: Gen 3 wins in true dark, red/NIR-heavy conditions (deserts, jungle, mountain wilderness). Photonis 4G narrows the gap in city-lit or moonlit conditions, thanks to better blue response.

Our analysis aligns with manufacturer statements: Harris emphasizes Gen 3’s superiority “in the true dark of night” — conditions where Gen 2/4G can’t keep up. Meanwhile, Photonis claims good urban and mixed-light performance, which the SMI numbers support.

We also evaluated two spectral outliers:

• Beyond 950 nm: Photonis retains some response into SWIR (up to ~1100 nm), unlike Gen 3 GaAs, which drops off sharply past 920 nm. This could matter in niche applications like detecting 1064 nm NATO designators.
• UV range (~350–400 nm): Photonis has some sensitivity where Gen 3 has none. If exotic sources (UV chemlights, bio-luminescence) are used, this could offer unique detection capabilities — but such UV sources are rare in operational environments.

In typical night settings, these fringe wavelengths contribute little to the photon budget. The meat of the night sky — and tactical relevance — still sits between 500 and 950 nm, where Gen 3 is dominant.

---

Halo and Bright-Source Performance Anomalies

Besides pure sensitivity, real-world performance includes how each tube handles bright points of light (street lamps, headlights, flares) in the field of view. A known metric here is HALO size – the radius of the glow around a bright light source in the intensifier image. Photonis 4G tubes exhibit smaller HALOs (on the order of 0.7 mm) compared to early Gen 3 tubes (~1.0 mm HALO). Over time, and largely due to gated power supplies, modern Gen. 3 tubes, including those without the ion barrier film, have been able to close the gap. Both technologies are currently neck-and-neck at ~0.7mm.

Differences in MCP Design

A smaller HALO reading means that lights “bloom” less, preserving more of the surrounding scene — a clear tactical advantage in environments like cities with oncoming headlights or sudden strobes. A natural question follows: does the color of a light affect HALO size? For example, would a red and blue LED of equal power produce different blooms?

The answer lies in the photocathode. Image intensifiers do not process color — they convert photons into electrons, and those electrons are colorless. HALO is caused by a scattering of electrons between the photocathode and the microchannel plate. Once a photon is absorbed by the photocathode and it kicks out an electron, the electron’s angle and path determine whether it hits the MCP cleanly or spreads outward in a diffuse cone. This scattering is what dictates the size of the HALO.

Importantly, color only matters if the photocathode is sensitive to it. A blue light (~400 nm) might trigger a Photonis 4G tube and produce a HALO but go completely undetected by a Gen 3 GaAs tube, resulting in no HALO simply because no photoelectrons were produced. Likewise, a deep infrared source (~900 nm) might trigger a Gen 3 HALO and be invisible to Photonis. But for wavelengths both tubes can detect, Photonis generally exhibits a tighter electron cone and therefore a smaller HALO footprint⁹. For a system with zero HALO, you simply have to leave the lens cap on.

Environment-induced anomalies in HALO are more likely related to intensity and contrast than wavelength per se. For instance, cloud cover can diffuse light and reduce HALO impact, whereas very clear air might make point sources sharper. Both modern Gen 3 and 4G are auto-gated to dim during very bright exposures, but the fact that there is more red/infrared light in the world at night skews the number of sources that may produce a HALO effect towards Gen 3. It is simply a case of seeing “more” — both good and bad.

Another environment-related effect is fixed-pattern noise (FPN) or the “honeycomb” grid that can appear at higher light levels. Photonis claims their 4G tubes have less FPN due to MCP design. In scenarios like full moon or near urban light, a Gen 3 tube image might start to show a faint hexagonal MCP pattern (noise) when it is near saturation, whereas a Photonis tube might maintain a cleaner image¹¹. This is a secondary effect, but it is relevant when comparing “image quality” in moderate-bright environments (where, as we saw, pure sensitivity differences are minimal anyway).

Tube Size and Photon Collection Area

One often overlooked factor in comparing intensifiers is the photocathode size. Standard Gen 3 tubes have an 18 mm diameter active area. The majority of Photonis tubes are also 18 mm. However, some are offered in a smaller 16 mm format (notably the Photonis echo 16 mm used in some ultralight monoculars). A smaller diameter means less total light collection if the lens and field of view remain the same. For a given f-number lens, the amount of light (number of photons) reaching the sensor is proportional to the sensor area. A 16 mm tube has about (16/18)² ≈ 0.8 (roughly 79%) of the area of an 18 mm tube – so about 21% fewer photons are intercepted from the image plane. In low-light conditions, losing 20% of the photons can translate to a noticeable drop in SNR or a need for slightly more gain. Manufacturers mitigate this by designing optics specifically for 16 mm so that, practically, the FOV or magnification is adjusted, but fundamentally there is a photon loss with smaller photocathodes. This is not a spectral issue per se, but it disproportionately affects performance in extremely dark, photon-starved environments (where every photon counts). In longer wavelengths (say >800 nm), natural sources have very low photon flux; a 16 mm tube might capture 20% fewer of those scarce IR photons compared to an 18 mm tube, exacerbating the shortfall. Thus, if one compares a 16 mm Photonis versus an 18 mm Gen 3 in identical optical setups, the Gen 3 has an inherent aperture advantage.

Photonis's counterargument is that their tubes of both sizes can achieve similar FOM and performance specs – which is true in terms of per-unit-area sensitivity¹¹. The 16 mm and 18 mm Photonis echo, for instance, have the same typical SNR and resolution specs. But the 16 mm will yield a dimmer image unless the optic is adjusted (for example, some systems use a faster lens or a different eyepiece to compensate). In practical terms, most end-user devices with 16 mm tubes do tweak the optics to maintain brightness, at the expense of a slightly narrower true field of view or a different magnification.

For our comparison, we assume a like-for-like format (both 18 mm) to isolate photocathode differences. but it is worth noting: if a Photonis 4G tube in a given goggle is the 16 mm version, it may appear about ~0.2 log units less bright than an equivalent 18 mm tube under the same conditions. That gap would be most apparent under extremely low light (starlight) and less noticeable under brighter conditions (where gain control, not photon flux, limits the image brightness). Since nobody currently makes a 16mm Gen 3 tube, we cannot make a Photonis 4G vs Gen 3 performance comparison... yet.

Conclusions and Implications

Environmental favorability: combining all the above findings, Gen 3 tubes tend to dominate in environments where longer-wavelength (600–900+ nm) illumination dominates – essentially, very dark natural conditions such as overcast starlight, new moon nights in rural areas, or under heavy foliage. In these situations, a Gen 3 intensifier can produce a significantly brighter and higher-contrast image from the scarce IR photons¹² than a Photonis 4G can. This could mean the difference between recognizing a person’s shape at 200m vs only seeing darkness or very noisy output. On the other hand, Photonis 4G tubes find their advantage in scenarios with a relative surplus of shorter-wavelength light – for example, near urban settings with mercury/LED lighting, or perhaps at high altitude under certain twilight conditions. In those cases, the Photonis tube will capture some details (like faint blue signatures or instrument backlights) that a Gen 3 might miss or show much dimmer. However, the differences here are usually subtle – as noted, under partial lighting, both tubes perform well and the image quality appears similar to most users.

Magnitude of Differences

How substantial are the differences when one technology is “favored”? Quantitatively, the Spectral Match Index (SMI) concludes that Gen 3 yields on the order of 1.5× to 2× the signal of 4G in a truly IR-dominated environment (jungle darkness). This could translate to, say, an extra 0.5–1.0 in SNR or a somewhat less noisy image for Gen 3. In contrast, in a blue-rich environment, Photonis might equal Gen 3’s performance – it is rare to exceed it by a big margin because even “blue-rich” nights (short of pure artificial light) still have plenty of longer wavelengths where Gen 3 excels. One could contrive a scenario (e.g. a scene illuminated only by a 405 nm UV lamp) where Photonis would vastly outperform (Gen 3 would see almost nothing, Photonis would see something). But in organic scenarios, the differences are usually not enough to “dominate” – rather, Photonis narrows the gap. A fair assessment is: Gen 3 provides consistently better performance across most environments and a safety margin in the darkest of nights, whereas Photonis provides acceptable performance in most environments, some real advantages in urban environments that have not “gone dark,” and some niche detection abilities (e.g. 1064 nm lasers, UV glow).

Other Considerations

We should emphasize that both Gen 3 and 4G are top-tier image intensifiers and this was not about picking a "best night vision tube." High-spec tubes from each camp often have overlapping performance in standard metrics like resolution (typically ~64–72 lp/mm), FOM (~1800–2400), and SNR (~25–30). reliability and lifespan might differ (Gen 3 tubes tend to have longer proven life in field use, whereas Gen 2/4G historically had faster degradation – Harris claims Gen 2/4G tubes lose performance quicker over thousands of hours, though modern Photonis tubes are much improved)¹³. Gen 3 tubes are ITAR-restricted in many cases for their military night vision performance, whereas Photonis (European) can supply any FOM globally, which sometimes means 4G tubes in foreign markets can actually exceed the spec of exportable US Gen 3. These geopolitical nuances aside, when choosing technology for a given environment: if maximum low-light capability in no-light scenarios is the priority (e.g. special operations in remote areas, or wildlife observers far from any sky glow), Gen 3 is the winner. If operations include a lot of mixed lighting (e.g. law enforcement in semi-lit streets, or military ops on urban fringes), Photonis 4G will perform on par and will not be overwhelmed by as many “nuisance” lights (anything red). Additionally, Photonis tubes might be preferable for integration with digital sensors or exotic applications where detecting out-of-band light (UV or SWIR up to 1.1 µm) is desired alongside the intensifier function. At the end of the day, it is really all about night vision spectrum sensitivity and the spectrum available in the environment they are to be used.

We don't carry 4G products. The closest we have is this Gen. 2+ Goggle

For NVG-ALPHA Goggles with Elbit Tubes Go Here

Closing Example

Imagine a night patrol scenario in two parts – first through a dense jungle, then approaching a lit village. In the jungle, under triple-canopy, the Gen 3 goggles clearly show slightly more detail – the ground, rocks, and warm bodies are just a bit more distinguishable thanks to the tube making better use of faint deep-red light (the Photonis user sees more grain/noise). Upon nearing the village, dimly lit by bluish LED lamps, both operators now see well, but the Photonis user notices a subtle light on the UV spectrum (say a UV beacon or the glow of certain fluorescing material) that the Gen 3 user does not – a potentially critical piece of info. Both then get near a truck with headlights on; the Gen 3 user’s view blooms out around the lights more than the Photonis user’s (larger HALO), so the Photonis user can identify a person standing closer to the headlights that the Gen 3 user struggles with. These trade-offs illustrate that neither tech is universally “better” – it depends on context.

That could be YOU out there not only seeing in the dark, but seeing in colors that are UNSEEABLE by the human eye.

End Notes

1. Weyne, J. Difference Gen 3 vs 4G. Photonis White Paper, 2020.
2. Harris Corporation. The Gen 3 Advantage. White Paper v3, 2019.
3. Popow, M. et al. Influence of Moon and Clouds on Night Illumination in Two Different Spectral Ranges. Scientific Reports, 2021.
4. The Solar Spectrum in the Atacama Desert. Scientific Reports, Nature Portfolio.
5. Rosam, H. et al. Light Conditions in Forested and Mountain Regions. Journal of Remote Sensing, 2015.
6. Kolláth, Z. et al. The Colour of the Night Sky. Journal of Imaging, 2020.
7. Barentine, J.C. Artificial Light at Night in Urban Environments. Lighting Research & Technology, 2018.
8. RemoteSensing-10-00207. "Albedo and Reflectance Properties of Snow Fields." Remote Sensing, 2018.
9. Difference_Gen 3_4G_english_version.pdf. Photonis Internal Technical Summary.
10. Exosens. Photonis 4G Performance Data Sheet, 2022.
11. Photonis. Echo 16 mm Product Specification, 2023.
12. Harris Corporation. The Gen 3 Advantage. White Paper v3, 2019. (cited again for low-light IR performance)
13. Harris Corporation. The Gen 3 Advantage. White Paper v3, 2019. (cited again)