A Gimbal Glare Explainer

It's symmetric to the flashlight about the center of the image, which is typical for an internal reflection.
Same as this:
The little cluster of dots, on the other hand, moved around the screen when I moved the camera. In the example below, it appears as a tiny fleet of UFOs above the flashlight:

P.S. after you attach an image, in edit mode, click "insert" on the attachment icon, then select "full image" to display it in your post.
 
abstrakt-leuchtendes-licht-sonne-platzen-mit-digitalen-linseneffekt_3247-304.jpg
my personal definitions:
Lens flares = caused by the lens system, sharply defined artifacts, placed away from actual light source
Glare = caused by atmosphere, diffuse shape, around actual light source

The spike fits neither because
Spike = caused by the lens system, sharply defined, around actual light source
Blue is what I have always known as a diffraction spike.

Usually in telescope images the shape/number of the diffraction spikes relates to the configuration of the mirror supports, but they are often seen as desirable and maybe created using string etc.

https://www.celestron.com/blogs/knowledgebase/what-is-a-diffraction-spike
 
Blue is what I have always known as a diffraction spike.
Article:

See also​

  • Bokeh, a source of circles around out-of-focus bright points, also due in part to the internals of the lens.
  • Diffraction spike, a type of lens flare seen in some telescopes

Yeah, it's only Wikipedia...
Usually in telescope images the shape/number of the diffraction spikes relates to the configuration of the mirror supports, but they are often seen as desirable and maybe created using string etc.

https://www.celestron.com/blogs/knowledgebase/what-is-a-diffraction-spike

A diffraction spike is the light you see extend from a star in your astro-images. A diffraction spike is caused by how light bends or diffracts around an object or in this case the support beams in your secondary mirror in reflecting telescopes. This spike isn’t present with refractors as it doesn’t have a secondary mirror.
diffraction_Spikes_reflector_medium.jpg

You can manually create diffraction spikes [..] with strings you place on the front part of your optical tube.
diffraction_Spikes_all_large.jpg

diffraction_Spikes_camera_large.jpg

Content from External Source
 
Last edited:
abstrakt-leuchtendes-licht-sonne-platzen-mit-digitalen-linseneffekt_3247-304.jpg
my personal definitions:
Lens flares = caused by the lens system, sharply defined artifacts, placed away from actual light source
Glare = caused by atmosphere, diffuse shape, around actual light source

The spike fits neither because
Spike = caused by the lens system, sharply defined, around actual light source

The spike also affects the ability to see things behind it, and therefore satisfies the given definition of glare. However, that's because it's so broad it covers most things. I can think of at least 5 different things that could cause, or legitimately be described as, glare just on the top of my head:
- in-front-of-optics imperfections (which I think is the thing that rotates in these UAF vids);
- optical imperfections (the point spread function, and don't ever think "perfection" is achievable, even the best optics are somehow imperfect);
- aperture effects ([edit - oops, not Farey!]Airy rings, can sometimes be lumped in with the PSF, as it's just an intrinsic property of the system);
- internal reflections (typically "lens flare" where you see the shape of the aperture, but there are other types); and
- sensor overloading (bloom).

Most of those effects have some kind of rotational symmetry to them, mostly because there's no prefered direction for the effect to have. Lens flare being an internal reflection of course has the mechanics of the camera defining preferred directions, and hence the typical cascades of heptagons in straight lines. However, in-front-of-optics imperfections can be completely arbitrary, and if they are caused by an oily cloth wiping a surface can have a very strong directionality to them. Mike demonstrated this perfectly with I think a thumb-smear on a microscope slide between 2 iphones in one of his vids.
 
Last edited:

diffraction_Spikes_reflector_medium.jpg[/FONT]
Content from External Source

I love these, as they demonstrate the layering of the distortions - great find! Firstly, the bars/strings will cause a dotted diffraction pattern of ghost images (the red dot image), but the other unavoidable distortions in the system then smoosh that into the smooth spikes and stars that we actually see.
 
FatPhil mentioned blooming. This is caused in the CCD (or whatever semiconductor you use to capture the image). If you use a phone to demonstrate a rotating spot of light, blooming can also be the cause since you rotate both the optics and the CCD at the same time.
When shooting a bright light source such as direct sunlight, the source of the light may appear as a bright halo with column or line defects around the highlight. This phenomenon appears when the subject illumination is so strong that the sensor pixels become saturated, this saturation usually appears as a white area on the image which is proportional to the intensity of the illumination source.
Source

Example from source:
1236a.jpg
 
The definitions and usages of "glare" and "lens flare" vary in the optic literature, and are applied (often simultaneously) to different things.

However I'm using it to mean stray or spreading light around a visually small light source. This is generally described/modelled/measured in the literature using a "glare spread function" (also "point spread function") - which, quite simply, is a function that tells you how much radiation spreads around any given point in an optical system. This is discussed at: https://www.metabunk.org/threads/the-shape-and-size-of-glare-around-bright-lights.10596/

Example:



The glare spread function is often described as a one dimensional function (It's also often show with a logarithmic scale, as the center is thousands, sometimes millions, or times brighter than the transparent fall-off), but as you can see from the glare shape above, it's a more complex 2d function, here's a real-world measurement from https://graphics.stanford.edu/papers/glare_removal/glare_removal.pdf
2022-01-19_08-10-33.jpg
(Note streaks/rays of light, and the secondary reflection, a lense flare)

Very large "spikes" or "streaks" are most commonly seen with a lens (or window, or mirror) that has been incorrectly cleaned and has physical microscopic streaks in one direction. Wipe your phone camera with a finger (slightly greasy, touch your nose) to see the effect.

Smaller spikes come from the optical system of the camera. They are diffraction artifacts, and can show the shape of the aperture, or other edges and corners in the optical path that result in diffraction. My small Countor-HD camera has an oval-shaped glare even with clear front glass.

A "shaped glare" can come from a combination of things.
 
FatPhil mentioned blooming. This is caused in the CCD (or whatever semiconductor you use to capture the image). If you use a phone to demonstrate a rotating spot of light, blooming can also be the cause since you rotate both the optics and the CCD at the same time.

Source

Example from source:
1236a.jpg
This is getting off-topic. That image shows both spreading glare (from the optics) and sensor blooming (the streak). Glare is often mistaken for sensor bloom.

The bottom line here is quite simple. No matter what you call it, glares with shape and spikes appear in IR systems. We have examples.

 
The definitions and usages of "glare" and "lens flare" vary in the optic literature, and are applied (often simultaneously) to different things.

However I'm using it to mean stray or spreading light around a visually small light source. This is generally described/modelled/measured in the literature using a "glare spread function" (also "point spread function") - which, quite simply, is a function that tells you how much radiation spreads around any given point in an optical system. This is discussed at: https://www.metabunk.org/threads/the-shape-and-size-of-glare-around-bright-lights.10596/

Example:



The glare spread function is often described as a one dimensional function (It's also often show with a logarithmic scale, as the center is thousands, sometimes millions, or times brighter than the transparent fall-off), but as you can see from the glare shape above, it's a more complex 2d function, here's a real-world measurement from https://graphics.stanford.edu/papers/glare_removal/glare_removal.pdf

(Note streaks/rays of light, and the secondary reflection, a lense flare)

Very large "spikes" or "streaks" are most commonly seen with a lens (or window, or mirror) that has been incorrectly cleaned and has physical microscopic streaks in one direction. Wipe your phone camera with a finger (slightly greasy, touch your nose) to see the effect.

Smaller spikes come from the optical system of the camera. They are diffraction artifacts, and can show the shape of the aperture, or other edges and corners in the optical path that result in diffraction. My small Countor-HD camera has an oval-shaped glare even with clear front glass.

A "shaped glare" can come from a combination of things.
Yes. Also, the "Glare Spread Function" is actually just the BRDF (or BSDF) that is practically the amount of (stray) light in a steradian, as induced by the scatter properties of any surface (lens, windows, mirrors). And the curve shown is made up of a few common components, that means for instance the peak in the middle is the actual source, and the "bump" caused by overal roughness of the surface(s). Sometimes peaks can be seen, caused by regular patters for instance (wiping..).

(this just to add)
 
@Mick West : I have a comment/question before this new work is unleashed to the public.

If Gimbal is 30 Nm away, based on the FOV it's about 20m long. What is the size of a small plane engine? 1-2m diameter? Is it possible that an engine creates a glare an order of magnitude larger than its size ? I don't see that in the small plane FLIR shown above. This is the only one that really looks like Gimbal, but the glare is rather small and not obscuring the entire plane. Certainly not 20m wide

I know you're focusing on the rotation only, but your point will likely shift the discussion towards Gimbal being a plane, and it raises many other questions.
 
If Gimbal is 30 Nm away, based on the FOV it's about 20m long. What is the size of a small plane engine? 1-2m diameter? Is it possible that an engine creates a glare an order of magnitude larger than its size ? I don't see that in the small plane FLIR shown above. This is the only one that really looks like Gimbal, but the glare is rather small and not obscuring the entire plane. Certainly not 20m wide
That's a good question. My guess is we could only be certain by directly comparing a jet to Gimbal, as photographed on the same (or at least similar) system, using the same zoom setting. The example above isn't shot at .35°, I don't think. The ATFLIR also seems to be more sensitive (the sky is considerably darker in black-hot).
 
My point is, these glares are a couple meters wide, Gimbal is 20m wide (for a 30 Nm distance, worse if further away).
Judging by that familiar 'bonfire' case with the huge rotating diffraction spike, there isn't any close relationship between the size of a glare and the size of the hot object. In that case the 'spike' was many times wider than the fire.

I recall that previously someone did some work on the distance of the Gimbal object based on the 2 degree downward line-of-sight from the Navy jet. I think it was this post by jplaza , but there may be others:

https://www.metabunk.org/threads/gi...f-bearing-and-or-dcs.11836/page-8#post-261727

I did a few back-of-envelope calculations myself, simply using the relationship (tangent of 2 degrees) = x/y, where x is the vertical distance of the Gimbal object below the level of the Navy jet, and y is its horizontal distance from the jet. Given x, we can solve for y, or vice versa. For example, if x = 1000 feet, y = about 5.4 miles. (y = x/tan 2deg, = (1000/.035) ft = 28571 ft = ~5.4 miles) Similar calculations give:

x = 2000 ft y = ~10.8 mi
x = 3000 ft y = ~16.2 mi
x = 4000 ft y = ~22.6 mi
x = 5000 ft y = ~27 mi
x = 10000 ft y = ~54 mi.

The relationship is of course linear, as one would expect.

On a flat earth, if the altitude of the Navy jet is 25000 ft, as shown on the display, then the altitude of the Gimbal object would be (25000 - x) ft, e.g. for x = 2000 ft, the altitude of Gimbal would be 23000 ft at a distance of ~10.8 mi, or at a distance of ~54 mi it would be 15000 ft. On the globe, allowance needs to be made for the 'drop' with distance, which is not linear. This is allowed for in jplaza's model. My impression is that it doesn't make a huge difference until the distance is very large, say beyond 50 miles. Even at 50 miles it would only make a difference of a few thousand feet.

Of course all these figures assume the inclination is 2 degrees, but the figure is only given on screen to the nearest degree, so the true inclination could be anywhere between 1.5 and 2.5. For 1.5 all the estimates for y would need to be about 50% longer, and for 2.5 somewhat shorter, e.g. for x = 10000 the range would be between 43 and 72 miles.

One implication is that for relatively short distances, say 10 miles or less, the object would be only a few 1000 feet lower than the Navy jet, which would be consistent with it being another Navy jet, but rather low for an airliner at cruising height.
 
That's a good question. My guess is we could only be certain by directly comparing a jet to Gimbal, as photographed on the same (or at least similar) system, using the same zoom setting. The example above isn't shot at .35°, I don't think. The ATFLIR also seems to be more sensitive (the sky is considerably darker in black-hot).

What blows my mind is that after 4 years we still cannot have a definite answer to this simple question : are fighter pilots familiar with this kind of glare obstructing aircrafts in their ATFLIR targeting pod ?

We are stuck with endless speculations based on 3-4 videos, under different conditions, that we don't know if they use a comparable instrument. Aren't there anybody who know fighter pilots and can ask them this simple question ? Or maybe they do not use the ATFLIR that often ?
 
What blows my mind is that after 4 years we still cannot have a definite answer to this simple question : are fighter pilots familiar with this kind of glare obstructing aircrafts in their ATFLIR targeting pod ?

We are stuck with endless speculations based on 3-4 videos, under different conditions, that we don't know if they use a comparable instrument. Aren't there anybody who know fighter pilots and can ask them this simple question ? Or maybe they do not use the ATFLIR that often ?
It's likely not an issue in normal operations since these targeting pods are typically used at shorter ranges, where you'd be able to see more details of the aircraft being tracked. That's also something that's been consistently stated by the pilots "you should see the aircraft outline". Ryan Graves said the pilots believed the thing was within 10 Nm, where they would've had every reason to believe that the glare shape was actually the object. That situation is very much outside their common experience.

I'd like to hear the end of the sentence "Well the FLIR's looking...". If, as I expect, he was about to say the FLIR was looking in one direction while the radar indicated another, that could help establish that what they were seeing on the radar and what they were tracking on the FLIR were not the same object.
 
Is it possible that an engine creates a glare an order of magnitude larger than its size ?

Glare size depends on brightness, focus, diffraction effects, and exposure. At a very simple level, the answer is yes, because we have examples. Consider the size of these engines.

engine flares banked closeup overlay.jpg

Here's a video from my garage showing how this occurs


Here are some distant small planes obscured by their engine glares.


And a closer one (showing turbulence, as it's quite close)


And one showing a similar shape, but where the plane is still faintly visible behind it.
 
Thanks @Mick West , I'm still on the fence as it is hard to know how these videos really compare to the ATFLIR from a Super Hornet. I find it strange that such high-tech targeting pods would give you a big IR glare when tracking a simple plane, to the point that you are completely confused and don't know what you're looking at (as the pilots are in Gimbal).

I'd like to hear this is absolutely common, from fighter pilots themselves. I don't recall Lehto saying it is, but I have not watched all his videos.
 
Thanks @Mick West , I'm still on the fence as it is hard to know how these videos really compare to the ATFLIR from a Super Hornet. I find it strange that such high-tech targeting pods would give you a big IR glare when tracking a simple plane, to the point that you are completely confused and don't know what you're looking at (as the pilots are in Gimbal).

I'd like to hear this is absolutely common, from fighter pilots themselves. I don't recall Lehto saying it is, but I have not watched all his videos.
Isn't the whole point that it is not that common which is what lead to it being a UAP?
 
Isn't the whole point that it is not that common which is what lead to it being a UAP?
How could it be uncommon if IR glare is inherent to an ATFLIR system ? They are puzzled before it makes any rotation, so it seems it is an UAP to them just by the look of it on the FLIR image. If IR glares such as this are common, why are they puzzled ?
 
How could it be uncommon if IR glare is inherent to an ATFLIR system ? They are puzzled before it makes any rotation, so it seems it is an UAP to them just by the look of it on the FLIR image. If IR glares such as this are common, why are they puzzled ?
Because normally the target is too close to be obscured by the glare.

This is what they normally see:

Source: https://www.youtube.com/watch?v=vwdV7x_Rmqo


Imagine taking the Sukhoi into the distance so it shrinks almost to a point. The infrared from the engines is mostly directional, so it wouldn't decrease quite as much, and the glare (while shrinking in absolute terms) would become larger with respect to the jet, eventually obscuring it. The glare is something they're used to ignoring, but here it's all that's left!
 
The glare is something they're used to ignoring, but here it's all that's left!
That would mean tracking a plane far away from behind is something unusual then. I don't know, when I imagine training I suppose they try the ATFLIR on various targets, close and far. I have absolutely no knowledge of military operations but I feel like in a combat situation 30 Nm is not super far. At 350 Knots you get there (or it comes to you) in 5 minutes. I wish I knew a fighter pilot, I would take him/her for a drink and ask a million questions :)
 
The glare is something they're used to ignoring, but here it's all that's left!
There's also the asymmetry/shape. IR glare seems to be typically circular, or at least symmetrical. But here we have something that looks kind of like a teardrop in white-hot, and a flying saucer in black-hot.
 
It makes you wonder what kind of analysis AATIP performed, and as others have said the fact that video was named GIMBAL gives a clue to what it was meant for originally.
 
to the point that you are completely confused and don't know what you're looking at
if they had a different set of targets on the radar at a different range, that'd explain the confusion
At 350 Knots you get there (or it comes to you) in 5 minutes
if it is flying at you, you don't see its engines = no glare

if it is not flying at you, you can't catch up to it easily, as 350 knots is not the closing speed
 
I find it strange that such high-tech targeting pods would give you a big IR glare when tracking a simple plane, to the point that you are completely confused and don't know what you're looking at (as the pilots are in Gimbal).
Maybe they weren't really that confused.

Pilot 1: Dude, it's F#&*ing drone!
Pilot 2: There's a fleet of them.
 
Could it be that they instinctively flip out of IR mode and into visible light mode when the glare's too great in IR?
Maybe, and maybe it was hazy this day (hence no EO footage, though there are other possibilities (maybe they avoid switching modes so they don't lose the track).
 
That would mean tracking a plane far away from behind is something unusual then. I don't know, when I imagine training I suppose they try the ATFLIR on various targets, close and far. I have absolutely no knowledge of military operations but I feel like in a combat situation 30 Nm is not super far. At 350 Knots you get there (or it comes to you) in 5 minutes. I wish I knew a fighter pilot, I would take him/her for a drink and ask a million questions :)
Yes, it would be nice to get some confirmation of this. My understanding is that ATFLIR and similar systems are mostly used in A/A mode to satisfy certain rules of engagement that demand positive identification of the target before opening fire. This would make it slightly pointless to use at longer ranges. It likely depends on the culture as well -- the Russians have tailored their military technology to American stealth, for instance, so their aircraft and training emphasize use of IRST, which has been showing up in American jets only relatively recently.
 
there was some argument in connection with the rocket launch hypothesis over whether the encounter occurred by night or by day; I don't remember if we reached a definite conclusion.
That's a good point. I've been tacitly assuming the encounter happened during the day (which also puts some options on the table like a mylar balloon reflecting sunlight) but really we don't know!
 
I'll put this out there, I know it will be unpopular, but we're here to exchange ideas after all, you make what you want out of it.

I've been watching the rotation section of the video in repeat, and I think both sides of the argument may be right here. Meaning there is an artificial rotation due to the pod happening, but there is also real rotation of the object.

A giveaway to me is the bit I posted in the other thread, at 0'30. I upload the section I'm talking about below.

At 30' we see a clear rotation of the entire image. The clouds, the object, and the light patterns are moving together. This is in contradiction with the glare hypothesis, because one of the pillar of the glare hypothesis is that it should not rotate with the background.

This rotation is smooth, it is different from the "step" rotations happening three times : 0'27, 0'29 just before the smooth rotation, and 0'33. If it was all from the pod roll I don't see why the background would rotate with the object in one case, but not in the others.

My take on this : this object is not a glare but a real object that emits a strong IR signature. It rotates doing three sharp steps, but in the middle of it a gimbal roll happens (exactly between azimuths 1L and 1R, as you would expect from the pod passing the 0 azimuth).

This is why it's so confusing, because there is indeed rotation induced by the pod roll, but also real "step" rotations by the object.

This is also consistent with the shape of the object being very consistent when seen in black hot mode at the end. I would expect a glare to show a little bit of change in shape when rotating, but that does not happen here.


 
At 30' we see a clear rotation of the entire image. The clouds, the object, and the light patterns are moving together. This is in contradiction with the glare hypothesis, because one of the pillar of the glare hypothesis is that it should not rotate with the background.

This rotation is smooth, it is different from the "step" rotations happening three times : 0'27, 0'29 just before the smooth rotation, and 0'33. If it was all from the pod roll I don't see why the background would rotate with the object in one case, but not in the others.
It's different because it coincides with the jet banking angle going clockwise. Background rotates when the jet rolls (i.e. a change in bank angle), the glare rotates when the pod head rolls relative to the pod body.

There's a long smooth rotation here partly because it's at the steepest part of the curve, but also because of the jet rotation clockwise which has the effect of making the curve steeper (try toggling "Use real bank angle" to see the effect) and also rotating the background in the same direction.

There's nothing there that suggest an actual rotating object - all the rotation is still matching the expected glare rotation curve. Just that in one option of one of the pod rolls there's a jet roll in the opposite direction
 
It's different because it coincides with the jet banking angle going clockwise. Background rotates when the jet rolls (i.e. a change in bank angle), the glare rotates when the pod head rolls relative to the pod body.

There's a long smooth rotation here partly because it's at the steepest part of the curve, but also because of the jet rotation clockwise which has the effect of making the curve steeper (try toggling "Use real bank angle" to see the effect) and also rotating the background in the same direction.
I understand your point, but between 1L and 1R both the pod and plane rotate (bank). So I would expect the glare to rotate more than the background, but it seems to perfectly follow the whole image rotation.
 
I understand your point, but between 1L and 1R both the pod and plane rotate (bank). So I would expect the glare to rotate more than the background, but it seems to perfectly follow the whole image rotation.

It's not though. The glare is rotating about twice as fast as the background - which is essentially rotating the same as the artificial horizon.

Make the graph and the video large in the sim to examine exactly what is going o. Here it is at 1x and 0.25x speed. You can also single step.



The red line is the bank angle (i.e. the horizon angle in the video). The green line is the glare angle. The glare starts to rotate at 28.75 seconds. The jet (and horizon) start to rotate at 29.54 seconds, but at half the speed. The rotations end at about the same time, maybe a few frames earlier for the jet.

So really there's no coupling.
 
For anyone fixated on the idea that the apparent shape of the Gimbal object must represent its true shape, it should be sufficient to point to the example of the FLIR1 video. In that case the apparent shape of the object in white-hot IR mode is utterly different from its apparent shape in TV mode (both at x2 zoom for fair comparison). They can't both represent its 'true' shape, unless we suppose that it 'shape-shifts' at exactly the moment when the ATFLIR system switches from IR to TV mode. An alien spaceship might have a shape-shifting capacity, but how would it know when to shift? It doesn't receive any signal from the ATFLIR, which is a passive sensor (unless it fires its laser, which as far as we know it didn't in this case). Do we have to appeal to quantum entanglement??

We can only regret that the Gimbal video (at least in the publicly available form) never shows the object in TV mode. It might have saved a great deal of argument and effort if it had.
 
FLIR is not a glare right? It's an IR signature of a physical object. Here the question is not whether we see the true shape of the object, rather if it is its IR signature, or just a glare. The fact that the objects stays the exact same shape while rotating, at the end when seen under the same conditions (black hot, small azimuth) is what I find surprising for a glare.
 
It's not though. The glare is rotating about twice as fast as the background - which is essentially rotating the same as the artificial horizon.

Make the graph and the video large in the sim to examine exactly what is going o. Here it is at 1x and 0.25x speed. You can also single step.
I see that now, thanks.
 
Back
Top