Some Refinements to the Gimbal Sim

[TheCholla]

And I think Edward was also onto something
https://www.metabunk.org/threads/gimbal-blender-simulation-with-clouds.12209/post-263267

"In the original, I don't think we see those most distant clouds; the range of depth seems much more limited, with anything in the distance dark and blurry."

This is problem with using the cloud-sky line to estimate the change in elevation angle. Cloud motion is less ambiguous.

 

[Mick West]

Cloud motion is less ambiguous.

It brings its own ambiguity, as we don't know the position or orientation of those clouds. The way to disambiguate it might be to create a reconstruction that recreates exactly what we see.

I'm mulling how we might use the velocity differential across the clouds. But that's probably a longer-term project.

 

[TheCholla]

The key point is that the 3-D recreation have and have always relied on recreating cloud motion to estimate lines of sight. Slight errors in motion have great consequences on the lines of sight. You could not frame it better:

Or because the simulation isn't exactly modelling reality. Do you have a better sim that includes clouds?

Yet people here, including you, have claimed that there is a incredibly coincidental plane flight path in the lines of sight. A lot.
That was too compelling to even consider context and what the aircrew described (the unphysical/mathematical close path etc).

So where is the truth? The 3-D modeling is perfect to the tenth of the degree, or it isn't?

It objectively isn't, you show it yourself. And don't get me wrong, Sitrec is very useful and I've been really impressed following the modeling work you've done. It's awesome. I've worked with models enough to know that even an imperfect model (and they all are) is useful to learn about what it helps investigating.

 

[TheCholla]

Even with a decrease in elevation angle as small as -0.1°, potential plane paths completely disappear.

 

[Mick West]

Even with a decrease in elevation angle as small as -0.1°, potential plane paths completely disappear.

Why? All that is required for a changing elevation angle is a distant plane that's climbing or descending.

 

[Harabeck]

So where is the truth? The 3-D modeling is perfect to the tenth of the degree, or it isn't?

It isn't, and I think everyone (except you apparently?), understood that from the start.

Yet people here, including you, have claimed that there is a incredibly coincidental plane flight path in the lines of sight. A lot.
That was too compelling to even consider context and what the aircrew described (the unphysical/mathematical close path etc).

The video is the hard data we have to analyze. The "context" you're referring to are anecdotes subject to biases, misinterpretations etc. If we look at the data alone, and it can be explained by something normal, we have great reason to doubt exotic explanations.

 

[TheCholla]

It isn't, and I think everyone (except you apparently?), understood that from the start.

Right. I was in team "it's not perfect" from the very start, if it wasn't obvious.
Why it's beyond me how it could be claimed that there is a super compelling straight level path in the distance, with deep probabilistic discussions of how it could not be there by accident, when a -0.1 difference in elevation angle makes it descent 400 ft (if it was a linear decrease, but it isn't, the plane also needs to go up at the end).
But I'm glad we finally seem to agree.

The video is the hard data we have to analyze. The "context" you're referring to are anecdotes subject to biases, misinterpretations etc. If we look at the data alone, and it can be explained by something normal, we have great reason to doubt exotic explanations.

The close path that stops reverse direction (in the ground track, the one radar would show) is not only an anecdote, it's a result from the data.

 

[Mick West]

it's a result from the data.

Or maybe it's a result of the camera plane stopping and reversing direction (in the component of velocity perpendicular to the camera)

Which is what you'd expect if the camera was looking at something that's basically just flying away and to the left, about 30NM away.

 

[TheCholla]

Or something that was reversing direction at a closer range. Lines of sight are a result of the plane path AND object path. I don't understand your statement.

So now the compelling distant path is just something flying away to the left? I'll leave it there then, it's not really going in a productive direction.
If in the future you add new stuff about cloud motion/visibility, I'll sure check it out. Until then...

 

[Mick West]

Or something that was reversing direction at a closer range. Lines of sight are a result of the plane path AND object path. I don't understand your statement.

You must admit the paths look curiously synchronized? Close paths are curves that mirror the motion of the platform plane, far paths are unrelated paths that look like a plane flying.

So now the compelling distant path is just something flying away to the left?

Always has been. The point (going back to @Edward Current) is that distant physically plausible paths exist (and pretty much jump out of the data, once you have the LOS), and the close paths just look like weird physically implausible projections of those paths.

 

[TheCholla]

You must admit the paths look curiously synchronized? Close paths are curves that mirror the motion of the platform plane, far paths are unrelated paths that look like a plane flying.

Crazy right? Like if the pilot was adjusting turn to the path of the object they were seeing on radar.
And yeah unrelated paths are exactly that, unrelated.

EDIT to add an example of close path, top view
EDIT2: the plane path in cyan is also its ground track, air track is different

 

[TheCholla]

jump out of the data

No, jumping out of the data would have been a convincing straight level path, and it was never really convincing anyway.

You're just moving the goalposts. There is also no linear change in elevation angle, so if now the plane can go down and up, yes everything is possible. But then I'm not interested in following your work anymore, it's just pure speculation.

 

[Mick West]

Crazy right? Like if the pilot was adjusting turn to the path of the object they were seeing on radar.

And why would they do that?

 

[TheCholla]

Why would they increase turn chasing a random distant plane they don't even see on their instruments?

 

[Harabeck]

The close path that stops reverse direction (in the ground track, the one radar would show) is not only an anecdote, it's a result from the data.

In the only hard data we have, there are solutions that could be explained by a mundane aircraft. Therefore, only the anecdotes point to something exotic.

It seems to me that the main disagreement here is actually over standards of evidence. You seem to be saying that because there are possible tracks consistent with the anecdotes, that the possible far tracks are meaningless, and the video is good evidence of an exotic phenomena. I don't think that's a reasonable conclusion. You're weighting data we can't, and probably will never, have access to, over the data we actually have.

 

[TheCholla]

The problem here is not exotic or mundane. That would be next. Here it's evaluating the model's accuracy (topic of the thread after all).

Solutions for a mundane aircraft would need to be shown with a model that does what it's built to do, represent cloud motion accurately. Distant paths are too sensitive to small variations in angles to just hand wave solutions. So because the model is inaccurate (cloud motion does not look like the video), until proven otherwise so are the lines of sight and especially in the distance. The evidence now points to a non-linear decrease in elevation angle, which means descent then climb of a random plane if you put the object somewhere in the back. This is not my standard of evidence for a resolution, sorry.

I think this kind of review is what you'd get during a scientific peer-review process. But here because it's UAP-related everything can go loose. Why? Just a waste of time to build a super-detailed full 3-D model then.

 

[Mick West]

Why would they increase turn chasing a random distant plane they don't even see on their instruments?

They were turning. The purpose of the turn is unknown.

 

[Harabeck]

So because the model is inaccurate (cloud motion does not look like the video), until proven otherwise so are the lines of sight and especially in the distance.

But by the same logic, the close tracks are out too, and all we have is the 2D image.

The evidence now points to a non-linear decrease in elevation angle, which means descent then climb of a random plane if you put the object somewhere in the back.

Assuming that's true for the sake of argument, I'm not sure why you keep bring it up as though it's some preposterous scenario, when the alternative is far more so.

 

[TheCholla]

Yes the close tracks have to be off too. But not by the same extent because angle errors rapidly increase with distance.
And I know it's a horrible sin to say that here, and I apologize, but they also make just a bit more sense with context than the rest. But as Mick said they were maybe increasing bank just for the thrill of it!

Assuming that's true for the sake of argument, I'm not sure why you keep bring it up as though it's some preposterous scenario, when the alternative is far more so.

We went from "look, there is an incredibly coincidental straight/level path in the data" to "it could be a plane doing the rollercoaster, who cares".
It's been fascinating to follow.
I sincerely appreciate the discussion.

 

[Zaine M.]

It brings its own ambiguity, as we don't know the position or orientation of those clouds.

Currently the sitrec has them at 70-120nm, where as I cant resolve, via triangulation that distance in any realistic encounter. IE I dont believe anyone is saying the F-18 has a tailwind, its more likely than not a headwind (35nm to clouds)


Source: https://x.com/LathanielS5437/status/2009394380663345446

there are solutions that could be explained by a mundane aircraft.

If you are referring to distant plane, at 32NM, I cant sustain that claim.

1. The azimuth was smoothed via markus

The smoothed seems a little more faithful so I went with that, but it doesn't make too much of a difference. With that as the source for the gimbal azimuth we can finally plot something that makes sense!

2. using these figures gives this result (distant plane is changing its heading and velocity)


Source: https://x.com/LathanielS5437/status/2009394520325325104


3. the FOV spread, using match clouds, resolves on sitrec to a distant plane but it truncates the spread by 20-25 percent. (I think i have a video on metabunk demonstrating that but i cant find it atm)


Source: https://x.com/LathanielS5437/status/2009394704056766974

Always has been. The point (going back to @Edward Current) is that distant physically plausible paths exist (and pretty much jump out of the data, once you have the LOS), and the close paths just look like weird physically implausible projections of those paths.

So no, a distant plane, "flying a kinda straight path, whilst also descending in altitude, before increasing its altitude, just at the point the F-18 turns nose on to it" doesnt "jump out of the data". You need to smooth the AZ values, then truncate them by 20 odd percent, for clouds a 100 odd NM away.

[edit - tor doesnt like images]

 

[Zaine M.]

But briefly, as there has been a lot on other threads, Is @Mick West agreeing with bank/ tilt method for de-rotation now? or is the formula still preferred?

Asking to figure out if I need to check the motion tracked go fast, both versions for curvature

 

[Mick West]

Is @Mick West agreeing with bank/ tilt method for de-rotation now?

You'd have to explain again exactly what that is. I don't recall seeing a straightforward explanation. If there was one, can you link to it?

Things get lost mid-thread.

 

[Zaine M.]

I don't recall seeing a straightforward explanation.

To de-rotate the footage, we remove the planes bank and remove the value for camera tilt. Footage is now in the global up is up position.

But to this part

All that is required for a changing elevation angle is a distant plane that's climbing or descending.

the simplest explanation for the apparent increase in distance between the clouds and object, noting your stitch gets the same thing as us.


Source: https://x.com/MvonRen/status/1831696129832181771


Is that the camera is getting closer to the target still. Watch, in the footage Marik tweeted, how the horizon and object get further apart, as the chase plane gets closer to it.

Same thing I was illustrating here, generally.

https://www.metabunk.org/threads/some-refinements-to-the-gimbal-sim.12590/post-358922

[edit - context]

 

[Harabeck]

Yes the close tracks have to be off too. But not by the same extent because angle errors rapidly increase with distance.

That's an explanation for why a closer object would be easier to model, not why a closer track is closer to the truth.

And I know it's a horrible sin to say that here, and I apologize, but they also make just a bit more sense with context than the rest.

But the "context" is a set of extraordinary claims. That's why they must be critically examined. Anomaly hunting to show that the model isn't perfect seems like a shallow pretext to throw out the analysis and accept the original claims uncritically.

We went from "look, there is an incredibly coincidental straight/level path in the data" to "it could be a plane doing the rollercoaster, who cares".

Again, accepting that as true for the sake of argument, that still seems like a fairly mundane scenario. I'm not sure why the distant would have to be perfectly level and straight to demonstrate a plausible alternative to extraordinary claims.

 

[Zaine M.]

But the "context" is a set of extraordinary claims.

Claims from the 2 first hand witnesses is in the data, we test the claims for validity (as we have also been debating Formula V Bank/ tilt to get a better contextual understanding of what is actually occurring), other wise its just a bunch of cool stories.


Source: https://x.com/LathanielS5437/status/2009405502988669130

To this again,

Anomaly hunting to show that the model isn't perfect seems like a shallow pretext to throw out the analysis and accept the original claims uncritically.

Its not a shallow pretext to disregard, the amount of manipulation (manipulation is NOT bad, eg just leveling on the artificial horizon is manipulating the data, so put better manipulation of the data for reasons) of the data (smoothing the az values to plot something now, followed by a further 20 percent truncation of those values) to even fit a distant plane in (that happens to coincidentally start increasing its altitude at the point the F-18 goes nose on, like seriously, lets just descend this distant plane, opps to much decent, lets increase our altitude again).

[edit - additional context]

 

[TheCholla]

a set of extraordinary claims

I don't see what's so extraordinary about the claims. Stop/reverse on a dime on radar (ground track), is actually a turn (probably with altitude change here) in air track. The claim that they saw this on radar is not extraordinary given high-wind and how air track can differ from ground track.
See blue line below (air track) versus ground track in green (what was seen on radar).

Now, what object is behind the IR signature, that's intriguing. But to completely discard it because it's too extraordinary, I find it subjective.

 

[Zaine M.]

@Harabeck perhaps this is a better explanation?


Source: https://x.com/LathanielS5437/status/2009410945681133668


*IF* our "side", smoothed the figures out, then truncated the values by 20 percent and said "ta da! look, anomalous motion" (to your distant plane is mundane) everyone on metabunk would jump on that claim and point out how far from the actual values it is.

Which then talks to how the clouds were ranged originally, how the "bump" predictor missed the first bump, how there has been soo much opposition to it being a pod elevation change and insistence that "its a pilots comfort additional rotation" without even qaulifying that claim other than "i think the clouds look better like this" (yeah, thats subjective AF)

[edit- so i dont do another post]

 

[jackfrostvc]

What are the comments made my the Gimbal pilot and WSO about the Gimbal object motion?
Are there links to sources of their comments about this. If so, can they be provided here please

 

[Zaine M.]

What are the comments made my the Gimbal pilot and WSO about the Gimbal object motion?
Are there links to sources of their comments about this. If so, can they be provided here please

Theres two releases, different redactions, of the Range Fouler from this encounter. Reversal of direction, ruled out spoofing specifically etc.

https://www.secnav.navy.mil/foia/readingroom/CaseFiles/UAP%20INFO/UAP%20DOCUMENTS/RF%20Reports%20Redacted%20(202301).pdf


https://documents2.theblackvault.com/documents/navy/DON-NAVY-2022-001613.pdf
Combined it reads like this,

Background: In [redacted], LT [redacted] was serving as a Weapons System Officer in [redacted]. At that time, the squadron was participating in COMPTUEX with the [redacted] as part of the pre-deployment work up cycle. Towards the end of one of the night flights [redacted] was conducting during the at-sea period, he and his pilot detected an air contact via [redacted] coming from the east and heading towards the ship. Initially thinking it may be a simulated adversary aircraft as part of the COMPTUEX scenario, he took a [redacted] lock to investigate further. The contact was at approximately [redacted]. With a stable trackfile, the two aircrew determined that it was not a "false hit" and they were able to gain a lock via the [redacted], which further indicated the vehicle had a [redacted]. It became clear via the [redacted] that there were [redacted] air vehicles flying [redacted] type formation. The pilot maneuvered the [redacted] to maintain [redacted] lock in an effort to [redacted] and try to make an ID. The air vehicle appeared to be shaped like a [redacted], resembling some type of [redacted]. Maintaining [redacted] on what appeared to be the [redacted] craft, LT [redacted] and his pilot noted what appeared to be very [redacted] that it made. Because this event took place at night [redacted] being used and the closest point of intercept being approximately [redacted] with the craft(s) was never made. Instead, all of the aerial [redacted] and headed back to wards the east, away from the [redacted]. Closest the air vehicles came to the [redacted] was approximately [redacted]. Once LT [redacted] and his pilot were back on board the [redacted] was viewed [redacted] and [redacted]), but nothing more was ever discussed or analyzed about the event after it occurred.

 

[TheCholla]

@jackfrostvc , cued up to the segment about Gimbal

Source: https://youtu.be/YhUwuaxiiQU?si=IUqDJDcBss-TuOxd&t=1147

 

[Zaine M.]

Just to my remark,

how the "bump" predictor missed the first bump,

In this video, I demonstrate what I mean, noting that Mick made remarks about

"finally there's the incredible synchronicity between the calculated roll the white line and the roll that's simply the same as the angle of the glare the green line for the first 20 seconds they are within about 2 degrees no need to adjust".

(Top video is the hypothesis that its spoofing/ war games reason, even though the range fouler directly contradicted that claim)


Source: https://x.com/LathanielS5437/status/2009420497403122093

 

[Edward Current]

I don't see what's so extraordinary about the claims. Stop/reverse on a dime on radar (ground track), is actually a turn (probably with altitude change here) in air track. The claim that they saw this on radar is not extraordinary given high-wind and how air track can differ from ground track.

It's extraordinary because the path resembles a simple math function, one that is a projection of a nearly straight-line path at a distance. In the real world there are multiple variables going on, and drag and other things scale non-linearly with velocity. As a result, if you try to get an aircraft to make a vertical turn that comes to a ground-track stop in 120-knot winds (even in laminar flow), the path will unlikely be a smooth, non-changing near-parabola as seen in the models. That is why it is unphysical!

The path could be achieved if, for example, the aircraft's computer was trying to hit this particular near-parabolic path (which also happens to be a projection of a near-straight line in the distance). Is that the claim?

By the way, why is something turning vertically in high winds tens of thousands of feet over the ocean, anyway? To me, that right there is extraordinary!

 

[Zaine M.]

one that is a projection of a nearly straight-line path at a distance.

when the los spread is truncated by 20 odd percent, after already being smoothed out?

That is why it is unphysical!

I think of it as anomalous motion, thinking DARPA not aliens, but never before seen signature and capabilities.

To me, that right there is extraordinary!

Thats what the data indicates, noting the issues with distant plane idea, the newly highlighted elevation change from the pod, which reduces the climb impact, and flattens altitude for the object throughout the encounter, that in the correctly leveled footage we see its not glare but a real object, there's no rotation of the object by 11.5/ 14 degrees required as the pod changes orientation by those amounts.

 

[Mick West]

To de-rotate the footage, we remove the planes bank and remove the value for camera tilt. Footage is now in the global up is up position.

Up is up? So why isn't the horizon horizontal?

 

[Zaine M.]

Up is up? So why isn't the horizon horizontal?

Whilst I would like to highlight that there has been no commentary about how it was determined the natural motion of the clouds was wrong, to justify implementation of the formula


Source: https://x.com/LathanielS5437/status/2008355440690950246


The horizon is horizontal, and what we are looking into are the possibilities for the cloud angle, such as,
1. the camera is at its max resolution limit, I know you say, but we can see the moon,


Source: https://x.com/MickWest/status/1480793594327797760


but that's un-realistic, IMHO, given that these are clouds, and the max effective slant range, reported by Raytheon is only 40NM slant range.

2. the clouds could just genuinely be at an angle, we aren't viewing them straight on, at 90 degrees, but some other angle


Source: https://x.com/LathanielS5437/status/2009442944663851390

 

[TheCholla]

It's extraordinary because the path resembles a simple math function, one that is a projection of a nearly straight-line path at a distance.

That's if elevation angle is nearly constant. Of course if you make the F-18 close on an object while maintaining this angle constant, it's going to induce a parabolic curve that looks mathematical. But it's an artifact of the assumption being made, if the assumption is wrong.

How do you make the determination that elevation angle is nearly constant (to a few hundredths of degree), with high confidence?

EDIT: typo

 

[Mick West]

The horizon is horizontal, and what we are looking into are the possibilities for the cloud angle, such as,
1. the camera is at its max resolution limit, I know you say, but we can see the moon,


Source: https://x.com/MickWest/status/1480793594327797760


but that's un-realistic, IMHO, given that these are clouds, and the max effective slant range, reported by Raytheon is only 40NM slant range.

What is unrealistic? I mean saying "the camera is at its max resolution limit" is totally meaningless here.

Are you actually thinking the camera can't see things past 40NM? That's utter nonsense.

 

[Zaine M.]

What is unrealistic? I mean saying "the camera is at its max resolution limit" is totally meaningless here.

Clouds giving off enough IR energy at 120 nm is unrealistic, IMHO (noting i used that qualifier).

Raytheon doesn't say 50, 60, 70, 100 plus miles.


Source: https://x.com/LathanielS5437/status/2009501633525547180


https://web.archive.org/web/20220613161524/https://www.raytheon.com/capabilities/products/atflir

So to be clear, you are asserting that the maximum effective range of the ATFLIR is,
- 120NM?
- or is it 250,000 miles (to the moon),
- or is it actually 1AU (to the sun)?

Would you like to discuss the meaning of the word "effective" in the context of the cameras military usage? Or are you saying this pod is actually designed for Astronomical imaging?

Are you actually thinking the camera can't see things past 40NM? That's utter nonsense.

1. Can you sustain the clouds being around 120 nm away to justify your claim that its effective range is further?
2. Can you provide a situation where, the pod being able to see the moon, has military advantage?

For example, and to be illustrative, reading your post I hear,

"Pilot this is your CO. Mick West said the pod can see the moon, now your target is located somewhere between 120NM and 360 NM away, we need you to image that target for us, so we can see what battle space defensive positions are there, damn it Jim, it can clearly see clouds at a minimum of that range, so don't argue."

Noting,

1. I provided the multiple different items we are looking at to discover the true reason for the angle of the clouds. The camera reaching its limit of resolving imagery being only 1 of them.
2. Still no answer as to what in the gimbal footage demonstrates that its a "pilots comfort" rotation and not its natural motion due to pod elevation.

 

[Mick West]

So to be clear, you are asserting that the maximum effective range of the ATFLIR is,
- 120NM?
- or is it 250,000 miles (to the moon),
- or is it actually 1AU (to the sun)?

Would you like to discuss the meaning of the word "effective" in the context of the cameras military usage? Or are you saying this pod is actually designed for Astronomical imaging?

I'm sorry Zaine, but what point exactly are you trying to make? You think there's a limit on how far the ATFLIR can see clouds at night, based on some intuition?

The camera reaching its limit of resolving imagery being only 1 of them.

That sounds exactly like a Flat Earth argument. And using it as a reason for why the clouds are not at the same angle as the artificial horizon is ludicrous.

 

[Zaine M.]

I'm sorry Zaine, but what point exactly are you trying to make?

I was responding to you,

What is unrealistic? I mean saying "the camera is at its max resolution limit" is totally meaningless here.

Are you actually thinking the camera can't see things past 40NM? That's utter nonsense.

You are saying that the manufacturer doesn't understand the limits of their product, or that their product doesn't actually have limitations like "an effective range", and I was asking you what the effective range is, I gave you options based on, your moon comment.

That sounds exactly like a Flat Earth argument. And using it as a reason for why the clouds are not at the same angle as the artificial horizon is ludicrous.

Can you word that differently as,
1. i did not make any definitive statement,
2. I highlighted that there are many different reasons for the cloud angle being the way that it is and
3. you still haven't made representations about what in the gimbal footage demonstrates that a pilots comfort rotation is required.

So I appreciate the humor at, at least insinuating, that I am thinking like a flat earther. (I actually take that as a compliment, metabunk has nothing else and is going to name calling). But if you could,

1. can you clear up the actual effective range of the atflir
2. can you provide repeatable test that demonstrates the clouds are as far away as is claimed, Ive already illustrated where my figures come from a number of times, and demonstrating the effective range is way beyond what Raytheon represents.