Sitrec - Metabunk's Situation Recreation Tool - Development Discussion

Wow that flip is crazy. You can see the vector pointing consistently in the same direction basically. In your sim 340° of heading until the flip. So it's a straight line and then inverts direction.
It immediately motivates the description of the aircraft changing heading; and since this is 100% simulated, we know that the aircraft causing this vector to flip keeps flying straight, if I understand @Mick West correctly. It shows that the description need not be reliable, that it could be an honest mistake.
 
Last edited:
F18: 12m wing span, 37m² wing area, 45k ft.
Citation: 18m wing span, 49m² wing area, 51k ft.
GlobalHawk: 35m wing span, 50m² wing area, 60k ft.

I'm not confident that this would affect the IR profile much, especially from a horizontal
aspect; if you see the aircraft from below or above, they're obviously quite distinct. The point is, if all you have is a few pixels of glare, they're not that different, since the engines are close to the centerline. And it answers this:
When I say you need big wings what I mean are many things. For example aspect ratio is crucial.



The F-18 has an AR of 4.
Cessna Citation X Aspect ratio: 7.8
B52: 8.56

That's twice the aspect ratio with a much wider wing spans. Makes a HUGE difference in performance.

For reference:
U2 has an AR of 23 with a ~70k ft service ceiling
SR-71 has an AR of 1.939 ~85k ft (Like I said.... it was basically a ram rocket... no wings needed!)
 
When I say you need big wings what I mean are many things. For example aspect ratio is crucial.

If all you have is a few pixels, the glider wings are going to be harder to see (blending into the background more) than the Cherokee's wings (provided the wing area is comparable).
 
It immediately motivates the description of the aircraft changing heading; and since this is 100% simulated, we know that the aircraft causing this vector to flip keeps flying straight, if I understand @Mick West correctly. It shows that the description need not be reliable, that it could be an honest mistake.
Wait I'm lost... the flip is just simulated in the video and not the result from a path in sitrec?

If the vector flips like that it means you changed heading 180° and are now heading in the opposite direction. You can't be flying in a straight line or your vector would always be pointing to the same heading.
 
Wait I'm lost... the flip is just simulated in the video and not the result from a path in sitrec?
"Simulated" means "derived from the simulation data", aka the "path in sitrec". It's not a contradiction.

What I meant by "100% simulated" is that there is no ambiguity as to how this display came to be; there's no actual sensor data involved because sitrec is 100% synthetic. But the SA display is supposed to work as if the data was real. So when sitrec simulates a straight flight path for the target and the SA page flips the vector, the same thing should've been displayed in the cockpit if that was the real situation.

Which begs the question, would observer pilots realize the target is flying straight when they see its vector flip?
 
Wow that flip is crazy. You can see the vector pointing consistently in the same direction basically. In your sim 340° of heading until the flip. So it's a straight line and then inverts direction.
It's programmed to be a straight line in that example as that's the only way you can fit an overhead view with the description of stopping, and then reversing direction.

However, I think this raises a lot of questions as to how the narrative emerged from the SA page. Next step would be to add the "fleet"
 
Wait I'm lost... the flip is just simulated in the video and not the result from a path in sitrec?

If the vector flips like that it means you changed heading 180° and are now heading in the opposite direction. You can't be flying in a straight line or your vector would always be pointing to the same heading.
The is the result of a path - one that viewed from the side looks like this:
2022-05-16_04-06-11.jpg
It's climbing constantly, and the horizontal velocity slows to zero and reverses. The aspect vector is only showing horizontal velocity.
 
Toggling the "MAP" button on the SA will now add the full tracks of the jet and the target. This is the default (10 NM, Constant Speed)
2022-05-16_04-48-32.jpg

And this is a straight line solution (heading of 0)
2022-05-16_04-50-41.jpg

The amount of actual movement is very small, and it's hard to see how the narrative of flying along, stopping as a fleet of five other targets swung around it, and then heading back, would have been discernable.

I've asked Graves to draw a diagram of what the movement would look like on the SA, but he has been reluctant to do so.
 
If all you have is a few pixels, the glider wings are going to be harder to see (blending into the background more) than the Cherokee's wings (provided the wing area is comparable).
Agreed. But usually long wings = long fuselage too (for balance).

The reason I'm saying all this because we should take into account only realistic flight paths. We can disqualify a lot of solutions to remain within the flight path of known aircraft types.
Also: we know the indicative size of the object. If it's a b52 we would visually see a difference. Huge thick wings with engines under them would show up on IR. I would exclude that. Business jet or drone is the most viable option here if this was a misidentification but for each combination/scenario we need to take into account speed altitude and envelope.
The is the result of a path - one that viewed from the side looks like this:
2022-05-16_04-06-11.jpg
It's climbing constantly, and the horizontal velocity slows to zero and reverses. The aspect vector is only showing horizontal velocity.
Ok thanks!
That's an Immelman basically (https://en.wikipedia.org/wiki/Immelmann_turn)?

We can exclude a lot of aircraft types if that is the actual manoeuvre here.

At what altitude? Might be tuff even for an F-18 at max power. Can you share the settings in sitrec to get this path?
(EDIT: Ok got it and I understand what you mean maybe.... you are using the "straight line" LOS traversal method to force the path to be a straight line. For some reason I can't figure out the altitude by myself though. The graph for altitude seems broken and I don't see it go up when the Immelman happens).
 
Last edited:
Here I'm trying to replicate what was described. It needs the fleet to go up to 4200 knots
 
If by "basically" you mean "following the same math function from beginning to end," which according to measurements taken from my model is a 3rd-order function (backward–upward acceleration that increases at a constant rate).

As I've said elsewhere, I'd like to know how any kind of aircraft, whether conventional or unconventional, achieves such a trajectory in a real world with gravity, wind, and drag.
 
Graves does mention "airspeed bands." What do those look like? Could they indicate a "stop"?
Do you have a quote in context? I don't think he is referring to something visual on the SA. I think you need to place the cursor over a target to see speed and angels. He probably meant "bands"="ranges". "We noticed the objects were usuamoving at airspeed ranges between mach x and mach y"
 
Do you have a quote in context? I don't think he is referring to something visual on the SA. I think you need to place the cursor over a target to see speed and angels. He probably meant "bands"="ranges". "We noticed the objects were usuamoving at airspeed ranges between mach x and mach y"
It's from his AIAA talk.
slide formation.jpg
 
Doing a bit of image analysis coding, and thought I'd start with this line detection, loosely based on a similar tool in the IPACO suite
https://metabunk.org/sitrec/?sit=linetest
2022-05-26_17-27-42.jpg
Drag an image onto the left side (or use the default sample), then drag a box around a region, then click "Best Angle, Full 180" to try to align it with a line

The idea is to detect barely visible lines holding things up, or strings under balloons.
 
Here I'm trying to replicate what was described. It needs the fleet to go up to 4200 knots
This is an admirable attempt to reconcile the description given by Graves with reality.
Doing a bit of image analysis coding, and thought I'd start with this line detection, loosely based on a similar tool in the IPACO suite
https://metabunk.org/sitrec/?sit=linetest
2022-05-26_17-27-42.jpg
Drag an image onto the left side (or use the default sample), then drag a box around a region, then click "Best Angle, Full 180" to try to align it with a line

The idea is to detect barely visible lines holding things up, or strings under balloons.
That's pretty cool!

Guess you are getting bored waiting for Graves to get around to providing the clarification that proves the greatest UFO sighting ever was legitimate, guess he's got other things to do eh.
 
Added the bare bones of Aguadilla.
https://www.metabunk.org/sitrec/?sit=agua2022-06-06_10-02-08.jpg

2022-06-06_10-03-13.jpg

This introduces the ability to pull in maps (using Mapbox, or OpenStreetMap) with terrain elevation.

Lots to do still. In particular, I need a 3D track editor. Also probably need to do the video in a different way here, as it's 7x longer than the ATFLIR videos, so loading all the frames individually might not be the best approach.
 
I like it. I have some of the data (such as the SCU object flight path) modelled in GE if its any help...?
 

Attachments

  • Screen Shot 2022-06-06 at 19.44.25.png
    Screen Shot 2022-06-06 at 19.44.25.png
    1.9 MB · Views: 115
  • Aguadilla Object Analysis.kmz
    8.9 KB · Views: 150
Took a detour to add a couple of new ones.

The view of San Jacinto from Pt. Dume, with the ability to flatten the earth (by increasing the radius)
https://www.metabunk.org/sitrec/?sit=dume
I just did this one because it occurred to me it would be pretty easy.
2022-06-11_10-29-51.jpg

And the St.Ives Hayle beach drone (no drone yet, just setting up the view). Double click to make one of the view full-window
https://www.metabunk.org/sitrec/?sit=hayle
2022-06-11_10-31-16.jpg

2022-06-11_10-30-40.jpg

I'm moving towards things being data driven. The entire configuration for Dume is now just:
JavaScript:
export const SitDume = {
    name: "Dume",
    azSlider:false,
    jetStuff:false,
    animated:false,

    lookFOV: 10,

    terrain: {lat:  34.001856, lon:-118.806196, zoom:9, nTiles:8},

    fromLat: 34.001241, // point dume
    fromLon:-118.806459,

    fromAlt: 300,
    fromAltMin: 300,
    fromAltMax: 1000,

    toLat: 33.814029,
    toLon: -116.679419,
    toAlt: 3302.2,


}
 
2022-07-24_08-37-40.jpg

  • Added a menu selector for all available sitches (also renamed "sit" as "sitch")
  • Made it not crash on non-chromium browsers and instead just displays a warning instead of video
  • The sitches with longer video work on Chrome, Edge, and theoretically, Opera
 
Hey @Mick West - I'm on holiday this week in Hayle, staying about 200m from that location. Let me know if you need any ground truth gathered....!
Cool! Not super important at this point - but if you see any drones being flown there, you could maybe get a shot of them, even video, from the path. Also a shot of the camera?

But really, don't worry about it, have fun!
 
Just been down to ground zero at Hayle. Here's some pictures for you @Mick West . Didn't see anyone flying drones, but the Cove Cafe seems to me (as a drone flyer myself) to be the perfect spot to fly from.

Me on the webcam
Screenshot_20220808-162939_Chrome.jpg
Panorama of the area

Looking back towards the webcam
20220808_163011-01.jpeg

Webcam is top left of this image
20220808_163032.jpg

The Cove Cafe20220808_163427-01.jpeg
 

Attachments

  • 20220808_163834.jpg
    20220808_163834.jpg
    167.4 KB · Views: 242
Last edited:
With the ability to use KML ADS-B Track, it made it very easy to add this Lake Michigan Tic-Tac - Just needs two tracks (camera and target) and a start time and some display tweaks like sky color and fov.
https://www.metabunk.org/sitrec/?sitch=lakemichigan



Other sitches can be created with just variations of this code:
JavaScript:
import {SitKML} from "./Sit-KML";
import * as THREE from "../../three.js/build/three.module";

export const SitLakeMichiganTicTac = Object.assign(Object.assign({},SitKML),{
    name: "lakemichigan",
    sitch: "lakemichigan",

    planeCameraFOV: 5,

    frames: 1200,
    terrain: {lat: 42.933424, lon: -86.129730, zoom: 9, nTiles: 8},
    files: {
        KMLFile: './FlightAware_UAL267_KBOS_KDEN_20211017.kml',
        KMLTarget: "./FlightAware_DAL2474_KORD_KDTW_20211017.kml"
    },
    startTime: "2021-10-17T15:50:27.766Z",

    videoFile: "videos/118963_submitter_file1__IMG2065.mp4",
    brightness: 100,
    skyColor:new THREE.Color().setRGB(0.53, 0.81, 0.92),  // daylight sky

})
 
With the ability to use KML ADS-B Track, it made it very easy to add this Lake Michigan Tic-Tac - Just needs two tracks (camera and target) and a start time and some display tweaks like sky color and fov.
https://www.metabunk.org/sitrec/?sitch=lakemichigan



Other sitches can be created with just variations of this code:
JavaScript:
import {SitKML} from "./Sit-KML";
import * as THREE from "../../three.js/build/three.module";

export const SitLakeMichiganTicTac = Object.assign(Object.assign({},SitKML),{
    name: "lakemichigan",
    sitch: "lakemichigan",

    planeCameraFOV: 5,

    frames: 1200,
    terrain: {lat: 42.933424, lon: -86.129730, zoom: 9, nTiles: 8},
    files: {
        KMLFile: './FlightAware_UAL267_KBOS_KDEN_20211017.kml',
        KMLTarget: "./FlightAware_DAL2474_KORD_KDTW_20211017.kml"
    },
    startTime: "2021-10-17T15:50:27.766Z",

    videoFile: "videos/118963_submitter_file1__IMG2065.mp4",
    brightness: 100,
    skyColor:new THREE.Color().setRGB(0.53, 0.81, 0.92),  // daylight sky

})
Hi Mick, this seems to crash my Chrome after a few seconds. Not sure if you need any info to debug..

Also would be cool to simulate this one https://www.metabunk.org/threads/reddit-pilot-sees-a-ufo-from-n14aq-probably-dal2369.12561/
 
Hi Mick, this seems to crash my Chrome after a few seconds. Not sure if you need any info to debug..
If you could see if there's a red message in the Javascript console, or if it differed significantly from this at the end:
2022-09-03_08-42-51.jpg

View -> Developer -> Javascript Console

Also what OS and Chrome version
 
If you could see if there's a red message in the Javascript console, or if it differed significantly from this at the end:
2022-09-03_08-42-51.jpg

View -> Developer -> Javascript Console

Also what OS and Chrome version

64index.f1d36a2b68653a50109f.bundle.js:1 Added via init promises
64index.f1d36a2b68653a50109f.bundle.js:1 ++++++++++++++++++++RESOLVED SEAMS
index.f1d36a2b68653a50109f.bundle.js:1 Terrain loaded, executing callback
www.metabunk.org/:1 WebGL: CONTEXT_LOST_WEBGL: loseContext: context lost
index.f1d36a2b68653a50109f.bundle.js:1 DOMException: Decoding error.
error @ index.f1d36a2b68653a50109f.bundle.js:1
index.f1d36a2b68653a50109f.bundle.js:1 THREE.WebGLRenderer: Context Lost.

Windows 11 and Chrome Version 105.0.5195.102 (Official Build) (64-bit)

It's the main playback window that crashes..
 
t's the main playback window that crashes..

Of course, now it's NOT crashing for me.

This might be because of my rather brute-force video rendering. Works for me on my M1 Mac Mini, 16GB, but seems slow in Windows (Parallels). Takes quite a while before the video pops up.

Alternatively, it might be the rendering of the long tracks.

Does Aguadilla work for you? Is it slow? How much memory do you have?
https://www.metabunk.org/sitrec/?sit=agua
 
Of course, now it's NOT crashing for me.

This might be because of my rather brute-force video rendering. Works for me on my M1 Mac Mini, 16GB, but seems slow in Windows (Parallels). Takes quite a while before the video pops up.

Alternatively, it might be the rendering of the long tracks.

Does Aguadilla work for you? Is it slow? How much memory do you have?
https://www.metabunk.org/sitrec/?sit=agua
I have 16GB, Aguadilla also crashes after few secs as well
 
Back
Top