help with Eulerian Video Magnifacation Please

OK I received a personal message from one of our members and to be quite frank some of this stuff is over my head and well above my knowledge base... So.... With some of the incredibly knowledgeable members we have, I thought I would post and see what turns up..

You are my favorite thread producer for 9/11 and related
topics. I can not post anything about this for certain reasons
but if one were to analyze the available good stable HD video
of the towers smoldering with the open source Eulerian video
enhancement software many things we've never seen should become apparent. Same with Mandalay.

So maybe you know someone who would be interested?
Certainly not asking you to do it, just drives me nuts no one has done it, we just need the idea out there.
Check out Eulerian Video Magnifacation.

Hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm..............................




Well isnt that something. For WinXP I have a demo version of police / FBI photo enhancement software that I used to do post stuff like with the pentagon CCTV video for example, that was unlike any other images I ever seen without it. Havent been able to use it forever, but it appears these guys have truly figured out that kind of logic with video.

Requires MATLAB Runtime(1.3GB):
www.mathworks.com...
NOTE: I installed 8.3 but the executable still bitching it cant find matlab 8.0...

A rabbithole of work to spiral out of this thing. So many videos to consider like with 9/11. You'll want to be using the most original renditions possible (raw uncorrupted video files). Problem is, unless you've been archiving and really paying attention over the years, by the time you swipe such an old video off youtube its been 'stepped on' many times over (re-rendered again and again and again). All the 'no WTC planes' videos I've seen are based on this, although they tend to deliberately exploit the concept. I'm imagining their tribe will be having a blast tweeking this "Eulerian Effect" to really squeeze out BS results to push their COINTELPRO efforts. And perhaps they have been as this executable has been available since 2013. I was tuned out when this hit so I missed the memo until now. I have a friend well connected to the 9/11 scene, very software technically oriented, has the most complete raw 911 footage archive about possible, I'll hit him up later see if he's heard abut this thing and seen any results...

In other news, I'm sure Google and friends have been busy integrating this into the AI gods, and all sorts of behavior monitoring sorts of evil go go gagdetry already by now. We're f&**)ed.

a reply to: 727Sky

Good Lord! The ramifications of that work are boundless across all fields of mechanics and flesh and blood.

When I tried this, my difficulty was in obtaining an HD and stable enough vid of the buildings prior to and during the collapse. Anyone have one?...
Also, the program has a few parameters to tweak: I need a bit of practice at that .

For the Eulerian program to work best, an absolutely still camera is essential; such as cctv. And very high definition also.

a reply to: ecapsretuo

you would just have to stabilize the video with post processing first, then run it through Eulerian

a reply to: NobodiesNormal

Seems that would be adding long term defects.

You want RAW video to achieve natural results is my instincts.

Like with calculators when doing higher math, "Garbage In Garbage Out" the more calculations stemming from a wrong input the more dramatically incorrect the answer is the further you push on with it.

There is a TED talk on this subject ...Or I mean the subject comes up in a Ted talk . There is some incredible things they can do using light ...

a reply to: IgnoranceIsntBlisss

i thought you were more familiar with video editing tech then that....

stabilizing the video would center the frame on the object of your choosing.

eulerian would emphasize movements occurring within the chosen object,

so no stabilizing it would not "add long term defects" rather it would lower the amount of obscuring artifacts present on the object in question. (which in the given example would be the twin towers)

its a good idea really, but yes its true you'd need pretty decent quality footage, but not super great as was implied

a reply to: 727Sky

I was under the impression that this works only with flesh and blood

a reply to: NobodiesNormal

I'm all self-taught when it comes to video work, fall asleep int he keyboard if I attempt to do coding stuff, but I do have a degree in electronics engineering math so I've seen how totally math can get in a calculator.

This topic here it's mostly instincts I have to go by...

What we're calling "video" here, is pure Math, and this new software tech is something like the most advanced computer video mathematics out there.

I havent read their papers, and I'm not a mathematician.

I have dabbled in encoding /decoding and rendering video. I've seen where bizarre artifacts that dont exist in the video preview emerge after rendering, and have hassled with trying to remove them to no avail, on occassion.

Note I mentioned the No planers and their shady games, as 'all' of their "evidence" is in the deliberate video propagation of such artifacts, which then they stampede around screaming it as proof.

So looking at this thing, it seems you'd want the cleanest 'math sample' (video clip) possible.

Meaning if you go and 'stabilize' the video sample, I'd anticipate stabilizing the subtleties right out of it (or worse). This is a matter of tainting evidence except for when and IF you nail it just right (which you might not even be sure if you do). Keep tweeking the settings eventually you'll see what you want to see, kind of thing.

Could be wrong, and I'm stunned by what they've accomplished, but it has to have limits.

Need original video, not what is uploaded to Youtube.
I can do the stabilization and apply this process.

a reply to: Deaf Alien

How is it original if you stabilize it?

Good luck getting 9/11 clips that arent only in youtube. Some of them were probably first uploaded and then downloaded back on Google Video, and then posted to youtube later (if not rendered in and back out of compilations / documentaries since).

Then there's the lacking tech of the era, even with the big fancy news cameras.

But for news clips the closest thing to raw you can go and direct download would be the archives of newscast from that day, at Archive.org. I suspect they have an added step of conversion, or rather that all of them do, but there is a chance that the largest file they offer hasnt been stepped on.

In 2006, my friend and I caught wind that they had in their internal systems the pre-release files for the set, and we met by we were both in there trying to download as many as we could before they shut it down and then I went into a forum trying to connect to people doing the same. I had got several, he had got a lot more, in the like less than a day window we had to run at it. They're about 1024MB each file if I remember correctly (400 or so files). He ended up finding damn near if not all of the complete set. This was the DC newscast footage archive, they might have added more. The DC set was mostly complete for a day or so before and several days after. There were only few slices of time missing from the whole thing (I have threads about it if they're still up in here). This was back in late 2006. Friend he still has them and a whole lot more, but its a full hard disk long distance so I havent hassled trying to get it from him. They should have videos up covering all the files we after, and they look nice the raw files, but it still isnt original video because the news took what they had and ran it thru their processing equipment all those details during the broadcasts, and then the stock footage archive house recorded it and then 'whatever' into their format they store it as. Then we dont know what else happened to the 'math' by the time they uploaded everything into archive.org by the time we caught on. Then most people can only get 'these as close to raw as you could hope for' now ran through archive.org's automated systems.

Here's a quick search result a guy talking about what happens when you upload good raw video to youtube:


"I'm often asked why GoPro video that looks gorgeous when played back from original files on your computer, looks like crap when played back on YouTube. It's YouTube's over-compression - period. In this short video I explain my strategy for taking high frame rate HD video from a GoPro Hero3+, and converting it so it will look at least decent on YouTube. "

Now imagine clips that were worked on in an editor, rendered, then uploaded you youtube, downloaded (possibly converted during this step), and now loaded into this MATLAB tool.

I fear the 'hidden elements' that would 'appear' would be the artifacts themselves, from all of that stepping on.

originally posted by: NobodiesNormal

stabilizing the video would center the frame on the object of your choosing.

You still have to tilt/skew the frames to line them up. How much mess that would or would not trigger further down the line I'm not sure.

Then you still have to render it.

I suppose if you rendered it in the most RAW method possible, which wouldn't be hard for one test clip but damn hours worth of them would probably grow unfeasible, then this might work.

But this method of 'stabilization', you'd have to match the frames with utmost precision, or I'd expect that the 'hidden details' that would pop out in MATLAB would be the tampering er the flaws in the frame matching.


Another headache would be a lot of clips you might try to throw at their tool wouldn't even be compatible without running the clip thru a convertor. I know damn Vegas after all these years it still has problems opening clips. I have project folders from years ago the source clips I used in older versions of Vegas I have to convert them for them to even open in new Vegas. I even have videos that were rendered in older Vegas that now wont open in the new one! Been hassling with this conversion stuff like crazy lately...


All of this isnt to say that the available version of the tool isn't capable of showing us some eye poppers from the world of clips out there. But I suspect with this first generation of the tech it'll take a lot of different people trying their hand at a lot of different clips......

a reply to: IgnoranceIsntBlisss

I will take a look later. I've done the stabilization and the process before.
Hope the result will be interesting.

Heres a compilation of the 2nd plane hitting. The videos are fairly stable.

At the 3:37 mark, you can see a helicopter on the left of the screen. I bet he got the best video of the 2nd event.

a reply to: Macenroe82

i believe the idea was to run video of the collapse, not the planes hitting,

stabilize a video of the collapse, run it through eulerian and you might see evidence of demolition charges detonating within the building as it goes down. stabilizing would not remove this since this would be within the centered object not outside it....



a reply to: IgnoranceIsntBlisss


you dont need programing knowledge to run eulerian, the programing is all done they made it for you, its already a completed program, you only download it and run it like all your other video editing programs.

last i used it though it just didnt have a fancy ui, had to run it through a bat command but it was simple enough.
i also have zero programing knowledge and i managed to run eulerian with no problems im sure you could too, its weird you didnt try, i would have thought such a tool would have excited you with all the uses you can put it too, conspiracy wise...

all your ideas about why it might not work are assumption on your part. try the damn program, its free. im sure once youve tried it and come to understand how it functions you will then realize as i have, stabilizing the towers then running it through would not compromise the effort, rather it would assist it.