It looks like you're using an Ad Blocker.

Please white-list or disable AboveTopSecret.com in your ad-blocking tool.

Thank you.

 

Some features of ATS will be disabled while you continue to use an ad-blocker.

 

Roswell - New scans of the Ramey Memo : Can it now be enhanced/deciphered?

page: 3
157
<< 1  2    4  5  6 >>

log in

join
share:

posted on Nov, 30 2015 @ 10:47 PM
link   
I've downloaded the images of the memo and will have a look at them in Photoshop. Don't know if I will get to them tonight, but if I cannot sleep I will def give it a go!





posted on Nov, 30 2015 @ 11:58 PM
link   
I've been in the digital media field now for 18 years and what I see in the downloaded photos I can assure you is indecipherable. I used every tool I have and I'm afraid that there is nothing hidden in the image, no filter or image manipulation that can be done to it to figure out what the paper says. What we have in the photo are pixels that map a fuzzy traditional photograph image, no matter what you do to the pixels, the fuzzy image will remain. That's just my opinion, I'd love to to have someone prove me wrong.



posted on Dec, 1 2015 @ 03:03 AM
link   



posted on Dec, 1 2015 @ 04:43 AM
link   

originally posted by: WP4YT
gizmodo.com...


Thanks. I've tried SmartDeblur without any joy on this particular photo (unlike the success we had with the Roswell Slides placard). Of course, others may have more success trying different settings with the same software or (more likely in my inexpert opinion) some other package.
edit on 1-12-2015 by IsaacKoi because: (no reason given)



posted on Dec, 1 2015 @ 05:06 AM
link   
a reply to: IsaacKoi
Need to post so I don't loose this thread. I have seen on one of the tv shows where someone had enhanced the picture and interpret what was said. I would really like to see what the geniuses here can do.





posted on Dec, 1 2015 @ 05:10 AM
link   

originally posted by: clt1994
I have one question for you, with enough funding and todays technology could this image in theory be processed fine enough to make at least 90% of the words legible?


The honest answer is simple - I don't know.

I doubt that 90% of the words can be made legible, but I think it is possible that some enhancement can be achieved.

Basically, I think it is worth a bit of playing around with different software packages and filters. I may bother some experts by email as well.

But I'm not suggesting any efforts to raise funding.



posted on Dec, 1 2015 @ 05:30 AM
link   
a reply to: IsaacKoi

Your dedication to the cause is outstanding Isaac

I think this is the way to go:



I may bother some experts by email as well.



posted on Dec, 1 2015 @ 07:26 AM
link   
This is great stuff Isaac!

OK, here's what I'm seeing so far... (Not necessarily very accurate!)

_ AND THE FINDING (or VENTING) OF THE _ WAS CONNECTED TO THE
_BACK AT EYES, MOUTH? _
_ON THE "SLOT" (or CLOTH). SHOT USED NEW TAK_
_
_ THEY SHOW ONLY _ BY WEATHER BALLOON(S)_ _
_HAS LAME_

But I need better software and skills!! to get more.
BigG



posted on Dec, 1 2015 @ 07:31 AM
link   

originally posted by: IsaacKoi

Section E : Possible enhancements



For example, given the fact that the paper is folded and clearly is not straight, I wonder if a puppet warp mesh could be used to straighten some of the lines – although I think the main problem is seeking to improve the apparent resolution of the text rather than its orientation.


Took this one: MS 21st nAnd5more_fused Color Range.psd

Did this to it:


Comparison to explain:


This is ...

Forensics: Mapping a quadrangular region to a rectangular one

VFX: Projecting a 2D texture onto a 3D rectangular plane

Normal people speak: I really lazily took the bit I wanted and placed the information in a way so it was easier to look at

It's a really common procedure most often done with license plates. Fundamentals of Digital Image Processing referenced it in 1989. That's not actually the first reference of it.

You could try and average the images together combining this technique with well ... averaging. This would make the signal stronger, but if you only have one signal then you're just going to get one really really solid pile of gibberish instead of a better signal.

The ones I looked at all seemed to be of a single image. Maybe I'm mistaken.

Hope it helps.

P.S - if anyone actually wants a less lazily uncompressed perspective corrected version from one of the tif files, can do that I suppose.

Nothing leaps out at me that would improve the situation for the time being without the original camera and even then. Automated character recognition barely recognizes that text might be there. Maybe some wizard comes.

Can provide books / references as per usual. There are far more practical ones on the subject (i.e. ... not journal articles).

Kthxbai.
edit on 1-12-2015 by Pinke because: P.S



posted on Dec, 1 2015 @ 07:36 AM
link   
This is what I've managed to tweak and read so far:



Full size: files.abovetopsecret.com...



posted on Dec, 1 2015 @ 08:15 AM
link   

originally posted by: IsaacKoi

originally posted by: WP4YT
gizmodo.com...


Thanks. I've tried SmartDeblur without any joy on this particular photo (unlike the success we had with the Roswell Slides placard). Of course, others may have more success trying different settings with the same software or (more likely in my inexpert opinion) some other package.


Your source is not blurry but low resolution. No software will be able to fix that.

Cryptanalysis might have a better chance. www3.nd.edu...



posted on Dec, 1 2015 @ 09:24 AM
link   
a reply to: moebius
Agree with Moebius.

Things like Lucis (adaptive histogram equalization) and the like won't particularly do much either, will just push numbers around. There are some papers around on text blob recognition but they're mostly aimed at recognizing where text is during video so a mobile phone can take a higher resolution photo to analyse etc ...

Someone has been working on a similar approach I guess: www.roswellproof.com...

This is why the original camera would help. You would actually have something to compare to even if its a long shot. Could attempt to virtualize the camera but the whole film processing thing would be an issue.



posted on Dec, 1 2015 @ 09:37 AM
link   
a reply to: Pinke

hello, this is my first post just to put this link here
scraping.pro...

these are tools for solving captcha, and some you can even submit your own images too. this would be a great way to decipher what is written on that paper



posted on Dec, 1 2015 @ 10:16 AM
link   
Hello Isaac and great thread as usual!

I was asked as well to try to enhance this Ramey memo using our IPACO software. Unfortunately, the deblurring module is still under development and its kind of painful/complicated (In particular it needs lots of tests) to determine the best algorithm that could give the best results.

Anyway, be sure that as soon as the first achieved version will be ok, I'll try it on the memo and post the results here, even if I do not put too much faith in the final result (see the reason why below).

The Super-resolution imaging technique is interesting but I'm afraid that the results of its use of the memo will be disappointing.

Indeed, the difference between the enhanced examples on the Super-resolution page, the "Roswell slides placard" and the Ramey memo is that the text wasn't preserved the same way. The creating process of these documents wasn't the same at all, and this alone determine the quality of the final documents that we can see now.

At the contrary of the Roswell Slide placard and the Super-resolution example, the text on the memo is considerably damaged at the point that lots of letters are completely missing and some others are absolutely unreadable, whatever enhancement you use. And when there is zero information (in terms of pixels features), you can use whatever software you want, zero information will still be zero information. So what's left at best is guesses...

Anyway, like you rightly said, it worth a try.



ETA: when I said guesses, it includes of course one of the main Rudiak method which was to reconstruct words from (the best readable) known letters.
edit on 1-12-2015 by elevenaugust because: (no reason given)



posted on Dec, 1 2015 @ 11:21 AM
link   
Dropbox link is down.
When it's back up I'm going to have a play.



posted on Dec, 1 2015 @ 02:01 PM
link   
a reply to: IsaacKoi

If resolution can't be improved you don't need it. It is now a Cipher. You can treat the blurred letters as images, and build a glossary.

Start with a basic 3 letter word, most likely "the" or "and" and save each "fuzzy" letter (image ) and match it with other likely candidates. Concentrating on letters is a lost cause. These are now blurry images.

But the good news is they are all uniform, from the same typewriter, so they can be matched...roughly.

If you can crop and cut each "letter" it's possible



posted on Dec, 1 2015 @ 02:15 PM
link   

originally posted by: Taggart
Dropbox link is down.
When it's back up I'm going to have a play.


The original link is temporarily suspended by Dropbox to a large volume of downloads. (This is the first time I've encountered a limit on Dropbox - but I don't often share 27GB archives...)

Here is a secondary link to the same files:
preview.tinyurl.com...
edit on 1-12-2015 by IsaacKoi because: (no reason given)



posted on Dec, 1 2015 @ 02:43 PM
link   
Hello Isaac,

I saw this on Kevin Randle's blog and had a feeling you would be putting something together for us on ATS.

I am sure everyone is in agreement that making the picture publicly available is the best way of attempting to solve this mystery. Even if the resolution of the original pictures proves to be too low to decipher anything then at least the work people put in will all be out in the public domain for everyone to see.

I trust Jaime Maussan is not involved in this project?


Kind Regards MM



posted on Dec, 1 2015 @ 03:01 PM
link   
Well, the new scans are interesting because at first glance they make one thing simultaneously more clear and more problematic. The word that so many people think is "DISC" reads more like "CISZ" (or "CIAS" -- possibly relating to the newly-formed CIA). Anyway, I thought it was interesting that the closer you look, the more the "DISC" vanishes, literally and symbolically.



posted on Dec, 1 2015 @ 03:06 PM
link   
a reply to: mirageman


The solution will have to come from a combination of image enhancement and contextual analysis software. So far I'm only seeing the image enhancement used to its apparent limits.




top topics



 
157
<< 1  2    4  5  6 >>

log in

join