Earlier today Polarik commented here in the original Stretch Marks post. The original Stretch Marks post is a bit old and now on the second page, so I EMailed Polarik and asked him if he’d mind if I started a new post and he could re-post his comment there.
I haven’t heard back from him so I decided to start a new post and reply to his comment in a comment here. It’s approaching dinner time here but I should have it done in the next hour or two. Stay tuned.
k
Koyaan's Weblog 
73 comments
Comments feed for this article
August 11, 2008 at 5:18 pm
koyaan
Polarik wrote:
The reason the Kos COLB is screwed up and Michele is not is because the forger used PD’s COLB for the template.
Yes, so you’ve claimed.
You can get a copy of PD’s COLB on her daughter’s Valleehill Genealogy website. The “D” stands for “DeCosta.” Need any more hints?
Nope. I’ve already had the PD COLB for some time now.
OK, it should be patently obvious that this COLB document was scanned with the top off (or the top was flimsy) because the image shows the two dep folds in the document.
Yes.
Those folds are key because they cause the height of the border to be shorter than it really is. With the height effectively short, we no longer have the same aspect ratio as every other COLB that has been scanned from a COLB document placed flat against the glass.
That would be the assumption, yes, if you also make the assumption that printed height of every other COLB is identical.
Someone created this animated GIF, who like youself, thought that the Kos image is genuine because PD’s COLB is genuine. This, they reasoned, is why the Kos image is genuine.
Excuse me, but I have never made the argument that the Kos image imust be genuine simply because the PD COLB is genuine.
Here’ their GIF:
Yes. Seen it already.
AH…not so fast, Grasshopper.
Yes, Master Po.
Look at it carefully and what do you see??
Someone who’s seen a few too many Kung Fu reruns?
Remeber what I said about the COLB being scrunched together because of its two large folds as it sat against the scanner glass?
Yes.
I said that it was these folds that made the height of the PD COLB image to come out a lot shorter than normal.
Yes, given the qualification I pointed out previously about all COLBs having been printed with the same height.
There is only one other COLB image with the same exact aspect ratio, the same exact border location, the exact border thickness, the same placement of the text, and the same location of the masthead.
And that is a 1963 Pontiac Tempest.
My Cousin Vinny fan too?
That COLB is the Kos COLB image!!!
So you say.
Screetch! That’s the sound of your theory making a panic stop.
No, I’m afraid it’s your theory that will be making a panic stop here shortly.
The Kos COLB image looks like it was LAID FLAT AGAINST THE SCANNER GLASS.
Yes, it does.
The Kos COLB imag has only one, thin fold in it, so not only did it have less folds than the PD COLB, or any other COLB in existence, this fold had no effect on the scanning of the document.
One would think.
I’ll say it again.
If you must.
The Kos COLB image looks like it was LAID FLAT AGAINST THE SCANNER GLASS.
Got that out of your system now?
HOW DOES A COLB IMAGE THAT WAS ALLEGEDLY LAID FLAT AGAINST THE SCANNER GLASS PERFECTLY MATCH ONE THAT WAS RAISED OFF THE SCANNER GLASS?
It doesn’t. Assuming that both COLBs were originally printed with borders of the same height and width.
AND, THE REASON WHY MICHELE’S COLB DOES NOT HAVE THE SAME ASPECT RATIO AS THE KOS COLB IMAGE, IS BECAUSE MICHELE’S COLB IS TOTALLY REAL, WHILE THE KOS COLB IS TOTALLY FORGED FROM PD’S COLB.
So you claim.
That’s it, Koyaan. There are no other explanations for why all of the printed elements on the Kos COLB totally match the same, printed elements on the PD COLB.
Except for the one eensy weensy little fact that they don’t.
Unless you can also bend Space-Time, you have no other alternatives left, but the most parsimonious one — that the Kos COLB is a knock-off, a fake, a forgery, a fraudulent image.
Except for the alternative that… you are just flat out wrong.
Ok, my turn.
The fact is that the Kos border does not have the same aspect ratio as the PD border. In spite of the loose folds of the PD COLB when it was scanned which you claim would have to make its height shorter than that of the Kos border, when matched to the width of the Kos border (preserving the aspect ratio), the height of the PD border is actually taller than the Kos border.
Here is the overlay of the PD COLB over the Kos COLB:
obamapdoverlay.jpg
As you can see, the border is taller and the horizontal alignment of the text gets progressively worse as you go down below “CHILD’S NAME.” Also, the “ANY ALTERATIONS…” line is significantly wider in the PD COLB compared to the Kos COLB.
If you match for height instead of width, you end up with a PD border that’s narrower than the Kos border:
obamapdoverlay2.jpg
The only way to get the aspect ratios and the text to line up between the two is to change the aspect ratio of one image or the other. Either squeeze the PD image vertically making the height of the border shorter, or stretch the Kos image vertically making the height of the border taller.
Furthermore, and again completely contrary to your claim, the aspect ratio of the Kos border is identical to the aspect ratio of the Michele border. Also, except for the placement of the text between “CERTIFICATE NUMBER” and “OHSM…”, everything else lines up much better between the Kos image and the Michele image, and the “ANY ALTERATIONS…” line is the same width, as shown here:
obamamicheleoverlay.jpg
BTW, as you may have gathered, I have a much different take than TechDude’s.
Yes. But you seem to have something that’s not so different from Techdude; fabricating your analysis.
k
August 12, 2008 at 12:56 pm
polarik
My comments will come after I make the following HTML test here:
bold
line break
italic
August 12, 2008 at 2:35 pm
koyaan
Will they come today perchance?
k
August 12, 2008 at 3:39 pm
polarik
Yes, they will. But, what I can tell you now is that I never got your email. I certainly would have responded earlier, even though I’ve been up to my neck in alligators.
Please resent it again, if you like. but I’ll be happy to respond to your comments here.
First of all, I have to thank you for making this commentary on my work. It’s caused me to rethink my assumptions.
I thought that it would be relatively easy for others to replicate. Now I know that I need to add a caveat.
One of the problem areas for us (and probably for others) is distinguishing between the actual, paper document, and the image made from a scan of it.
For the audience’s sake, we should point out when we are talking about images, and when we are talking about the actual scanned documents that became images.
Believe it or not, we agree on a lot more points that we disagree, and if we were to be one the same page, vis-a-vis, the image/document dichotomy, I think you’d find there is very little difference.
Well, except for the diametrically-opposed positions from which we start. It’s a little more complicated than you say, “po-tay-to,” and I say, “po-tah-to.”
Basically, what I can also tell you now is that all COLBs, since 2001, were printed by Vital Records, on one networked printer, and on the same brand of printer when they upgraded. I know this because it was I who first contacted the OHSM about it.
In fact, there are a lot of “firsts” that I can claim, but I had to defer to you-know-who for the reason that I do not like to air dirty laundry.
OK. I’m almost done with my responses to your comments and queries.
I’ll post them sometime this evening.
Regards,
Polarik
August 12, 2008 at 4:35 pm
koyaan
Yes, they will. But, what I can tell you now is that I never got your email. I certainly would have responded earlier, even though I’ve been up to my neck in alligators.
Please resent it again, if you like.
Nothing to bother to resend. It simply said “Would you mind if I start a new thread to discuss this? Stretch Marks is pretty old and now on page 2.”
First of all, I have to thank you for making this commentary on my work. It’s caused me to rethink my assumptions.
You’re welcome. Though why do you make assumptions?
I thought that it would be relatively easy for others to replicate. Now I know that I need to add a caveat.
I could have easily replicated your results, but I’d have had to change the aspect ratio of one or the other image to do so. And you had said the aspect ratios of the two borders were the same.
One of the problem areas for us (and probably for others) is distinguishing between the actual, paper document, and the image made from a scan of it.
But neither you nor I have the actual, paper document. So all we can speak to are the images.
Believe it or not, we agree on a lot more points that we disagree, and if we were to be one the same page, vis-a-vis, the image/document dichotomy, I think you’d find there is very little difference.
Perhaps. Though as I said, neither of us have the actual, paper document so I don’t see how either of us can say anything relevant with regard to it.
Basically, what I can also tell you now is that all COLBs, since 2001, were printed by Vital Records, on one networked printer, and on the same brand of printer when they upgraded. I know this because it was I who first contacted the OHSM about it.
Ok. Perhaps that’s true, perhaps it’s not. But assuming that it is true, I’m not sure what you feel the relevance is.
I’ll post them sometime this evening.
Okie doke. I’ll be here.
k
August 12, 2008 at 5:37 pm
rayinaus
In 2007 Hawaii used a border that is identical to the Obama COLB border.
August 12, 2008 at 6:08 pm
koyaan
In 2007 Hawaii used a border that is identical to the Obama COLB border.
Great. Let’s see one.
k
August 12, 2008 at 7:45 pm
polarik
Before I begin, let me assure you that it will be a very professional exchange from me. I have dispensed with making any raunchy retorts, flippant flames, sarcastic statements, and denigrating denouements.
Not bad, eh?Let me also add that when I discuss comparing two COLB images, I’m talking about orienting both to be perfectly parallel to each other, and cropping both of them to the very edges of their respective borders. This is the only honest way to compare COLB images
OK, with that out of the way, let’s proceed. As per your format, I will put your comments in bold, and mine in plain type.
The reason the Kos COLB is screwed up and Michele is not is because the forger used PD’s COLB for the template.
Yes, so you’ve claimed.
True. When I mentioned the PD COLB, your response was
Nope. I’ve already had the PD COLB for some time now.
Also true, and which lends credence to the fact that the PD COLB has been posted online for the past two years. The other important fact about the PD COLB (which will be explored below) is that, before the Kos image was ever made, the PD COLB was the only COLB to be posted on the Internet.
OK, for the audience sake, we should acknowledge that we are talking mostly about images, and less so than the actual scanned documents that made them.
I continue with this statement:
OK, I said that it should be patently obvious that this COLB document was scanned with the top off (or the top was flimsy) because the image shows the two deep folds in the document. Here I am talking about the physical condition and position of the paper document on the scanner usede to produce the image.
Well, we both agree on this outcome.
Then I said that these folds are key because they cause the height of the border to be shorter than it really is. With the height effectively short, we no longer have the same aspect ratio as every other COLB that has been scanned from a COLB document placed flat against the glass.
Your reply:
That would be the assumption, yes, if you also make the assumption that printed height of every other COLB is identical.
See? So far we are copacetic when we are talking about the actual printed paper COLBs and not their images. But, before I respond, we need to agree on an operational definition of “identical” and also agree to limit our discussion, just for this part, to aspect ratios rather than focusing on one of the two dimensions. So, I do hold the assumption that the aspect ratio of every other COLB is “identical,” providing that we are talking about all COLBs printed from 2001 until now. The reason for the qualification is because the COLB form was revised in November 2001, and I have never seen a pre-2001 COLB.
First order of business is to formulate an operational definition of “identical.” I’m a statistician by trade, and there are no real absolutes in statistics. I work with things that vary, like tolerance, margin of error, and so on, to acount for random variation. So when you say “identical,” I need to know if you define it as I do. Do you mean that the printed height and width should have “zero tolerance,” as in “no deviation from a benchmark size,” or do you mean then they should not vary beyond what is acceptable and expected? I would select the second definition, and until you tell me otherwise, I’ll take it that you also mean the latter.
So, let’s agree that the aspect ratio — the ratio of height to width — is a constant within a tolerance of +/- 3 to 5 pixels out of an image measuring 2550 x 3300 pixels, or an uncropped image copy of an 8 1/2 x 11 inch COLB document. I ahould also tell the audience that the top 3″ of the paper COLB is detachable, leaving an 8 1/2″ square COLB.
I think that I can demonstrate to you that the aspect ratio has not changed from 2002 to 2008, as well as reinforcing our “joint view” that the folds in the document change the aspect ratio.
p>OK. I’d like for you to look at the following two COLBs that belong to Jason Tomoyasu and Jeremy Smith. They are from the same range as PD’s COLB (2001-03). In terms of aspect ratios, please tell me which one of these would be closer to the Kos image and which one would be closer to Michele’s image (both images are show at their original image size). Also tell me how you made the choices (the answers are at the end).
Smith:
Tomoyasu:
So, yes, I do acknowledge, as a matter of fact, that the aspect ratios of all the printed COLB documents from 2002 to 2003 and from 2006 to 2008 are the same. I will further assume that the aspect ratios for all other COLBs, from 2001 until 2008, are also the same. When I say, “other,” I am referencing those COLBs that I have not personally seen.
My essential argument, to recap, is that, with its height effectively shortened as a result of the folds, the PD COLB no longer has the same aspect ratio as every other COLB that has been scanned from a COLB document placed flat against the glass. And you agreed, providing that…
That would be the assumption, yes, if you also make the assumption that printed height of every other COLB is identical.
Which I have noted. Not just the printed height, but also the printed width, and taken together they form the same aspect ratio. [Note to audience: we are talking about the printed paper document, and not the image made from the scan of it.]
Well, nothing lasts forever, and so our short “honeymoon period” ends when I made the observation that “you thought that the Kos image is genuine because PD’s COLB is genuine.”
Boom! You fired back:
Excuse me, but I have never made the argument that the Kos image must be genuine simply because the PD COLB is genuine.
You’re excused. Now, let’s have a reality check.
In response to my comment that the folds are what caused the height of the border to be shorter than it really is, you provided the same observation in your July 23rd rebuttal (albeit not as clearly stated as your other comments.)[My explanatory comments are in brackets]:
“The only other known COLB I’m aware of which still retains enough of the background pattern to be useful is the DeCosta COLB. And since the change in aspect ratio is pretty much entirely horizontal, the DeCosta COLB will do just fine (it was apparently scanned without putting the lid down on it so the folds are still very slightly folded which screws up the vertical image).”
What you’re saying here is that, on the DeCosta image, the change in aspect ratio [from what other COLBs have] resulted from a drop in the position of the top horizontal border because the folds screwed up the length of the vertical borders that support it.
Or, what I originally said to you. Thanks for agreeing with me on the change in the aspect ratio of the image (and not the document). Where we differ is in the significance of that change. Which is why I wish that you would have used “aspect ratio” instead of “printed height” in this comment of yours:
“I noted before that the Obama image, save for its funky borders, was quite consistent with all the previous COLBs whose provenance isn’t really questionable. ”
Since the DeCosta COLB was the ONLY previous COLB available for viewing when the Kos image was posted, your statement confirms what I said about the Kos COLB being genuine becauses it fully matches the PD COLB — save for the big word you used, namely “provenance.”
What does provenance mean?
prov·e·nance (pro(v’?-n?ns, -näns’) n.
1. Place of origin; derivation.
2.
1. The history of the ownership of an object, especially when documented or authenticated. Used of artworks, antiques, and books.
2. The records or documents authenticating such an object or the history of its ownership.
So, to rephrase your statement using one of the synonyms of “provenance,” you are, in effect, saying that:
“I noted before that the Obama image, save for its funky borders, was quite consistent with all the previous GENUINE COLBs.
Unless you wish to contest that GENUINE is not a synonym of AUTHENTIC, or PROVENANCE”, you’ve been duly corrected.
Not too bad, so far. What follows after this exchange are several paragraphs of my comments followed by your acknowledgments, and a gag reference to a Pontiac Tempest (Funny you should mention it, but I used to drive a ’64 Goat tripower.) I need to mention one comments here for continuity purposes. Afterwards I’ll skip ahead past any comments in agreement with mine, as well as any pithy remarks that are not germane to the subject.
What I have been saying throughout is that there is only one other COLB image [besides the PD COLB image] with the same exact aspect ratio, the same exact border location, the exact border thickness, the same placement of the text, and the same location of the masthead. I said, “Kos,” you said, “Pontiac Tempest.” (OK, I’ll allow you one pithy patter).
I need to point out again that we are talking about images, and not the actual documents that were scanned. We should agree, contingent on the assumption being true, that all printed COLBs since 11/2001 have the same shape, with consistent heights, widths, and aspect ratios. Now, we have not talked about the thickness of the borders, but let’s assume, for the moment, that they also remain the same on the printed document.
The point of contention centers around the image of the Kos COLB (not the printed original if ever there was one) versus the image of the PD COLB (not the printed original which we know did exist), and the observable match between them. What makes this image match so special is that they are the only two COLB images, out of all the COLB images posted on the Net, that have the same, exact, idiosyncratic aspect ratios. Plus, they are the only two COLB images that match each other on every printed component of the images while being unlike the other COLB images posted on the Net.
Hopefully, you and the audience will be clear on the distinction I made regarding the image match, and not on any match between them as paper documents. But, lets assume, for the moment, that both documents would match each other if both existed in the real world (and not just in the virtual world of images), why would a document that had its height shortened by folds such that it produced a scanned image with a distorted aspect ratio, exactly match a document that had no fold distortions such that it should have produced a scanned image with a totally normal aspect ratio?
My original question to you, which reflects the same conundrum, was “How does a COLB image that was allegedly laid flat against the scanner glass perfectly match one [another COLB image] that was raised off the scanner glass?
Now, I thought that I had made it clear that I was talking about images and not paper documents. However, I think you mixed them up when you said:
It doesn’t. assuming that both COLBS were originally printed with borders of the same height and width.
Koyann, I was discussing image matches and not matches between any printed paper documents. If the height, width, and aspect ratios of the printed paper documents match in the real world — which is what you said about printed paper documents, then why would they still match in the virtual world if one of the images was made from a deeply-folded document that was not scanned flat and the other image was made from a lightly folded document that was scanned flat?
To put it another way, one of the two images got its distorted aspect ratio because, in your own words, “it was apparently scanned without putting the lid down on it so the folds are still very slightly folded which screws up the vertical image).” So, I ask you, “What is the reason why the vertical part of the Kos image, which was not “screwed up,” match the vertical part of the DeCosta image, which was definitely screwed up?”
The bottom line is that the aspect ratio of the Michele image did not “get so messed up.” It has the same aspect ratio as every other COLB image that was made from a perfectly flat, scanned document.
There’s nothing wrong with Michele’s COLB image. The same cannot be said for the Kos COLB image. Yet, you didn’t see the conundrum:
“Well, the image I’d been working with was the image I pulled off Polarik’s photbucket blog. Maybe he boogered it up somehow.”
I’m sorry to say that the only thing that got “boogered up” was your interpretation of images — they were based solely on your understanding of real paper documents, when they should have been based on an understanding of how images can reflect, or not reflect, reality. The DeCosta COLB image reflects the reality of how folds in paper distort the images made from it. There are no reasons why the Kos COLB image should be similarly distorted.
“Whereas the images used to prove the Obama image “a horrible forgery” had aspects to them which were consistent with each other, but not consistent with any of the other known COLBs.”
Well, for starters, “Horrible forgery” is not part of my vocabulary here. When you said, “any of the other known COLBs,” you didn’t elaborate, but let’s say that one of them is the PD COLB — which you “already had for some time now.” Recall that the PD COLB image was the only known COLB image before the Kos COLB image was, pardon my French, “manufactured.”
“And now apparently there’s yet another aspect about them which is consistent among them, but not consistent with any of the other known COLBs. A mess up aspect ratio.”
Did they “mess up” Massa? Sorry..couldn’t resist it, Why would you call three, perfectly good-looking COLB images, “messed up” when the really “messed up” image was the authentic, PD COLB, whose “vertical image was [also] screwed up?”
Now your turn.
The fact is that the Kos border does not have the same aspect ratio as the PD border.
Would you care to bet on that, sir? Remember, we are talking about IMAGES, not PAPER DOCUMENTS. If you resize one image by making its width equal to the width of the other image, WHILE KEEPING THE ASPECT RATIO CONSTANT, the heights match. Alternatively, If you resize one image by making its height equal to the height of the other image, WHILE KEEPING THE ASPECT RATIO CONSTANT, the widths match. They do for me, and they should for anyone else in the real world. Remember, we are talking about IMAGES, not PAPER DOCUMENTS. These two images are not sized the same, but they have the same aspect ratio. They have the same height-to-width relatonship. So, since they have the same aspect ratio, by keeping that constant, you can make one image to be the exact size of the other by CHANGING ONLY ONE IMAGE DIMENSION.
Let me go back to what I said in the beginning. When I discussed comparing two COLB images above, I was talking about orienting both of them to be perfectly parallel to each other, and then cropping both of them to the very edges of their respective borders. This is the only honest way to compare COLB images. Unfortunately, as shown by the images you supplied, you have been talking about matching up the uncropped images. And, I agree that they will not match up.
I thought that you would have known that from my animated GIF’s, which are cropped right down to the last pixels lying at the outer edges of the COLB borders.
At this point, I really cannot comment any further on what you said because we are, indeed, comparing apples to apples. To drive that point home, let’s discuss the COLB image comparisons I asked you to make earlier. Now, like I said, to do this properly, you need to zoom on in, down to the pixel level — or as close as possible — and make sure that both are perfectly aligned in portrait configuration before you crop each COLB image to the outer edges of their borders.
The COLB that has the deep folds in it, will be a closer match to the PD COLB and the Kos COLB, when you set one COLB image dimension to that of the cther COLB image dimension, while keeping the aspect ratio constant.
Conversely, The COLB that has the minimal folds in it, will be a closer match to MICHELE’s COLB image and the other post-2005 COLB images, when you set one COLB image dimension to that of the cther COLB image dimension, while keeping the aspect ratio constant.
To conclude, the only way to be sure that you have two comparable images, is to scan both of them perfectly flat. To the extent that their folds prevent them from being scanned perfectly flat, the image that results will have shorter vertical borders than normal, lower horizontal borders than normal, and an aspect ratio that is abnormal.
I’m out of gas, and out of time. If you are still having problems wrapping what I said around your head, please ask your loyal readers to weigh in.
Thanks for the opportunity to clarify things here.
August 12, 2008 at 7:58 pm
polarik
I meant to say, “that at this point, we are talking about apples and oranges.”
Sorry….I’m tired
August 12, 2008 at 9:32 pm
polarik
Koyaan, to replicate my comparisons, let me give you the cropped-to-the-border image sizes of the three original COLB images. {NOTE: all images are set to 300DPI]
Kos: 2373 x 2312 pixels
PD: 845 x 823 pixels
Michele: 2392 x 2327 pixels
Also, please read through ALL of my post above rather than responding piece by piece. I spent a lot of hours in writing it, and I hope you appreciate that.
RP
August 12, 2008 at 9:35 pm
polarik
A final note before I hit the sheets:
Please remember to deskew them counter-clockwise, BEFORE cropping them, so that all three are arranged in a perfectly parallel, portrait orientation.
Hasta manana.
August 12, 2008 at 9:39 pm
polarik
Oh, and BTW, I fabricate nothing. What you see is what you get
August 12, 2008 at 9:56 pm
koyaan
Let me also add that when I discuss comparing two COLB images, I’m talking about orienting both to be perfectly parallel to each other, and cropping both of them to the very edges of their respective borders. This is the only honest way to compare COLB images
I’m sorry, but this is simply nonsense. There’s absolutely no reason whatsoever to crop anything.
The only honest way to compare two images, is to compare the two images. Period. Not by creating two new images by cropping them from the original images.
If you can’t make your case by comparing the two original images, then you simply haven’t got a case. And as long as you insist they must be cropped, then I can’t see this discussion continuing any further.
k
August 12, 2008 at 11:40 pm
rayinaus
I couldn’t bear to read through all the above text while knowing that it is all a complete waste of time because it is completely irrelevant and wrong – because, as I keep saying, Hawaii had a different border in 2007, however I have a few comments for when you finally get a grip and start comparing apples with apples, i.e. a 2007 certificate with wide cross-hatching with another 2007 certificate with wide cross-hatching.</b
(1) It’s a waste of time to crop the scanned images to isolate the printed images because the printed images can be lined up by simply moving them on a larger canvas.
(2) It’s obviously a good idea to align the two printed images at the left side, and to rotate one of them if necessary.
If I thought it would do any good I’d put my 2007 Marriage certification on the net (with the same wide cross-hatched border) for everyone to play with, but I already know what would happen if I did — nothing. Everyone would kep arguing nonsensically about nothing.
Play on!
Ray
August 13, 2008 at 12:01 am
johnque
Let’s get some facts on the table:
“the PD COLB was the only COLB to be posted on the Internet.”
Source? How do you know this to be true?
“OK, I said that it should be patently obvious that this COLB document was scanned with the top off (or the top was flimsy) because the image shows the two deep folds in the document.”
I completely disagree. If this was the case, the border pattern at and around the fold locations would show signs of being inconsistent with the “flat” portions. In fact, taking the Decosta COLB into Photoshop and zooming the border to 1600% with a 1 x 1 pixel grid turned on shows little or no evidence of such distortion. You’ve assumed that the Decosta COLB was scanned without flattening it but the borders say otherwise.
I would also expect to see some shadowing at the locations where the paper wasn’t flat on the glass surface of the scanner. Here’s an example where this occurs and the text isn’t even distorted:
a more distinct shadow and distortion of text:
The absence of a shadow doesn’t prove the Decosta COLB was flat when it was scanned but the evidence shows little or no evidence that it wasn’t.
As for the Michele COLB, the image was given to you by TexasDarlin who has zero credibility with me. I wouldn’t count on it being any more or less legitimate than any other image out there.
August 13, 2008 at 12:17 am
johnque
One other point that discounts the “shrinking document” theory. If the Decosta document was not flat when it was scanned, the text would be shifted upwards on the document compared to a document scanned flat due to the distortion of the document surface. But if you compare the Decosta to the Michele COLB, you can see that in fact the opposite is true. The text on the Decosta scan is consistently lower on the page, which is the opposite of what would happen if the document was not flat when it was scanned.
August 13, 2008 at 5:13 am
rayinaus
I agree with what johnque said in the above two posts, but I think some consideration should be given to the fact that scanners are not like conventional cameras which have very small depths of field for close-up work.
Scanners can capture images that would normally be slightly out of focus (or even completely out of focus) in a process camera image.
August 13, 2008 at 6:24 am
johnque
Fair point ray. One other thing that occurred to me thinking about the scanning process. If this document was scanned with the top off or the document not pressed down firmly on the document, one would expect to see quite a bit of darkness around the edge of the document from outside light interfering with the scanning process. This effect is easy to replicate by scanning a document with the scanner top off or up (depending on the model you are using). While the Decosta COLB has been fairly closely cropped, the bottom edge was not. But there’s no sign of light leakage at the sides where you would expect to see that happen. Again, there’s ways that this could be prevented even if the document wasn’t flat (document fit exactly to scanner edge, a larger piece of paper placed over original when it was scanned, etc.) But in my eyes, this is one more sign that the Decosta COLB was flat when it was scanned.
August 13, 2008 at 7:58 am
rayinaus
johnque wrote:
Fair point ray. One other thing that occurred to me thinking about the scanning process. If this document was scanned with the top off or the document not pressed down firmly on the document, one would expect to see quite a bit of darkness around the edge of the document from outside light interfering with the scanning process.
Yes of course we would.
This effect is easy to replicate by scanning a document with the scanner top off or up (depending on the model you are using). While the Decosta COLB has been fairly closely cropped, the bottom edge was not. But there’s no sign of light leakage at the sides where you would expect to see that happen.
I was going to mention that several times along the way, but the whole concept of using the DeCosta COLB for a fabrication seemed so silly that I didn’t bother responding to most of the points.
Again, there’s ways that this could be prevented even if the document wasn’t flat (document fit exactly to scanner edge, a larger piece of paper placed over original when it was scanned, etc.) But in my eyes, this is one more sign that the Decosta COLB was flat when it was scanned.
Something white had to be behind the certificate – and what’s the handiest white thing available? The bloody lid of the scanner of course!
——
Nearly everything about Obama looks very straight forward to me, even though he has had an unusual life, and if we look at every bit of evidence that has been supplied (much of it before his run for the Presidency) without trying to read more into it or distort it, it nearly all fits together. The same applies to the scanning of the DeCosta — no one should be inventing NEW ways to scan a document JUST because it is connected to an Obama smear.
An interesting bit of trivia regarding scanners and depth of field: In my work when I first got a scanner before transparency holders were on the market, I occasionally had film negatives and positives up to about 8 x 10 inches which I scanned on a paper scanner by laying them between sheets of glass stacked on books with an overhead light – sometimes a Kodak beehive 100v point-source light, although the first time I simply used a strong torch from about 2 feet above.
Another quick tip for copying newsprint etc where you don’t want the back to show through – use black backing paper and expose a bit more (if the scanner allows it). Photographing artwork of any type is NOT about the camera or scanner “seeing” the image – but “seeing” the background colour – usually white – so if it’s not pure white we need to increase the exposure to compensate for that “reflection density” difference.
Ray
August 13, 2008 at 8:31 am
polarik
Koyaan. When I said, “adjust my assumptions,” I was specifically referring to you and others who have difficulty in conceptualizing objects spacially, suich as how the printed document might look when placed on the scanner glass.
I was also referring to the twisting of my words, and denying things that they said they didn’t.
Koyaan: you mentioned “printed document,” more than once in your comments, so you cannot claim that you’re only tslking about images.
You cannot have it both ways, Koyaan. You can’t bitch about others not giving you the instructions on how to replicate their work, and then refuse to follow them.
Koyann, what possible justification do you have for not following my directions? How exactly do you make a fair comparison of three different images, all sized differently and cropped differently?
If you do not wish to follow my directions exactly, as you refused to do with mine last night, then that means you are afraid of proving me right.
So, do it my way. What have you got to lose?
What I propose is an experiment with one control and two experimental donditions:
To do this, you need to make some scans of one document under three document conditions: (1) with deep folds, (2) with narrow folds, and (3) with no folds.
Then, you will have two scanner conditions, scans with cover on, and scans with cover off.
In all, you will have six experimental conditions. As per Koyaan’s specious complaint about my croppings, all images will be scanned uncropped.
For the test documents, print out three copies of Michele’s COLB onto 8 1/2″ x 11″ sheets of paper.
Take two of them and tri-fold them so as to get them ready to be placed inside a #10 envelope. Leave the third one unfolded. Take one of the folded sheets, and trifold it again, but at slightly different sections of the document so that it has deep folds in it.
For all scans, set the output resolution to 300DPI and do a full scan of each sheet.
The first set of scans will be with the scanner cover off.
The second set of scans will be with the scanner cover on.
There are six experimental conditions, comprised of three document conditions by two scanner conditions Lable the images accordingly
Now you should have nine printed copies of a single COLB that are all the same — provided that your printer feeds blank sheets in uniformly.
This test is to see what happens to the resulting images under different experimental conditions while keeping the document variables constant, such as their print resolution, their height, width, and aspect ratio.
The dependent variable is horizontal border height, as measured from the center of the image, from the outer edges of the bottom border to the the outer edges of the top border
The null hypothesis is that all scanned images of the same COLB will be identical.
The alternate hypotheses are that
(1) COLB’s with deep folds and with the scanner cover off, will have shorter vertical borders than COLB’s with narrow folds and with the scanner cover off.
(2) COLB’s with deep folds and with the scanner cover on, will have shorter vertical borders than COLB’s with narrow folds and with the scanner cover on.
(3) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(4) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
(5) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(6) COLB’s without folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
What I predict is that the null hypothesis will be rejected in favor of the first five alternative hypotheses, but not for the 6th hytpothesis.
For one additional experiment to confirm that document folds change the vertical border length (or horizontal border height), take the deeply-folded COLB, and flatten it out, by bending it back against the folds and using an electric iron across the back of it.
Then do two more scans, one with the scanner cover on, and one with it off. Then compare your results to the scans of the same documents when they were scanned while still being folded.
Doing these experiments will answer the questions that Ray and John Q posed above.
But, I can respond verbally to John Q’s conclusion. Because the PD COLB image has a lower height to width ratio while Michele’s COLB has a higher height to width ratio, by resizing the width of the PD COLB image to match the width of Michele’s COLB image, the border height of the PD COLB will still be lower than the border height of the Michele COLB, as would the text.
Even if you unlocked the aspect ratio of the PD COLB image as to allow you to resize both its width and height to match Michele’s COLB image, you are also changing the relationship between text position and border position.
The text position on the PD COLB would not “hop along for the ride” when you increase the border height because you are changing both the border width and height. Making the border wider flattens the image even more, and when you also make the border higher, you are counteracting the effects of that flattening such that the relationship of text to border remains pretty much the same.
This is why I proposed the experiments, to see for yourself what happens,
August 13, 2008 at 8:45 am
polarik
Sorry, I made a typo on the alternative hypotheses/ They should be
(1) COLB’s with deep folds and with the scanner cover off, will have shorter vertical borders than COLB’s with narrow folds and with the scanner cover off.
(2) COLB’s with deep folds and with the scanner cover on, will have shorter vertical borders than COLB’s with narrow folds and with the scanner cover off.
(3) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(4) COLB’s with narrow folds and with the scanner cover on, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(5) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(6) COLB’s without folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
August 13, 2008 at 8:48 am
polarik
Oh, crap. I mistated #5. It should be
(5) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
August 13, 2008 at 8:54 am
polarik
One more time. Please ignore the previous experimental conditions. My wife was bugging me to get off the ‘puter.
Here is the correct order:
Sorry, I made a typo on the alternative hypotheses/ They should be
(1) COLB’s with deep folds and with the scanner cover off, will have shorter vertical borders than COLB’s with narrow folds and with the scanner cover off.
(2) COLB’s with deep folds and with the scanner cover off, will have shorter vertical borders than COLB’s with narrow folds and with the scanner cover on.
(3) COLB’s with deep folds and with the scanner cover on, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
(4) COLB’s with narrow folds and with the scanner cover off, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(5) COLB’s with narrow folds and with the scanner cover on, will have shorter vertical borders than COLB’s without folds and with the scanner cover off.
(6) COLB’s with narrow folds and with the scanner cover on, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
(extra) COLB’s with folds flattened and with the scanner cover on, will have shorter vertical borders than COLB’s without folds and with the scanner cover on.
The predicted outcome is that any COLB that has folds in it, whether deep or narrow, will affect the aspect ratio such that the vertical borders are shortened.
Removing said folds from those COLB’s that have them, will eliminate the change in aspect ratio
Hopefully, I got it right this time.
August 13, 2008 at 9:09 am
johnque
I think you are missing the points I stated.
1) If folds affect the borders, those should be visible in the images we’ve seen so far, especially the Decosta, since that is the one that is claimed to have deep folds. Forget the image ratios, etc. I’m talking about the physical appearance of the borders. With Photoshop, we can zoom into pixel level detail. The border pattern is consistent enough to reveal even small variations in the border. We don’t see it in the Decosta. That’s evidence to me that folding the paper has not affected the borders nor is there evidence that the folds affected how the document was scanned.
2) IF the paper is folded to a degree that it prevents the document from laying flat, that should be apparent in more than one way (border distortion, shadows on paper, etc.) We don’t even have to deal with image analysis, ratios, etc. Those are physical signs that the folds have affected the scans. Again, there’s no sign of that.
August 13, 2008 at 9:17 am
koyaan
polarik:
Koyann, what possible justification do you have for not following my directions?
The fact that there’s absolutely no justification for cropping the images.
How exactly do you make a fair comparison of three different images, all sized differently and cropped differently?
Because, if you’re comparing such things as the aspect ratios of the borders, the images being sized differently and cropped differently are irrelevant. Size and cropping doesn’t change the aspect ratio of the borders.
If I take an image and draw a rectangle somewhere in the middle of it that’s 3″ x 5″, that rectangle has an aspect ratio of 0.6:1. If I size the image to be twice as big, and the rectangle is now 6″ x 10″, it still has an aspect ratio of 0.6:1. If I crop the rest of the image out, leaving just the rectangle, it still has an aspect ratio of 0.6:1.
So again, there is absolutely no need to crop any of the images.
k
August 13, 2008 at 9:20 am
polarik
For Ray and John Q.:
I know that the PD image was the only COLB image on the Net because I did an extensive Internet search on June 13th, using full and partial matches to Certification of Live Birth.” If it cannot be found, then you cannot use it.
Doing the experiment above will help[ to clarify what I mean. If you wish to dismiss it out-of-hand, and refuse to do what I did, then you, like Koyaan, have no right to criticize my work.
Let me try and restate my argument for using ONLY the printed COLB itself, everything from the border on in, and to not include the paper portion of the image that lies ouitside of the borders.
The paper portion outside of the COLB itself serves no useful value. Only the printed COLB itself has any value. So, what’s the big woof about removing those parts of the image that are of no useful value???
How does removing extraneous parts that are not germane to the discussion change the nature of the COLB itself?
Think about an image of a typewritten document. Are you saying that the blank margins on the page actually have relevance to interpreting the meaning of the writing itself, or for analyzing one writing sample to another?
Essentially that is what you are saying by insisting on including the blank margins around the printed COLB matter itself.
That’s absolutely ludicrous.
Either you replicate my experiments, exactly as described, or you don’t do them at all. and prove to everyone that you have no stones.
Explaining them away is also a gutless move.
August 13, 2008 at 9:26 am
polarik
Koyaan, do you ever read what you write and make sense of it??
Because, if you’re comparing such things as the aspect ratios of the borders, the images being sized differently and cropped differently are irrelevant. Size and cropping doesn’t change the aspect ratio of the borders.
So, using your same logic above, cropping [images at the borders] doesn’t change the aspect ratio of the borders.
So, what’s the problem with cropping them at the borders if it doesn’t change the aspect ratio?
Answer: there is no problem.
After all, it’s the aspect ratio that gets “messed up” when you scan a document with folds and with the cover off.
August 13, 2008 at 9:28 am
polarik
Keep in mind that I am not cutting into the borders in any way. I’m only removing the portions of the margins lying outside the borders.
August 13, 2008 at 9:38 am
rayinaus
It won’t hurt to crop each COLB – even if it serves no useful purpose.
August 13, 2008 at 9:50 am
koyaan
So, what’s the problem with cropping them at the borders if it doesn’t change the aspect ratio?
Because as I said, the only honest way to compare two images, is to compare the two images. Again, you don’t create two new images and then compare the new images and call it an honest comparison.
But if you want cropped, here’s cropped:
obamapdcrop.jpg
k
August 13, 2008 at 10:20 am
rayinaus
Good job!
August 13, 2008 at 11:27 am
koyaan
One other point.
What’s the point of using something as a template, if you don’t use it as a template?
As I’ve noted before, while “CERTIFICATION OF LIVE BIRTH” match up pretty well between the Obama image and the PD image, the “ANY ALTERATIONS…” line in the PD image is significantly wider than that in the Obama image.
If you scale the PD image to get a good match on “ANY ALTERATIONS…” then the border sizes end up being really out of whack.
alterations.jpg
As a side note, this image also illustrates the different aspect ratios of the two borders as evidenced by the different offsets between the horizontal and vertical planes. In other words, the amount of offset between the two right vertical portions of the borders is greater than the offset between the two bottom horizontal portions (the two images were aligned at the upper left corner).
So, given the fact that the aspect ratios of the borders are different, that “ANY ALTERATIONS…” don’t match up (nor does “STATE OF HAWAII” all that well), I see absolutely nothing to indicate that the PD image MUST have been used as the template for creating the Obama image.
And has been noted before, simply scaling the Obama image, the Michele image gives a far better match in terms of border aspect ratio, “CERTIFICATE OF LIVE BIRTH,” “STATE OF HAWAII,” and “ANY ALTERATIONS…”
To that end, I see no compelling evidence that the PD image was used as a template. And there would simply be no need for that to be the case anyway as even if it is a forgery, they obviously already had something far far better to work with than the PD image.
k
August 13, 2008 at 2:47 pm
polarik
God, Koyaan, you are un-be-freakin’-believable!
You simply refuse to even try anything I suggested…even for grins.
Yu make this off-the-wall comment, “What’s the point of using something as a template, if you don’t use it as a template?”
That only refers to your intransigence to even try what I suggested.
You didn’t deskew the borders. You didn’t crop them , even after I gave you the actual crop dimensions. You’ve got this bizarre notion that cropping an image is creating “two new images.”
It’s tantamount to you saying that I cannot compare the relationship of 4/8 to 1/2 because I changed the numbers.
Hats off to Ray. He gets it when he said, “it won’t hurt to crop each COLB.”
But not you.
What you insist on doing is like taking a printed image on letter-sized paper and comparing it to the same image, printed on the same printer, but using legal-sized paper, and then refusing to cut out just the image, because you want to compare letter-sized paper to legal-sized paper.
You also “messed up” the aspect ratio in the image you created.
You cannot even do your own thing honestly by your own standards. You arbitrarily cropped both the Kos image and PD image, to create this obamapdcrop crap that does not retain the same aspect ratios as the COLBs.
Then, you also mess up the aspect ratios of yet another arbitrary comparison to prove your point, which is no proof at all.
Look again at the comparison between two similarly-sized COLBs that were made by orienting them straight up and down, by cropping them right to the pixel edges of the borders, and then setting the width (or height) of one image to the other while keeping the aspect ratio the same, there is a perfect match of the PD COLB to the Kos COLB:
There is a match between the word, ALTERATIONS, on each image.
When you follow the same procedures for straightening Michele’s COLB, cropping it to the border, and then setting the width (or height) of one image to the other while keeping the aspect ratio the same.there is NO match to the Kos image:
You,Koyaan, are just, plain chicken to do anything my way because you know — yes you do — that by doing it my way, you will be forced you to admit that I was right.
So, instead, you put together these f*cked up images which you used “brute force” to match them, SOLELY to try and prove me wrong.
You will not budge an inch. You will be thickheaded regardless of what I do. You will continue to live in Fantasyland while I’m in the World of Tomorrow.
You’re hopeless. I’m out.
August 13, 2008 at 6:56 pm
rayinaus
polarik wrote:
“God, Koyaan, you are un-be-freakin’-believable!”
I suppose I’d better do the right thing and comment here after being so outspoken everywhere else lately. Yes, Koyaan has got a few unbelievable quirks, but I’ve got to stick with him for now because he’s one of the few who is producing some good work on many fronts.
Why not give it another shot in case Koyaan didn’t see why your test needed to be followed.
Ray
August 13, 2008 at 7:14 pm
koyaan
Oh for cryin’ out fucking loud.
Ok. One last time.
I took each image and adjusted rotation so that the leftmost edges of their borders were as perfectly vertically straight as possible as referenced from the top leftmost edge to the bottom leftmost edge.
I then adjusted the skew of each image to get the most perfectly square corners at all four corners
I then cropped each image as close as possible to the outside edge of the borders without cutting off any of the border.
I then overlaid the cropped PD image on the cropped Obama image, set the PD image’s transparency to approximately 50%, and scaled the cropped PD image so that the width of its border matched the width of the border of the cropped Obama image referenced to the top left and right corners of the borders.
I then aligned the PD image both horizontally and vertically, again also referenced to the top left and right corners of the borders.
Finally, I exported the composite image at 300 DPI.
Here’s the result:
deskew.jpg
The aspect ratios still do not match. The border of the PD image is still taller than the border of the Obama image. And the horizontal alignment of text below “CHILD’S NAME” still gets progressively worse until it’s of by a significant amount by the time you get to “OHSM…”
Software used was Xara Xtreme Pro 3.2.
I don’t know what more you want.
k
August 13, 2008 at 7:45 pm
koyaan
What you insist on doing is like taking a printed image on letter-sized paper and comparing it to the same image, printed on the same printer, but using legal-sized paper, and then refusing to cut out just the image, because you want to compare letter-sized paper to legal-sized paper.
In addition to his insisting that the borders have to be cropped out of the uncropped images, this is a good example of how Polarik demonstrates that he simply has no business whatsoever talking about graphics and imaging.
If I draw a 3″ x 5″ rectangle on a sheet of legal size paper, and a 3″ x 5″ rectangle on a sheet of letter size paper, I still have two 3″ x 5″ rectangles which still have the same aspect ratios.
The size of the piece of paper is absolutely irrelevant with regard to making any comparisons between those two 3″ x 5″ rectangles. They’re still 3″ x 5″ rectangles. And if I chop off some of the paper, again, they’re still 3″ x 5″ rectangles.
So I’ve no idea why Polarik seems to think that anything has to be cropped.
k
August 13, 2008 at 8:33 pm
koyaan
From TexasDarlin:
polarik
I love it. I love it. I can’t get enough of it.
O-Bomb-Out is going down. This is going to make the 1968 DNC in Chicago look like a picnic.
Great work, TD!
There seems to be little doubt now why Poarlik has worked so hard at concocting a bunch of cockamamie theories about Obama’s COLB being a forgery.
He started out wanting it to be a forgery and would simply not consider anything to the contrary.
k
August 13, 2008 at 8:34 pm
rayinaus
Cropping it is just a preferred way of working.
Incidentally, I cannot see why you changed the proportions if Polarik is not doing it.
August 13, 2008 at 8:36 pm
rayinaus
(preferred for Polarik with this job)
August 13, 2008 at 8:48 pm
koyaan
Cropping it is just a preferred way of working.
Yeah? Why’s it preferred?
Incidentally, I cannot see why you changed the proportions if Polarik is not doing it.
What do you mean? Changed the proportions of what?
k
August 13, 2008 at 9:42 pm
rayinaus
koyann wrote:
[Ray]: Cropping it is just a preferred way of working.
[Koyaan]: Yeah? Why’s it preferred?
Different brain wiring or something. Lots of people have their own way of doing things if it’s new territory.
[Ray]: Incidentally, I cannot see why you changed the proportions if Polarik is not doing it.
[Koyaan]:What do you mean? Changed the proportions of what?
The only thing we’ve been talking about – the COLB.
Ray
August 13, 2008 at 10:06 pm
koyaan
Different brain wiring or something. Lots of people have their own way of doing things if it’s new territory.
Heheh. Well graphics and imaging sure seem to be “new territory” for Polarik.
The only thing we’ve been talking about – the COLB.
Sure. But what is it that you say I’ve done that Polarik hasn’t done that you’re meaning when you say I’ve changed proportions?
In other words, can you be a little more uh… specific?
k
August 13, 2008 at 10:24 pm
rayinaus
I didn’t bother reading all the material above, but I got the impression that Polarik just wanted you to lay one image on top of the other without stretching it and altering the proportions of one of them.
August 13, 2008 at 10:29 pm
rayinaus
[Ray]: Different brain wiring or something. Lots of people have their own way of doing things if it’s new territory.
[Koyaan]: Heheh. Well graphics and imaging sure seem to be “new territory” for Polarik.
The ‘new territory’ is his style of analysis – so he was just making some guidelines so that everything would be identical or close to it. Why am I telling you this – it’s obvious.
August 14, 2008 at 12:15 am
koyaan
I didn’t bother reading all the material above, but I got the impression that Polarik just wanted you to lay one image on top of the other without stretching it and altering the proportions of one of them.
Oh. Well, that’s impossible to do. The Obama image and the PD image aren’t the same DPI.
Even though the image information in the PD image says it’s 300 DPI, it could only have been 300 DPI if the certificate that was being scanned was about 3″ x 3″.
If you assume that the scan is of the full width of an 8.5″ sheet, then it’s really about 105 DPI.
So in order to do an overlay, you have to scale one or the other. Either enlarge the PD image, or reduce the Obama image. And since there’s no way to truly tell exactly what DPI it is, I’d just been scaling to the widths a the top of the borders.
By the way, I’ve since discovered that there is a PD image which does indeed have virtually the same aspect ratio as the Obama image.
Just for the hell of it, I went back over to the Valeehill website only to find a new scan of the Patricia DeCosta COLB. In some ways its a better scan (higher resolution) but in others its pretty crappy (the contrast is pretty washed out).
Here it is overlaid on the Obama image:
obamapdoverlaynew.jpg
As you can see, it’s a nearly perfect match, both with respect to the border’s aspect ratios and the horizontal alignment of the text.
Of course this does not vindicate Polarik’s claims because this is not the same image that was posted on the Valeehill website at the time all this was being discussed previously nor is it the same image that Polarik has been using to support his claim that the borders in the Obama image and in the PD image have the same aspect ratio. Though if you look, you’ll see that they’re still using the thumbnail of the older image (and the new image has the same file name as the old).
And if you overlay the old PD image on the new PD image, we have the same situation that we have when overlaying the old PD image on the Obama image:
pdpdoverlay.jpg
Just as before, the height of the border in the old PD image is taller than the border in the new PD image. As well, the horizontal alignment of the text gets gradually worse as you progress down from “CHILD’S NAME.”
Now you can’t make any arguments here because the two images are scans of the same document.
I don’t know just how Polarik screwed up in order to get the borders in the PD image to appear to have the same aspect ratio as the borders in the Obama image, but there’s absolutely no doubt that he screwed something up.
k
August 15, 2008 at 3:30 am
rayinaus
[Ray]: I didn’t bother reading all the material above, but I got the impression that Polarik just wanted you to lay one image on top of the other without stretching it and altering the proportions of one of them.
[Koyaan]:Oh. Well, that’s impossible to do. The Obama image and the PD image aren’t the same DPI.
Even though the image information in the PD image says it’s 300 DPI, it could only have been 300 DPI if the certificate that was being scanned was about 3″ x 3″.
If you assume that the scan is of the full width of an 8.5″ sheet, then it’s really about 105 DPI.
Originally it was 200 ppi so I would have thought that re-sampling from 200 to 300 would make it equal.
August 16, 2008 at 8:15 am
polarik
Koyaan, Ray is right and you’d do well to listen to him. DPI only pertains to the printed image, and has nothing whatsoever to do with the quality of the image displayed on your screen.
All computer screens display images at either 72DPI or 96DPI regardless of what the actual DPI happens to be. A 300 DPI scan will not look any sharper than one scanned at 96 DPI. Like Ray said, they will be printed out with different dimensions.
Maybe Ray or John Q can tell me why you still refused to follow my explicit instructions. Are you so stubborn that you block out anything not fitting your own mindset?
Ray or John, please jump in here and maybe use the dimensions I provided/
Koyaan’s deskew image is 2389 x 2329
Yes, those dimensions are of a different aspect ratio.
Here are my suggested crop measurements:
Kos: 2373 x 2312 pixels
PD: 845 x 823 pixels
Michele: 2392 x 2327 pixels
After copping them, please change the width (smaller number) of the PD COLB from 823 to 2312.
Thanks,
RP
August 17, 2008 at 8:08 am
polarik
John Qur said:
If folds affect the borders, those should be visible in the images we’ve seen so far, especially the Decosta, since that is the one that is claimed to have deep folds. Forget the image ratios, etc. I’m talking about the physical appearance of the borders. With Photoshop, we can zoom into pixel level detail. The border pattern is consistent enough to reveal even small variations in the border. We don’t see it in the Decosta.
Here are closeups of the border where the folds ran through them. You can see exactly how the folds messed up the height of the borders.
August 17, 2008 at 8:22 am
polarik
For comparison purposes, here are the same areas on the Kos COLB image:
August 17, 2008 at 12:33 pm
polarik
Actually, Ray and John, the scanner top off was not the basis for my argument. The reason I mentioned it is because the PD COLB document was not scanned perfectly flat against the glass. Koyaan was the first one to mention the “cover off” scenario as to why, in his observation of the image, its vertical borders got messed up. I just followed suit.
What’s really relevant here is that the PD COLB has some serious folding in it, and I proposed the experiment above to prove my point. But, try as I may, I cannot seem to get anyone to even try it.
Before going into that, I need to address this “cropping the image” non sequitur, by giving you guys a “pop quiz:”
Q: What was the original size of the Kos COLB image?”
A: 2474 x 2369
B: 2550 x 3300
C: 1024 x 1000
D: All of the above.
E: none of the above.
Q: What was the original size of the PD COLB image?”
A: 900 x 921
B: 2550 x 3300
C: 843 x 825
D: All of the above.
E: none of the above.
Q: What was the original size of Michele’s COLB image?
A: 2474 x 2369
B: 2550 x 3300
C: 1024 x 1000
D: All of the above.
E: none of the above.
Extra credit question. You and your partner just received your Business License, which you both want to display it at your place of work. The License was printed on an 8 1/2″ x 11″ sheet of paper. On one end is a 3″ stub that is detachable along its perforations. When detached, it creates an 8 1/2″ square certificate. You both decide to make a xerox copy of it, frame it, and hang it in your respective offices.
Now, you make a copy of the entire page and post it with the stub attached. However, your partner makes a copy of it after removing the blank stub. Have you or your partner changed the aspect ratio of the certificate?
—————–
I await your answers.
August 17, 2008 at 2:16 pm
polarik
Ray says:
I couldn’t bear to read through all the above text while knowing that it is all a complete waste of time because it is completely irrelevant and wrong
Wow, Ray. Have you added psychic to your already-padded resume? I wish you had told me, so I wouldn’t have to write so much.
because, as I keep saying, Hawaii had a different border in 2007
Yes, indeed. You are definitely psychic. How use can you see something that nobody else has seen?
Can you do a Vulcan mind-meld with the rest of us so that we can see it, too?
however I have a few comments for when you finally get a grip and start comparing apples with apples, i.e. a 2007 certificate with wide cross-hatching with another 2007 certificate with wide cross-hatching.
Which one is the other 2007 certificate? It can’t be Obama’s.
It’s a waste of time to crop the scanned images to isolate the printed images because the printed images can be lined up by simply moving them on a larger canvas.
Oh, really? How, exactly do you compare one image to another that’s four times its size?
It’s obviously a good idea to align the two printed images at the left side, and to rotate one of them if necessary.
Wouldn’t also be obvious that you need to enlarge one of them or shrink the other to make a point-to-point comparison?
If I thought it would do any good I’d put my 2007 Marriage certification on the net (with the same wide cross-hatched border) for everyone to play with, but I already know what would happen if I did — nothing. Everyone would keep arguing nonsensically about nothing.
If you want to talk the talk, then you have to walk the walk. Or, you can continue talking “nonsensically about nothing.”
August 17, 2008 at 2:18 pm
polarik
<b..I didn’t bother reading all the material above, but I got the impression that Polarik just wanted you to lay one image on top of the other without stretching it and altering the proportions of one of them.
YES you are, Mighty Carnac!
August 17, 2008 at 2:21 pm
polarik
John Que:
If folds affect the borders, those should be visible in the images we’ve seen so far, especially the Decosta, since that is the one that is claimed to have deep folds. Forget the image ratios, etc. I’m talking about the physical appearance of the borders.
Please let me know if you require a closer view than what I provided above.
Like, maybe at the pixel level.
August 17, 2008 at 2:25 pm
polarik
Koyaan pondered:
f I draw a 3″ x 5″ rectangle on a sheet of legal size paper, and a 3″ x 5″ rectangle on a sheet of letter size paper, I still have two 3″ x 5″ rectangles which still have the same aspect ratios.
So, in other words, if you drew 3″ x 5″ rectangles on the DeCosta image, the Tomoyasu image, the Smith image, the Michele image, and lastly, the bogus Kos image, then they would all have the same aspect ratios.
Makes sense to me.
August 17, 2008 at 2:33 pm
polarik
I’m all for comparing “apples to apples,” provided that both of them are real instead of one being real while the other is made out of wax.
August 17, 2008 at 2:33 pm
koyaan
Q: What was the original size of the Kos COLB image?”
F: Doesn’t matter
Q: What was the original size of the PD COLB image?”
F: Doesn’t matter
Q: What was the original size of Michele’s COLB image?
F: Doesn’t matter
Now, you make a copy of the entire page and post it with the stub attached. However, your partner makes a copy of it after removing the blank stub. Have you or your partner changed the aspect ratio of the certificate?
Either this is an intentional obfuscation or you demonstrate once again that you have absolutely no business talking about graphics and imaging.
The issue was the aspect ratio of the PRINTED BORDER. And the aspect ratio of the printed border DOES NOT CHANGE depending on what else is included in the image.
Draw a 3″ x 5″ rectangle.
Print it on a sheet of 8.5″ x 11″ paper somewhere in the bottom 2/3 of the page.
Scan the entire page. Call it Image 1.
Cut off the top 3″ of the piece of paper.
Scan what’s left of the page. Call it Image 2.
Take either Image 1 or Image 2 and crop the rectangle from the rest of the image. Call it Image 3.
The aspect ratio of the rectangle in Image 1, Image 2 and Image 3 will all be the same.
k
August 17, 2008 at 2:34 pm
polarik
Oh, and Ray, just for your benefit, I’m keeping all my subsequent posts nice and short.
August 17, 2008 at 8:10 pm
johnque
“Please let me know if you require a closer view than what I provided above.
Like, maybe at the pixel level.”
No, I’ve viewed the images at the pixel level as I stated in my posts. My version of Photoshop allowed me to zoom to 1600% and overlay a 1 x 1 pixel grid over the image. Not sure how much closer I could get zoomed in but it doesn’t matter from my viewpoint. As I stated, there’s no evidence of the folds affecting the image.
I can see from Polarik’s comments and additional images that I didn’t make one point clear. I’m not questioning whether there are folds in the PD COLB. Those are quite clear. I would expect most, if not all, COLBs to have some evidence of folds. My point is that there’s no evidence that the folds in any way distorted the image in the way that Polarik has previously stated has happened Go back and read my previous comments above. If the folds had distorted the image, the borders would be inconsistent or shortened at the folds. There’s no evidence of that. There’s also no sign of shadowing that would be evidence of the paper being scanned not lying flat on the scanner. In summary, there’s no evidence that the image was scanned in a way that it wasn’t lying flat on the scanner.
August 18, 2008 at 10:00 pm
polarik
John Q:
If the folds had distorted the image, the borders would be inconsistent or shortened at the folds. There’s no evidence of that.
There would be if you took the time to do at least two parts of my proposed experiment.
Where is your proof that the distortions in the folds (which are clearly visible) did NOT affect its aspect ratio?
Koyann showed that the Kos image did not have the same aspect ratio as Michele’s COLB, but did match the aspect ratio of the PD COLB, especially with the new image.
Here’s the thing that drives me up a wall. You guys will run your mouth off without producing any visual evidence to support your claim. to his credit, Koyaan has, but you, John have showed SQUAT!
Do you really think yopu can score points by saying that the Moon is made out of green cheese without serving up a slice?
You do not deserve to be on the same blog post with Koyaan. You’re a bullsh*t artist, and nothing more.
WTF? I'[m supposed to take your word for it simply because you think you know more than I? Jesus, man, If you had any balls, you would take my challenge and do at least a part of what I asked you to do, seeing as how you’ve already written off everything that I say or do.
So, either you Man up, or shut up.
August 18, 2008 at 10:36 pm
polarik
Well, there goes a wasted acknowledgment to Koyaan.
Which part of the quiz did you NOT understand? If you took a driving test like you did mine, you’d be walking back and forth to your daycare center.
Like I said to John Q, apparently you are the biggest bullsh*tter of the pack. It’s your lack of knowledge of graphics that scares the piss out of you because the last thing you would ever do is something that would prove me right.
So, we have another gutless wonder here who cannot fathom even the simplest of instructions.
I have more experience with graphics than all of you combined, and that has all of you running scared, and refusing to even try what I suggested.
So, allow me to force-feed it to you. The borders on Patricia DeCosta’s REAL COLB are distorted in four places. The main body of this COLB is also distorted in two places — the top third and the bottom third.
These two folds created two physical channels running right through the entire width of the COLB.
And it is these physical abnormalities that caused the COLB image to be foreshortened. To say otherwise is total bullsh*t.
Now, since neither of you answered the quiz, it tells me that neither of you have the slightest clue as to the original sizes of all these Kos images.
Koyaan is so far removed from reality that he cannot see how we’ve got five different COLB images (Kos, PD, Tomoyasu, Smith, and Michele) that were arbitrarily cropped to different image sizes.
The only way to compare them is to convert all images to transparencies, pick one dimension from one of the certificates, and then set that one dimension to be the same on all of the other images. For example, making all of the certificates to have the same width. Then see what happens when one is laid on top of the other.
Keep in mind, however, that all of the images have been arbitrarily cropped to different sizes, you absolutely cannot simply change the canvas size of each image because no two images are the same in size, because they were cropped differently, and each has different margins surrounding the certificate itself.
Either you have to set all of the margins to be the exact, same amount — which would require enlarging or reducing them (something that Koyaan stubbornly maintains will “change” the images).
I asked you, “How do you compare an image of a printed certificate to every other image of a printed certificate , when the certificates sre not the same size, and the cropped margins are not the same size?
If you cannot at least overlay one on top of the other with both having one dimension the same, then you cannot do any analysis of their aspect ratios.
Cropping an image to the edges of the certificate such that it has zero marginsis no differentthan arbitrarily cropping an image to “whatever floats your boat.”
How you can miss this basic concept is beyond me.
If you are going to go back to your non sequiturs and ad hominem arguments, then it tells me that you have even less knowledge about graphics than I give you credit for having.
The aspect ratio of the PD COLB matches the Kos COLB, even though there are large folds, visually plain to the eye that distort the f*cking text fields. If you cannot see that, then please pay a visit to your local optometrist.
August 18, 2008 at 11:32 pm
koyaan
The only way to compare them is to convert all images to transparencies, pick one dimension from one of the certificates, and then set that one dimension to be the same on all of the other images. For example, making all of the certificates to have the same width.
That’s not necessary at all.
If you’re comparing the borders to see whether their aspect ratios are the same or not, it doesn’t matter at all what the overall image sizes are or that the overall sizes are the same. You’re only comparing the borders. So you only need to scale one of the images so that one dimension of the borders are the same on all the images.
Once you’ve done that, then you’ll know whether or not the borders you’re comparing have the same aspect ratio. If they have the same aspect ratio, then both border dimensions will be the same. If they don’t have the same aspect ratio, they won’t be.
Simple as that.
And that’s exactly what I’ve one. I scale one image so that the widths of both borders are the same and then see if the heights are the same. If the heights are the same, then the borders have the same aspect ratio. If they’re not, they don’t.
Again, simple as that.
And again, the aspect ratios of the Kos border and the original PD border are NOT the same. The aspect ratios of the Kos border and the new PD border as well as the Michele border ARE the same.
That you seem to think that either the heights or widths of the overall images need to be matched instead of just the borders shows that you haven’t any business talking about graphics and imaging. That you still keep saying this even after I’ve provided you with the actual images showing that the aspect ratios are not the same shows that you’re either being disingenuous or you’re mildly retarded.
k
August 19, 2008 at 12:24 pm
polarik
Yee-ha! We’ve had a breakthrough, Steve. We are in agreement, and the only thing that separates us are the margins around the borders.
I was trying to get you to say how, exactly, do you measure just the borders on a larger canvass. Either you use a rectangular selection tool to measure them, or a pixel ruler stretched along each dimension.
If you were using a rectangular selection tool, then the only thing I did differently than you is to crop the borders with it.
The advantage of doing it my way, however, is that you automatically know what is the size of the borders by looking at the file information. You do not even have to open the file.
Different strokes for different folks.
Now, hear me out. Regardless of whether I go ahead and crop the selected border, or you simply measure the border, we should get the same starting measurements for each image.
To the extent that we disagree, I will be more than happy to use your starting measurements, if you, likewise, use my starting measurements.
Then, we will have two sets of images — that’s assuming if our measurements differ more than whatever pixel precision you desire.
If by some quirk of Fate that we have the same measurements to begin with, so much the better.
But, assuming for the moment, that we have different measurements. Then we will be working with two pairs of images.
For both sets of measurements, we need to agree whether we are going to to set the larger image width to that of the smaller image width, or vice-versa. Then, it’s only a matter of comparing them to see where they do not match, if at all.
I think that’s a fair trade off. Don’t you think? I will be as receptive to your ideas as you are to mine.
In fact, I’ll even make a concession — a gesture of my good will — by assuming that the Kos border is what a 2007 border looks like.
Shall we commence?
I hope that, after we do this exercise, that we will dispense from doing any name calling.
Agreed?
August 19, 2008 at 12:34 pm
polarik
I should like to add that, I will scale the image as you do, if you tell me exactly how you do it.
I’m not a bad guy at all when you get to know me.
August 19, 2008 at 12:55 pm
koyaan
I was trying to get you to say how, exactly, do you measure just the borders on a larger canvass. Either you use a rectangular selection tool to measure them, or a pixel ruler stretched along each dimension.
I do neither.
I should like to add that, I will scale the image as you do, if you tell me exactly how you do it.
As I said before, I import both images, overlay one on the other, set the top image’s transparency to approximately 50% so I can see the image under it.
I scale the top image so that its border’s width is the same width as the other border. Then I horizontally and vertically align the top left and top right corners of top image’s border with the other border.
Then I simply look at the heights of the two borders to see if they’re the same or not. If they’re the same, they have the same aspect ratio. If they’re not, they don’t.
I’ve already provided you the image showing you this.
obamapdoverlay.jpg
As you can see, the width of the PD border is the same as the width of the Kos border. However the height of the PD border is greater than the height of the Kos border. Hence, they don’t have the same aspect ratio.
I don’t know what more can be done to get this through to you.
k
August 19, 2008 at 1:26 pm
koyaan
Here, here’s a step-by-step pictorial of what I’m doing.
Square one.
process1.jpg
Import Obama image (FactCheck).
process2.jpg
Import original PD image.
process3.jpg
Set PD image to approximately 50% transparency.
process4.jpg
Scale PD image so that the width of its border is the same as the with of the Obama border, and align with the upper left and right corners.
process5.jpg
Check the heights of the two borders.
process6.jpg
And as I’ve said numerous times before, the height of the original PD border is greater than the height of the Obama border. Meaning the aspect ratio of its border is not the same as the Obama border.
If you still don’t get it, then you’re just going to have to not get it. I’ve done all I can do to illustrate this fact to you.
k
August 19, 2008 at 6:17 pm
polarik
As you can see, the width of the PD border is the same as the width of the Kos border. However the height of the PD border is greater than the height of the Kos border. Hence, they don’t have the same aspect ratio.
I don’t know what more can be done to get this through to you.
I don’t know either. What about trying some couples therapy first before we split up due to irreconcilable differences?
For starters, weren’t you the one complaining to me about changing thr size of an image…like how that creates another image? From where did you get the extra green border? It’s not a part of PD or Kos.
You’ve changed the sizes of the borders, as well as changing the image dimensions. I can’t see what you did here? Like, besides down at the bottom, where does the Kos begin and PD end? And, vice-versa?
This is why I needed to know the values of your border sizes — especially now that you boogered them up.
The dimensions of your overlay is 2537 x 2599. WTF?
The original Kos image is 2427 x 2369. Setting it equal to the width of your overlay should have made it 2663 x 2599 — not 2537 x 2599. It’s way off.
The original PD image is 921 x 900. Setting it equal to the width of your overlay made it 2543 x 2599. Closer, but no cigar.
I thought that you weren’t going to screw around with the dimensions of the benchmark image. What happeed to “I scale the top image so that its border’s width is the same width as the other border.” That’s not what you did here. You created two different images.
My way is way easier.
August 19, 2008 at 6:40 pm
polarik
Sorry..I did not get to read your update before I sent mine.
I’ll look at it later today,
Rhanks
August 19, 2008 at 9:28 pm
koyaan
For starters, weren’t you the one complaining to me about changing thr size of an image…like how that creates another image?
No. I said CROPPING the borders out of each images creates two new images. It’s a wholly unnecessary step. Better to compare a complete image to a complete image.
From where did you get the extra green border? It’s not a part of PD or Kos.
There is no extra green border.
You’ve changed the sizes of the borders, as well as changing the image dimensions.
All I did was scale the entire image, maintaining its aspect ratio.
Like, besides down at the bottom, where does the Kos begin and PD end? And, vice-versa?
Besides down at the bottom, the borders begin and end in the same place because they’re directly overlaid on each other.
This is why I needed to know the values of your border sizes — especially now that you boogered them up.
I’ve boggered up nothing. I simply scaled the PD image so that the width of its border was the same width as the Kos border.
The dimensions of your overlay is 2537 x 2599. WTF?
The original Kos image is 2427 x 2369. Setting it equal to the width of your overlay should have made it 2663 x 2599 — not 2537 x 2599. It’s way off.
The Kos image was cropped closer to the edges of the border than the PD image. So when the PD image is scaled up to make its border the same width as the Kos border, then its overall size is greater than the Kos image.
The original PD image is 921 x 900. Setting it equal to the width of your overlay made it 2543 x 2599. Closer, but no cigar.
You’re aiming at the wrong target.
Again, the Kos image was cropped closer to the edges of its border than the PD image and the PD image is what ultimately determines the overall image width and height.
The original PD image is 900 pixels wide. 2537/900 = 2.819. 2.819 x 921 = 2596.3.
I thought that you weren’t going to screw around with the dimensions of the benchmark image. What happeed to “I scale the top image so that its border’s width is the same width as the other border.” That’s not what you did here.
That’s precisely what I did here. The benchmark image, i.e. the Kos image was not touched. Only the PD image was scaled.
My way is way easier.
Your way doesn’t directly compare two complete images.
k
August 20, 2008 at 12:03 am
johnque
“So, allow me to force-feed it to you. The borders on Patricia DeCosta’s REAL COLB are distorted in four places. The main body of this COLB is also distorted in two places — the top third and the bottom third.
These two folds created two physical channels running right through the entire width of the COLB.
And it is these physical abnormalities that caused the COLB image to be foreshortened. To say otherwise is total bullsh*t.”
WRONG. Is there evidence of folds? Yes. Did they foreshorten the image? No.
“Where is your proof that the distortions in the folds (which are clearly visible) did NOT affect its aspect ratio?”
The proof is in the borders themselves. You can clearly see that while the border image shows evidence of the fold, the border itself is not distorted. It’s uniform in size both above, below and through the area of the crease. If the fold was so great that it actually caused the image to be shortened, the border image would not be uniform through that area. But there’s zero evidence of this. Go count the pixels. Show me how the border above and below the fold is different in pixel count than the border where the fold occurs. If your theory is correct, you should have no problem showing that.
For some reason, you are confusing the fact that the toner on the paper was distorted by the fold with the paper itself actually being distorted. If I take a document printed on a laser printer and fold or crease it through the printed area, more often than not, the toner on the paper will show evidence of that fold or crease. If I then put that document down on a scanner, ensuring that the paper is perfectly flat, the evidence of the fold or crease will still appear in the scan. Does that mean that the paper somehow became distorted despite my best efforts to scan it flat? No. It means at the point on the paper where the fold or crease occurred, the print toner has been disturbed and no matter how flat I make the document, that disturbance will still show. There’s no doubt there’s evidence of the fold. But there’s no evidence that the fold affected the image size.
August 20, 2008 at 11:09 am
polarik
Koyaan:
No. I said CROPPING the borders out of each images creates two new images. It’s a wholly unnecessary step. Better to compare a complete image to a complete image.
You’re aiming at the wrong target.
You’ve got the right target, but your aim is off.
Again, the Kos image was cropped closer to the edges of its border than the PD image and the PD image is what ultimately determines the overall image width and height.
OK. I’ll do it your way.
The original PD image is 900 pixels wide. The original Kos image width is 2369.
So, the calculation of the percentage increase should be done this way: 2369/900 = 2.632 (or 236.2%). The rescaled image height would then be 2.632 x 921 = 2424.2.
By incorrectly rescaling Kos and PD images, you’ve unknowingly created two new images.
Using your procedure, I rescaled the width and height of the PD image by the percentage enlargement factor (2.632, and not 2.819).
The results show a PD image enlarged to 2424 x 2369, which, for all intents and purposes is identical to the original Kos image (2427 x 2369).
The results also match my border cropped image (2425 x 2369), demonstrating that either way will yield the same results. The reason why it produced the same result, is because the canvas size is irrelevant. You demonstrated that fact by producing a merged image having a differently-sized merged image than what the Kos image originally had.
To summarize, as per your instructions, I calculated the enlargement factor , for comparing one image to another, by dividing the width of the original Kos image by the width of the original PD image. This factor was then multiplied against the height of the original PD image, to make it comparable in size to the Kos.
This method
August 20, 2008 at 11:16 am
polarik
This method also confirms that the aspect ratio of the Kos COLB image is identical to the aspect ratio of the PD COLB image.
August 20, 2008 at 12:48 pm
koyaan
There’s nothing left to say except you’re a stupid fucking retard.
k
August 20, 2008 at 7:42 pm
polarik
There’s nothing left to say except you’re a stupid fucking retard.
Now, don’t try to sweet talk me. Flattery will get you nowhere.