Friday, May 6, 2011

More Curious Than Anything Else

Before class started the other day, one of my students wanted to show me President Obama's long form birth certificate when she loaded it in Adobe Illustrator, to show the "layers" that it seems to be made of.  I explained how OCR software often does that--it tries to identify areas of a scan that are potentially convertible to text,
and while it looks bad, it is not a sign that the document is a forgery.  But while I was explaining this to her, I saw something that she did not see.

Diagonal lines, when scanned, produce a jagged set of pixels.  This is an artifact of how sharply drawn the line is, and the scan resolution (dots per inch). All things being equal, two lines drawn at the same angle should produce similar levels of jagged pixels.  Yet when I looked at the mother's maiden name, "Dunham" at 800% in Adobe Acrobat, I noticed that the diagonals on the "D" are very, very noticeably different in their pixelation than the diagonal lines of the letters in the rest of the name.


click to enlarge


The attendant's name (the doctor who delivered the child with the halo over his head) has the same jagged, low-resolution characteristics as the "D" in Dunham--but not the "unham."  (This was 600%.)




click to enlarge


I am not saying that the certificate is a fake.  I am assuming that it is real--because the alternative explanation is it took Obama more than two years to have an incompetent fake put together.  That makes far less sense than the Kenya birth theory.  I am mostly curious why it looks so different.  I look forward to an explanation. I guarantee that if freshmen in my classes are asking these questions today, there better be an explanation or this birth certificate matter will never end.

UPDATE: An interesting comparison is Louisiana Governor Bobby Jindal's birth certificate, which is here. One of the commenters who thinks that we are picking on Obama's birth certificate points to several things that he thinks are suspicious:

1. that Jindal's mother's address is listed as "Physics Dept. LSU";
2. that the city is identified as "BAton Rouge";
3. that the hospital is "Woman's Hospital."

None of these seem all that odd to me. I suppose if there had been questions about Jindal's place of birth--and Jindal had spent years and lots of money fighting to prevent its release--well, I might be a bit more curious. What I do find interesting is that when I blow up the image of Jindal's birth certificate, I do not see the oddities that are in Obama's birth certificate--such as signatures that seem to have been scanned at a different resolution than other parts. Why did Jindal's birth certificate come out looking like it was just a scan of a document--and Obama's looks like it is a composite?

11 comments:

Sean said...

Good catch. If you haven't seen them already, you should look at Karl Denninger's posts on his blog. He discovered similar oddities: http://market-ticker.org/akcs-www?post=185202

I agree that the Kenya birth theory makes little sense, but when I see someone willing to spend $1M+ to hide something, I have to assume that there's something inconvenient they want to hide.

LtDave said...

The baseline of the typed information also changes thru out the document. Compare it to the other exemplar Hawaiian documents available on the web. There is also inconsistency in the tab stops of the document.

hga said...

I believe the answer is fairly simple, but I don't know, since my professional experience with document imaging is from the first half of the '90s, long before we could routinely afford color images.

First, for reference, this posting claims to have all the separate layers.

What you've picked up on appears to be artifacts of mixed raster content based compression, where the document is first separated into layers before the most efficient methods are used to compress each.

There's a color background layer, with some unrecognized gray scale left behind like the lines of the form and most of the signatures. There's a gray scale main foreground which has most of that including most of his mom's signature. There's a bunch of black and white layers on top of that, the ones I've examined are 1 bit deep, the state registrar's stock signature (along with the rest of that block) is in the first one.

The split in her signature between the background layer and the first foreground layer is right between the 'D' and the 'u'; my experience is limited to OCR where the artifacts I see are reasonable (e.g. the frequent failure to pick up trailing characters). I'm assuming the varied results we see in recognizing the signatures is due to ICR, or at least limits based on what's cheap to license and computationally reasonable in whatever application was used to compress this image. I.e. unlike normal OCR/ICR the better the recognition the better the compression, but failures in the former are not errors, they just make the result larger.

Rob said...

I suggest you spend several hours a day for the next 2 to 6 years looking into these inconsistencies. Next, you should compare them to the birth certificates of the Nordyke twins for more vital clues! Then you can investigate these important proofs that it's a forgery!

Clayton said...

Who has time to devote to this? Especially since Obama is pretty much guaranteed re-election in 2012 anyway?

Mauser said...

Whoah, Clayton, don't be so pessimistic.

We can still pray for a meteor strike.

Now on the one hand, I can actually believe the different layers and compression could be part of a "Save for the Web" setting. OTOH, if the source was a print off a Microfiche, then the green background was unquestionably edited in. The attempt to make the document look "More Authentic" has had the opposite effect. And the AP had a version without that background, which doesn't help matters.

Clayton said...

Mauser, you know if the end of the world happens what the New York Times editorial headline will be: "End of the World Coming: Minorities, Women Most Affected."

Bubblehead said...

Gov. Jindal released his "Certificate of Live Birth". Why does it say his mother lives in a "Physics Department"? What is a "Woman's Hospital", what woman does it belong to. Why does "BAton Rouge" start with two capital letters? Why did the doctor not sign until 3 days after the mother signed?
Or let me guess, you guys don't have any questions about a Republican's birth certificate, do you?

Clayton said...

I am guessing that the street address was listed as the Physics Department because that was the most permanent address for a student. "Woman's Hospital" might indeed be the name of the hospital where Gov. Jindal was born. BAton Rouge is obviously a typo. Why did the doctor sign until three days later? Probably the doctor did not get the form until three days later.

I'm curious: how many newspapers have published accounts referring to Jindal as being born in India? How many conflicting versions of his birthplace have shown up? How long between when questions came up and Jindal supplying his long form certificate?

Clayton said...

Interestingly enough, when I blow up Jindal's long form, everything is consistent. You do not have some sections of the signatures that look like they were scanned at a different resolution than others. That does not mean that it is not a forgery. It does mean that if it is a forgery, it was well done!

By comparison, Obama's long form certificate is awash in oddities that, in light of the long battle to keep it secret, are just making the birthers even more sure that he has something to hide.

As I have said all along, the energy put into keeping it out of sight suggests that Obama has something to hide, even if it is not where he was born.

Bubblehead said...

I think he's trying to hide the fact that Birthers are idiots on an almost Trutheresque scale, and he laughs his butt off whenever someone who might run against him panders to the Birthers.
Of interest, it looks like Gov. Jindal won't pass muster with the "both parents have to be citizens to be a NBC" clowns.