You are on page 1of 7

Under the heading that its never been done before ...

I have successfully extracted three distinct .png images


from the Muscatine Iowa Journal PDF file 4db82608b486f.pdf
(in one extraction step). These three images are all
72 PPI x 72 PPI pixel resolution but each is a different
page size.
This result is not expected for a PDF file which was
purportedly created by a simple Photoshop edit of a
JPEG image file.
The stats for these three .png files are:
1.png 687 KB
W = 628 px = 8.722 in. @ 72 PPI
H = 747 px = 10.375 in. @72 PPI
2.png 1458 KB
W = 1043 px = 14.486 in. @ 72 PPI
H = 1243 px = 17.264 in. @ 72 PPI
3.png 5948 KB
W = 1739 px = 24.153 in. @ 72 PPI
H = 2071 px = 28.764 in. @ 72 PPI
The page size of each of these three images in the Muscatine Journal PDF file 4db82608b486f.pdf is:
W = 8.695 in. ; H = 10.355 in.
The 2.png image is equivalent to a pixel resolution of 120 PPI x 120 PPI (at the final PDF page dimensions).
Likewise the 3.png image is equivalent to 200 PPI x 200 PPI resolution (also at the final PDF page dimensions).
As I first reported, this Muscatine PDF file 4db82608b486f.pdf has a very unusual property. When the PDF is opened in Adobe
Illustrator CC, the displayed image has a pixel resolution of 120 PPI x 120 PPI. However, when the same PDF file is opened in
Adobe Photoshop, the pixel resolution of the displayed image is 200 PPI x 200 PPI. In both instances, the page size is 8.695 in. x
10.355 in.
See:
This result is theoretically possible because Adobe PDF files and Photoshop PDF files can be different at code level.
However, I can now add to my previous comments my instant findings that three separate .png images can be extracted from
4db82608b486f.pdf using non-Adobe software. The single extraction step doesnt require (or utilize) either Adobe Illustrator or
Adobe Photoshop. Moreover, the three images are all near-White background, have the same pixel resolution (72 PPI x 72 PPI) and
differ in page size. The page sizes are different when measured in pixels or in inches.
The Obot blogger NBC blamed the different resolutions (between Illustrator and Photoshop) on the operator accidently checking the
preserve Photoshop editing capabilities box before saving the PDF from within Photoshop. However, NBCs explanation is
incorrect. Based on my trials, which simulate the Photoshop edit workflow, the observed results are not obtained with NBCs
workflow.
There were two trials which differ only in the format of the graphic image originally loaded into Photoshop CC. In both cases, the
original image was derived from the ABC PDF file ap_obama_certificate_dm_110427.pdf.
In the first trial, the ABC PDF image was utilized directly. The ABC PDF file was opened in Photoshop CC as a Photoshop PDF file.
For the second trial, the JPEG image was first extracted from the AP PDF file ap_obama_certificate_dm_110427.pdf. The JPEG
was then opened in Photoshop CC.
In both cases the initial page size of the displayed image was 13.490 in. x 16.170 in. and the pixel resolution was 200 PPI x 200 PPI.
The page size in pixels was 2698 px x 3234 px.
For both cases, the image size was then reduced from (2698 px x 3234 px) to (1043 px x 1243 px) and the pixel resolution was
reduced from (200 PPI x 200 PPI) to (120 PPI x 120 PPI). In both cases, final dimensions after scaling were (8.692 in. x 10.358.)
These final page dimensions were close to the page dimensions of the Muscatine PDF image (8.695 in. x 10.355 in.). In each case, the
scaled and downsampled image was saved as a new PDF file.
For each scaled image, the new PDF code was parsed to extract all bitmap images including the Photoshop private (composite) image.
In all cases (except for the thumbnail) the image dimensions were 1043 px x 1243 px. In no instance were the pixel dimensions of the
private image different from the pixel dimensions of the non-private image. This was the case for both trials. This finding is not
unexpected because both images (i.e. private and non-private) are written to file when the PDF file is saved. At this instant, the image
that is displayed in Photoshop is the scaled image.
To the contrary, from my earlier results gleaned from my initial parsing of the Muscatine Journal PDF file 4db82608b486f.pdf, we
already know that the file contains a private image of page size 1739 px x 2071 px and a non-private image of size 1043 px x 1243 px.
Consequently, I conclude that NBCs proposed workflow for the Muscatine Journal PDF (which is based on a simple edit in
Photoshop CS2) cannot be the correct workflow.
To the contrary, the actual workflow would necessarily be more complex than NBCs proposed workflow. Most likely it involved
assembly of multiple images into one PDF file. The separate images would have been created separately using one or more digital
graphic programs.
I also extracted the bitmap images from the two trial PDFs using the same tool that produced the three unique images (1.png, 2.png,
3.png) shown at the top of page 1. As before for the Muscatine PDF, three images were extracted from each trial PDF. However, for
both trial PDFs, two of the images (2.png and 3.png) were identical. As before, all of the extracted images have near-White
backgrounds.
DEBATE HISTORY OF AP/MUSCATINE JOURNAL PDF
I first reported my findings (from my initial analysis of the Muscatine PDF) in a lengthy post about the LFCOLB forgery on NBCs
blog in July 2013.
NBC intercepted my lengthy post on the forgery and broke it up into several full-page posts of his own. He also took the liberty to
intersperse his own comments within each segment of my one lengthy post. He finally posted the final segment of my original post on
Jul 28, 2013. These high-handed tactics of NBCs has lead to much confusion in the interim. One objective of this report is to clear
up some of the confusion.
See: http://nativeborncitizen.wordpress.com/2013/07/28/hermitian-applewhite-ap-document/
As I first reported, the Muscatine PDF displays distinctly different images when the file is opened in Illustrator and Photoshop.
Hermitian...
When the Muscatine PDF image (4db82608b486f.pdf) is opened in Photoshop CC as a Photoshop PDF file it opens at 200 PPI X 200
PPI resolution. This image is the one created by Applewhite. When the same PDF is opened in Adobe Illustrator CC as an Adobe PDF
then it opens at 120 PPI X 120 PPI resolution. Thus a Photoshop PDF file opened in Photoshop does not necessarily produce the same
image when the same Photoshop PDF file is opened as an Adobe PDF file in Illustrator or Acrobat.
NBC began by thrashing around on the Muscatine workflow question last July. He did so in reaction to my initial post about the
AP, ABC and Muscatine Journal images.
See : http://nativeborncitizen.wordpress.com/2013/08/01/educating-the-confused-muscatine-and-ap-jpegs/
and http://nativeborncitizen.wordpress.com/2013/08/04/46588/.
NBC then offered up his murky workflow for the Muscatine Iowa Journal PDF 4db82608b486f.pdf on 08/04/2013.
NBC...
So now we can explain what happened.
The user imported the 200 PPI jpeg (26983234) obtained from the Associated Press or 13.49 by 16.17 and scaled/cropped the
image to 17382071 (or 8.6910.35 to make it closer to 8.511 format). This image was then used to correct for the bluish color
imbalance. Once satisfied, it was used to create a PDF @120 PPI and the PSD file was stored internally.
First of all, Photoshop does not store a PSD file internally as a private or composite image. The raw file extracted from the PDF
does not have the PSD header. Instead this internal file is usually a Flate-compressed binary file. This file is most easily extracted as
a type P6 .ppm bitmap file. Or it can also be extracted as a .bmp or .jpg file.
Of course graphic images can be saved to disk in the PSD file format (native Photoshop format) from within Photoshop. Also when
embedded bitmap images are un-embedded in Illustrator, the unembedded image is often written to disk as a PSD file. The
unembedded image (in Illustrator) is automatically linked to the external PSD file on disk when the unembed command is invoked.
Photoshop also stores proprietary internal files as unknown object streams. These are usually type 8BIM format files which cannot be
opened. These files can be large in size. I extracted one 8BIM binary file of 25.5 MB size from one trial Photoshop PDF. Evidently
these Adobe proprietary files are heavily compressed within the PDF. These 8BIM files are useful only to Adobe.
NBCs final post on the AP/Muscatine PDF creation repeated the same workflow as above albeit with somewhat greater clarity.
http://nativeborncitizen.wordpress.com/2013/09/01/final-report-on-the-ap-documents-and-the-muscatine-pdf/
Unfortunately for NBC, clarity and correctness are not the same thing.
However, even after much bloviating NBC still got it completely wrong. His workflow cannot produce the images found in the
Muscatine PDF file 4db82608b486f.pdf.
To prove that close only counts in horseshoes and hand grenades we cite NBCs last gasp on the subject of the AP/Muscatine
workflow here:
http://nativeborncitizen.wordpress.com/2013/09/08/muscatine-pdf-and-the-ap-jpeg-part-2/#more-47686
Ok, I did the following
1. Import the JPEG into Adobe Photoshop
2. Change image resolution to 200 ppi
3. Crop image (rough crop)
4. Save as PDF downsample JPEG to 120ppi, high quality, preserve editing and fast load, PDF 1.3 compatible, embedded
thumbnail
5. Extract JPEG
Too bad that NBC didnt bother to carry out the experiment. He was having difficulty in extraction the private image with his
freetoy extractor. If he had extracted all of the bitmap images from his trial PDF, then he would have discovered that, unlike the
Muscatine PDF, all of the images have the same pixel dimensions.
But thats what always happens when unbounded arrogance overtakes measured and surefooted effort. When this happens, then
claims run far ahead of findings.
The Muscatine Iowa Journal published a one-page PDF document containing a purported single near-White background image of the
Obama LFCOLB. When the Muscatine PDF is opened in Adobe Reader XI the image displays at 120 PPI x 120 PPI pixel resolution.
This image is not indicated to be a layered image when opened in Illustrator or Photoshop. Nevertheless, without using any Adobe
programs, three separate images of this single page were extracted from the file 4db82608b486f.pdf in one extraction step. These
three images all have the same pixel resolution (72 PPI x 72 PPI) but different pages sizes (measured either in pixels or inches). When
scaled to the same page dimensions as the PDF image (8.695 in. x 10.355 in.), these three extracted images (1.png, 2.png, 3.png) have
respective pixel resolutions of (72 PPI, 120 PPI and 200 PPI). Each of these three images has a near-White background. The 3.png
image is associated with an internal Photoshop private image. The private image is a dark-background image when extracted with
PDF parser. These findings are not consistent with a simple edit of a JPEG image in Photoshop.
Collectively, these findings indicate a complex workflow rather than a single edit of a JPEG image in Photoshop CS2.
So my original question still stands...
Why did someone at the Muscatine Iowa Journal conceal three different versions of the Obama LFCOLB image in a one-page
Photoshop PDF document. And how was he able to do so in a single-page document without using layers ?

You might also like