Perhaps our friend has not realized that I have been building the ‘report’ in the last few weeks, on this site. Did he really miss it all?
With examples of other Xerox WorkCentre PDF’s which all show the same embedded jPEG and quantization tables.
Hermitian has been incredibly helpful in building out this vast array of evidence that support a Xerox WorkCentre workflow.
The Cold Case Posse is aware of the work and it’s really up to them to either admit to their mistakes or explain why the work flow is insufficient.
Hermitian: Just so the readers will know exactly what I am referring to — I am referring to the Xerox 7655 scan to PDF image of your latest printout of the Obama WH LFCOLB that you claim shows the same quantization tables as the Xerox 7535 scan to PDF file that you have released. The only quantization tables that you both have released and that you claim came from a Xerox scan to PDF of the copy of the WH LFCOLB was produced on a Xerox 7535 and not on the Xerox 7655 WC.
Hermitian is wrong. I never claimed that my scans showed the same quatization tables as the 7535 or WH LFBC PDF. For obvious reasons: I posted the evidence that they did not match. However, in a stroke of luck, I looked at the WH 7655 Tax Return form scans of our President and VP, and noticed that they were created on a 7655 Work Centre and how their quantization tables matched. And contrary to Hermitian’s claim, I did publish that information.
If Hermitian has anything relevant to contribute, please let him do so but at the moment his ‘arguments’ have been reduced to totally irrelevant issues.
Until then, I would like to thank him again for his contributions.
Hermitian: The readers should note that NBC’s purported Xerox forger has never achieved the clean separation of the text from the background that the human forger achieved by manually moving the text from the background to the mostly text layer.
Well it was not really that cleanly as somehow many characters were missed but yes, the quality of the separation depends on the quality of the original. So noone is really expecting a perfect match.
Hermitian:Consequently any text that is not separated from the background in NBC’s Xerox scans is downsampled from 300 PPI x 300 PPI to 150 PPI x 150 PPI. Also any text that sticks to the background layer is converted from binary monochrome aliased text to Grayscale, anti-aliased text. As the comparative screen captures reveal, the degradation to the text image that results from the failure to separate from the background is great. Most of the capital letter “As” are converted to solid characters with the hole in the “A” completely filled with solid color pixels.
Yes, when you take a MRC compressed document and print it out and scan again using MRC, degradation is unavoidable but also expected. These are minor quibbles about the process which have no relevance to the overall success of the work flow as it explains ALL the artifacts, once claimed by the Cold Case Posse to be 100% proof that the PDF was somehow forged.
Hermitian may not appreciate these subtleties of the scientific method, but there are two competing hypotheses here: 1) These artifacts cannot be explained by simple workflow processes (aka argument from ignorance) 2) here are the simple steps which repeat all observed artifacts with all the details as to how to repeat it, and independent evidence from various sources and experiments.
If Hermitian wants to debunk 2) he has to do better than ‘my unknown forger’ for some reasons did better but not really and in fact managed to mimic the Xerox WorkCentre workflow in all aspects… So far NO evidence of a human forger intervention has been identified.
Hermitian: You may want to check out the images that I just posted which reveal dramatically how crappy your images really are compared to expected quality from a 600 DPI x 600 DPI scan to PDF.
Again Hermitian is confused. While the original was indeed scanned at 600×600 PPI, the maximum resolution for scanning, the image was significantly degraded by the MRC compression. So, the first step for Hermitian is to understand what happened when the document was scanned. And although I and others have explained this, it may be helpful to further explain.
The 600×600 PPI image is separated by MRC software into a single background and multiple foregrounds. The background is compressed with JPEG at about a 50% quality level (which is quite high) with the chrominance channels down-sampled a factor of 2 and then downsampled to 150×150 PPI resolution. The foreground is encoded in JBIG2 and down-sampled to 300×300 PPI.
Why Hermitian expect the document to be comparable to a 600×600 DPI(?) scan may need some explanation. Furthermore, the original document was already MRC compressed and of much inferior quality than the original document in possession of the White House. So it comes as no surprise that printing it out, and scanning it again, further degrades the document. Simple really.
Hermitiation: But you see that you have zero credibility on your claim to have recently found out that the quantization tables of the Xerox 7655 WC is now the same as all Xerox WCs. You could start redeeming yourself by posting a formal retraction of the different quantization tables that you initially reported for the Xerox 7655.
Again our poor friend does not understand the timeline. When I started my experiments, I reported that the quantization matrix found in my samples did NOT match the one in the WH LFBC PDF. Why would I have to retract these findings, are they are still factual? What I did find however is that the 7535 and other Work Centre PDF’s all matched the Quantization matrix and thus I compared it to the WH Tax Form scans which had been done on a 7655. And indeed, they showed the same quantization matrix as the WH LFBC PDF. It appears now that Hermitian totally missed these developments and now believes that I recently found out that the matrix matches…
In fact, he is many weeks behind… He blames me for his failures. That’s too bad. Why he wants me to retract my factual findings is beyond me. Science does not work that way. I report the data, positive or negative and define additional experiments to try to explain the observations. The fact that the tables match on the WH 7655 used for scanning the Tax forms of our President, help strengthen my workflow. Note that the exact nature of the scanner is of little relevance, other than that it a Xerox. My prediction of the 7655 as the culprit is based on logical inferences, and is icing on the cake. The same applies to detecting that the document was scanned upside down.
Perhaps Hermitian missed me posting the data for the 7655 Tax forms? Or perhaps our friend has trouble locating them online? Well, that is not exactly my problem now is it? Others have found it trivial to locate these documents. Perhaps you should pay more attention to people like RC, who have contributed immensely to the unraveling of the workflow.
Hermitian: The larger page size of all of NBC’s Xerox scans do not satisfy the 16 x 16 block alignment.
Neither does the WH LFBC my dear friend. Another strawman argument… Well done. But why not focus on the facts. And why there is 8 bit but not 16 bit alignment in 300 PPI. I am sure that eventually you too will come to see the light here.