The Museum PDF

The document is an 11×17 scan, in “portrait”, so no messy rotations but several features have been checked. The YCbCr comment string is, once again, present. The Quantization Matrices are the same and the foreground images align. The document also shows halos, x-ray effect and identical pixel elements.

museum_layers

Monochrome layers

Monochrome layers

20 thoughts on “The Museum PDF

  1. With so many Xerox MRC PDFs on the web, your could automate the whole examination process, and set a crawler off, collecting examples, building a database.

    I guess it’s fun to beat the birfers over the head, and elegant solutions, confirmations, vindications, are always satisfying … but it’s still hard to believe someone has had to go so far … just to “prove” something is perfectly ordinary!

  2. Forgot to add the Version is 1.6 and the 8-bit were filtered with FlateDecode/DCTDecode and the 1-bit were filtered with JBIG2Decode

  3. Oh yeah and a 7525 Xerox

    This is great.. Confirmation after confirmation and in the mean time Hermitian is struggling.

  4. NBC says:

    August 19, 2013 at 05:31

    “Oh yeah and a 7525 Xerox

    “This is great.. Confirmation after confirmation and in the mean time Hermitian is struggling.”

    Yes I’m struggling with why you haven’t posted your “trivial” explanation of the two unique strings that I found in the WH LFCOLB PDF. And the weird part was that the YCbCr comment always appears right at the end of one of these two strings — even when the strings are embedded within a Huffman Table.

    And I’m sure that you remember that I also found (with Google search) that these two strings are posted all over the internet. And that most of these postings have to do with problems that various code developers have had with uploading images or text files under PMP and the GD graphics utility.

    But I’m sure you are right and your “trivial” explanation is bulletproof.

    But I am curious why you haven’t yet Google searched your weird MAC OS translation of these same two strings to see if you get any hits Google at all.

    And maybe you should remember that I still have that Obama LFCOLB which has these same two strings but does not have the YCbCr label.

    Just saying ….

  5. NBC

    “The Museum PDF

    “Posted on August 19, 2013 by NBC

    I admit that the Museum would be a good place to put the Xerox forger to the test. Or maybe a good place to just put the Xerox forger.

  6. the strings are embedded within a Huffman Table.

    There you go! At least you’re on the right track! The string is part of one of the general-purpose Huffman tables listed in the JPEG standard. Every JPEG has a Huffman table (actually, minimum of two). So the reason it’s seen all over the place is because people tend to use the general-purpose Huffman tables when writing software. And sometimes, when they try to embed an image, they do it wrong, and they get an error message with the binary stream. See, trivial explanation!

    As to why the YCbCr comment comes after a Huffman table, the question might be better phrased as, what does the YCbCr string precede? The next chunk of data starts with FFC0, which is the Start Of Frame (Baseline DCT). According to the wikipedia article on JPEGs, this indicates that this is a baseline DCT-based JPEG, and specifies the width, height, number of components, and component subsampling (e.g., 4:2:0) – in other words, the size of the image (in pixels) and the color components. So the comment listing the colorspace precedes the marker that contains the information relevant to the colorspace. Like the Huffman tables, the Start of Frame marker is a required component of any JPEG, and Xerox decided to place this marker after the Huffman tables.

  7. BTW, the structure of a Xerox WorkCentre jpeg:

    Start of Image, Quantization Table 00, Quantization Table 01, DC Huffman Table 00, AC Huffman Table 00, AC Huffman Table 01, Colorspace Comment, Start of Frame, Start of Scan, [image data], End of Image

  8. Yes I’m struggling with why you haven’t posted your “trivial” explanation of the two unique strings that I found in the WH LFCOLB PDF. And the weird part was that the YCbCr comment always appears right at the end of one of these two strings — even when the strings are embedded within a Huffman Table.

    I was hoping that you could figure it out. Yes, you are on the right track, there is a Huffman table preceding the comment tag… Surely you can put 2 and 2 together now.

    And maybe you should remember that I still have that Obama LFCOLB which has these same two strings but does not have the YCbCr label.

    Just saying ….

    Yes, that does not surprise me, and neither should it surprise you…

  9. Start of Image, Quantization Table 00, Quantization Table 01, DC Huffman Table 00, AC Huffman Table 00, AC Huffman Table 01, Colorspace Comment, Start of Frame, Start of Scan, [image data], End of Image

    ROTFL, at last someone who understands jpeg formats…

  10. I admit that the Museum would be a good place to put the Xerox forger to the test. Or maybe a good place to just put the Xerox forger.

    And yet, another verification of the work flow. All Hermitian can do is attempt ‘humor’ 🙂

  11. And maybe you should remember that I still have that Obama LFCOLB which has these same two strings but does not have the YCbCr label.

    But I thought your experts said that non-Xerox pdfs would also have the ‘YCbCr’ comment… Maybe they aren’t so ‘expert’ after all?

  12. I wonder why Hermitian is undermining his own claims…

    And he is soooo close to understanding the strings and yet, he cannot make the final step…

Comments are closed.