Educating the Confused – Hermitian and 8 bit alignment

Let’s take small steps… Although I am sure I will have to break it down further

For the WH LFBC, it is clear that the images align at two sides with the 8 bit boundaries. I redid my calculations using illustrator as I assumed this would simplify matters for Hermitian.

Resolution = 300 PPI X 300 PPI / 8 X 8 Blocks
Top and Right Sides Obey 8 MOD 0
Left and Bottom Sides Align with Grid Lines(and also Touch Pixels)

Layer  
N (x,y)         (w,h)    (x+w,y)     ((x+w)/8,y/8)
1 (0,0)      (2552,3304) (2552,0)    (319,0) Background
2 (373,880)  (1819,1454) (2192,880)  (274,110) Mostly Text
3 (1270,2848) (778,199)  (2048,2848) (256,356) Onaka Signature
4 (710,2928)  (274,42)    (984,2928) (123,366) Onaka Date
5 (1836,2160) (228,123)  (2064,2160) (258,270) Reg. Gen. Date
6 (432,2240)  (216,47)    (648,2240)  (81,280) Loc. Reg. Date
7 (1458,1960)  (70,34)   (1528,1960) (191,245) Non
8 (735,2528)  (217,243)   (952,2528) (119,316) Bottom Speckle
9 (1050,32)   (142,132)  (1192,32)   (149,4)   Top Speckle

My findings:

Note how the y distances agree. The difference between my x-distance and Hermitian is that he measures from the opposite side but since the page is aligned at 8 bits this does not matter. For clarity I have added the calculation, as I am not sure Hermitian is up to it. The last column is 2552 – xoff which I underlined. Note how they match our friend’s calculations… x and y are calculated from xpt and ypt according to x=xpt*300/72 and y=ypt*300/72 to meet the 300 ppi coordinate resolution

                                       xoff  yoff           2552-xoff
             xpt    ypt     x      y  2551-x  y+2    /8  /8   x+1
Mostly text 525.84 210.72  2191   878   360   880   80  110   2192
Signature   491.28 683.04  2047  2846   504  2848   63  356   2048
Date Stamp  235.92 702.24   983  2926  1568  2928  196  366    984
Right Date  495.12 517.92  2063  2158   488  2160   61  270   2064
Left Date   155.28 537.12   647  2238  1904  2240  238  280    648
None        366.48 469.92  1527  1958  1024  1960  128  245   1528
Top speckle 285.84   7.2   1191    30  1360    32  170    4   2520

So let’s show an example in easy steps

Mostly text: Obtain the coordinates for the top right corner

mostly text524.84 by 210.72 or 2191 by 878

top right612.24 -0.48 or    or 2551 by -2

Distance 2551-2191 and 878+2 or 360 by 880. Divide by 8 and you get 45 by 110

22 thoughts on “Educating the Confused – Hermitian and 8 bit alignment

  1. How dare you use the anchors! They actually give accurate results, unlike rulers! ROFTLMFAO

  2. For Hermie, since he doesn’t seem to be aware of how to do this:

    Make sure to set up to use Smart Guides first.

    Measurement Labels
    Displays information for many tools (such as Drawing tools and Text tools) about the current position the cursor as you position the cursor over an anchor point. While creating, selecting, moving, or transforming objects it displays the x and y delta from the object’s original location. When you press Shift with a drawing tool selected, the starting location appears.

  3. Did anybody else notice that the purported Xerox added comment YCbCr happens to be placed at the exact end of a string that gets tons of Google hits all over the internet ?

    Yeah, I was wondering why Hermitian was all about rulers when you can just read off the values from the anchors. I assumed he was perhaps as new to illustrator as I am.

    Well, it never hurts to reintroduce him to Illustrator. But of course my calculations are not limited to Illustrator and were double checked by looking at the PDF data.

  4. Educating the Confused – Hermitian and 8 bit alignment

    NBC

    Posted on August 17, 2013 by NBC

    “Let’s take small steps… Although I am sure I will have to break it down further”

    “For the WH LFBC, it is clear that the images align at two sides with the 8 bit boundaries. I redid my calculations using illustrator as I assumed this would simplify matters for Hermitian.”

    So you used Illustrator to measure the WH LFCOLB PDF object boundaries and your results agree with my report which I posted on Scribd over a month ago.

    Well I suppose from your point of view that is progress.

    Now regarding your second table I assume that it is for the Xerox 7655 Preview saved PDF.

    Because you have yet to release this file I won’t be wasting my time trying to figure out your table. You haven’t bothered to define your variables and the column headers are misaligned. Just remember a picture is worth a thousand words.

    I prefer to spend my time comparing the Xerox 7535 Preview saved file to the WH LFCOLB.

  5. Well I suppose from your point of view that is progress.

    What is progress is that you finally understand that our numbers agree…
    What a painful trip that has been, even though the calculations were very straight forward.

    Well, now that we have you finally up to speed with how to properly calculate the distances, we can perhaps address why you claimed that they did not match with 8 bit boundaries…

    You and I have both shown the data that establishes these alignments for the WH LFBC.

    Perhaps you can share with us your 7535 data, since I have also provided these calculations.

    ROTFL, is it not hilarious how Hermitian’s confusions inevitably lead to him having to accept my finding.

    So desperate and yet unable to make a dent…

  6. e. You haven’t bothered to define your variables and the column headers are misaligned.

    Well I defined my variables but I can see why you do not understand this, as to headers being misaligned, I guess that you really cannot figure it out, even though it is trivially simple.

    My bad, I should have prepared my presentation for the level of my audience.

  7. For whatever it’s worth, the headers are aligned properly for me (Win 7, Firefox)

  8. I prefer to spend my time comparing the Xerox 7535 Preview saved file to the WH LFCOLB.

    And yet you have never managed to provide any evidence that you have done any measurements on the 7535 objects, Preview or non-Preview. Despite the fact that NBC has provided his measurements.

  9. I prefer to spend my time comparing the Xerox 7535 Preview saved file to the WH LFCOLB.

    I thought you had already done so… Wow…

  10. Just remember a picture is worth a thousand words.

    I understand that some need more visual support as the math may get too complicated too quickly. All those additions and subtractions, and even a multiplication and division…

  11. NBC

    “So let’s show an example in easy steps

    “Mostly text: Obtain the coordinates for the top right corner

    ‘524.84 by 210.72 or 2191 by 878

    “612.24 -0.48 or or 2551 by -2

    I prefer my method because the results are also graphically verified.

    1. Select the global rulers. This automatically sets the origin point to the upper-left corner of the background layer.
    2. Select / All on Active Artboard
    3. Using the direct selection tool move the origin point to an intersection point of two major grid lines. The major grid lines are set for either 8 x 8 or 16 x 16 blocks. (Sometimes this step is not required.)
    4. Double click the small square at the top left corner of the screen just outside the intersection point of the two rulers to move the 0 mark of both rulers to a coincide with the intersection point of a perpendicular (to each ruler) erected from the upper-left corner of the background and extending out from the corner to intersect each ruler.
    5. When a given rectangular object boundary satisfies the modulo condition then both the top and right sides of the rectangle are congruent with two major grid lines. The two major grid lines intersect at the top-right corner of the rectangle.
    6. Then the x,y and W,H values are read directly from the layers info panel. There are no offsets to worry about.

  12. 6. Then the x,y and W,H values are read directly from the layers info panel. There are no offsets to worry about.

    Good for you. But offsets are trivial to deal with. You do realize that the 8 bits refers to a 300 ppi resolution, not the 72 ppi of the PDF?.

    Given your description, you have forgotten this detail and indeed the alignments will be non existent.

    Cheers.

  13. 1. Select the global rulers. This automatically sets the origin point to the upper-left corner of the background layer.
    2. Select / All on Active Artboard
    3. Using the direct selection tool move the origin point to an intersection point of two major grid lines. The major grid lines are set for either 8 x 8 or 16 x 16 blocks. (Sometimes this step is not required.)
    4. Double click the small square at the top left corner of the screen just outside the intersection point of the two rulers to move the 0 mark of both rulers to a coincide with the intersection point of a perpendicular (to each ruler) erected from the upper-left corner of the background and extending out from the corner to intersect each ruler.
    5. When a given rectangular object boundary satisfies the modulo condition then both the top and right sides of the rectangle are congruent with two major grid lines. The two major grid lines intersect at the top-right corner of the rectangle.
    6. Then the x,y and W,H values are read directly from the layers info panel. There are no offsets to worry about.

    Missed a step there, bub. You need to redefine the grid lines for 300 ppi.

  14. Missed a step there, bub. You need to redefine the grid lines for 300 ppi.

    Oops.

    You can do the following

    Highlight the objects and add a stroke to the outline, make it 1 pixel and red.

    Export as png 300 ppi

    Now you can check that all the top and right boundaries align with 8 bit

    Just checked it.

    Illustrator makes some trivial things quite complicated… Especially if you do not have other tools.

  15. I think if he re-samples at 300 ppi he can then set the grids to pixels. He’ll need to make sure he doesn’t resize it at the same time. Normally I’d advise against re-sampling, but since all of the original image is at 150 or 300 ppi, it shouldn’t cause anti-aliasing effects (shouldn’t – then again, this is Illustrator we’re talking about — and Hermie). Just don’t save the changes!

  16. I did what Vicklund suggested and found again that the borders align quite nicely.

    Did Hermitian really forget this step?

  17. I’ve been wondering all week, but he never gave enough detail to be sure. BTW, please call me Kevin, you’ve earned that. Hermie, not so much.

  18. I’ve been wondering all week, but he never gave enough detail to be sure. BTW, please call me Kevin, you’ve earned that. Hermie, not so much.

    Well, ‘W’, may I call you ‘W’🙂, Hermitian surely is persistent, just not very effective. But he does force me to check and re-check and it is satisfying that all still adds up.

  19. Nobody gets to call me by my first name unless their last name is Vicklund or their immediate family member had Vicklund as a maiden name.🙂

  20. NBC says:

    August 17, 2013 at 23:11

    “”6. Then the x,y and W,H values are read directly from the layers info panel. There are no offsets to worry about.””

    “”Good for you. But offsets are trivial to deal with. You do realize that the 8 bits refers to a 300 ppi resolution, not the 72 ppi of the PDF?.””

    HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
    The 72 PPI is the resolution of the larger page size images that the forger created outside of illustrator. The smaller page size resulted when the forger placed the larger images into Illustrator and then applied the -90 degree rotation and the reduction scale factor of 48% or 24%. The images displayed by Illustrator are the smaller sized images. Consequently, the pixel resolution is either 150 PPI x 150 PPI or 300 PPI X 300 PPI. The 150 PPI x 150 PPI is used for the 8 x 8 Blocks. The 300 PPI x 300 PPI is used for the 8 x 8 blocks and the 16 x 16 Blocks. I thought that you knew that. Your table for the WH LFCOLB 8 x 8 blocks @ 150 PPI is identical to my first table in my report posted on Scribd.com here:

    http://www.scribd.com/doc/151738307/Analysis-of-Rectangular-Object-Boundaries

    If you had actually read my report, you would have known that my analysis utilizes both 150 PPI and 300 PPI resolutions. Consequently there is no missing step in my list. My report describes in detail how the grids are set.
    HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH

    “Given your description, you have forgotten this detail and indeed the alignments will be non existent.”

    HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
    Well you mentioned that your single table was identical to the first of my three tables. Was your single table nonexistent ? Or maybe was it just copied from my report ?
    HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH

    Cheers.

  21. “Given your description, you have forgotten this detail and indeed the alignments will be non existent.”

    Well you mentioned that your single table was identical to the first of my three tables. Was your single table nonexistent ? Or maybe was it just copied from my report ?

    We are talking about your claims about the 7535… Did you really forget when doing the ‘visual’ inspection to properly adjust the resolution… That would explain a lot my friend.

    You made claims about th e7535 not aligning… Now you are stalling… So I have no choice but to conclude that you indeed made the simple error. You are now complaining about ‘resolution’ and ‘grids’ not being reliable in Illustrator and yet you could have done the experiment to show me wrong…

    So what is it Hermitian? Did you forget?

Comments are closed.