In my discussions with Hermitian, he reminded me of my analysis of the JPEG data, especially the quantization matrix. I ran the tool on the extracted jpeg from the original White House document and the scanned xerox and I was happily surprised that the quantization matrices matched. I am such an idiot. This was the first fingerprint I was looking for but totally forgot about it. Note that Adobe for instance has very distinctive finger prints for its jpeg quantization matrix and the do look nothing like the ones below…
I’d say another nail in the coffin… But also a minor nail in me identifying 7655 as the culprit as it has very different quantization matrices. So we now have narrowed it to a 7535. Too bad, I had found the 7655 Xerox Work Centre… Now I have to see if they also have a 7535 work centre.
whitehouse pdf jpeg
Quantization table
Precision=0; Table index=0 (luminance)
8 10 11 13 11 9 14 13
12 13 16 15 14 17 22 36
23 22 20 20 22 44 32 33
26 36 52 46 55 54 51 46
50 50 58 65 83 70 58 61
78 62 50 50 72 98 73 78
86 88 93 94 93 56 69 102
109 101 90 108 83 91 93 89
Estimated quality level = 47.48%
Quantization table
Precision=0; Table index=1 (chrominance)
8 10 11 13 11 9 14 13
12 13 16 15 14 17 22 36
23 22 20 20 22 44 32 33
26 36 52 46 55 54 51 46
50 50 58 65 83 70 58 61
78 62 50 50 72 98 73 78
86 88 93 94 93 56 69 102
109 101 90 108 83 91 93 89
Estimated quality level = 47.48%
Average quality: 47.48% (47%)
|
Xerox scanned PDF jpeg
Quantization table
Precision=0; Table index=0 (luminance)
8 10 11 13 11 9 14 13
12 13 16 15 14 17 22 36
23 22 20 20 22 44 32 33
26 36 52 46 55 54 51 46
50 50 58 65 83 70 58 61
78 62 50 50 72 98 73 78
86 88 93 94 93 56 69 102
109 101 90 108 83 91 93 89
Estimated quality level = 47.48%
Quantization table
Precision=0; Table index=1 (chrominance)
8 10 11 13 11 9 14 13
12 13 16 15 14 17 22 36
23 22 20 20 22 44 32 33
26 36 52 46 55 54 51 46
50 50 58 65 83 70 58 61
78 62 50 50 72 98 73 78
86 88 93 94 93 56 69 102
109 101 90 108 83 91 93 89
Estimated quality level = 47.48%
Average quality: 47.48% (47%)
|
” But also a minor nail in me identifying 7655 as the culprit as it has very different quantization matrices.”
Have you compared the quantization matrices of the WH LFBC with those for the 2010 tax returns?
Let me check, I believe they did not contain jpegs…
I’ll be darned..
I may have set the compression to high on my xerox scan… Need to redo some experiments 🙂
This is cool…