Paul Weihe is the per­son in our stu­dio who does every­thing that no-one else does, i. e. quite a lot. Paul knows every­thing about print­ers, printer dri­vers, soft­ware licenses, the net­work. He is also our Barista and can draw com­pany logos and pic­tures of plants onto cups of cap­pu­cino, using frothy milk.

While doing one of his research projects, he came across a rumour, say­ing that var­i­ous man­u­fac­tur­ers printed an invis­i­ble code onto each page that comes out of one of their laser-printers. The US gov­ern­ment had sup­pos­edly requested that. The secret code was said to con­tain the date, down to a sec­ond, and the equipment’s ser­ial num­ber. The EEF.org (Elec­tronic Fron­tier Foun­da­tion) was reported to have cracked that code.

This is the result of Paul’s self-test:

A page A4, CMYK 0, 100, 100, 0, with 12pt Unit type on it, printed on our Xerox DC-12. A detail of 7.5 × 7.5 cm was scanned on our Quato scan­ner at a res­o­lu­tion of 3200 dpi (a 260MB file). With yel­low sat­u­ra­tion all the way up, all colours dark­ened, con­trast increased, this amaz­ing proof appeared on my screen:

The code does exist and is printed yel­low only, so hardly rec­og­niz­able with the bare eye. Blow­ing it up six times after adjust­ing all the other para­me­ters, one can actu­ally not only see a proper table, but also deci­pher it. The columns are num­bered across and the rows show the binary code, which can be deci­phered (the scan was com­bined with the orig­i­nal file in Indesign):

In my exam­ple the code can be read by adding the sums of the dots per col­umn, com­bined with the val­ues on the left (the bot­tom line shows the results per col­umn):
printed on the 22^nd day of the 11^th month at 14:38. The date is exact, the time not really, but that’s prob­a­bly down to them using a com­mon time zone. I don’t think they would have incor­po­rated a radio time clock…

Good to know that we can always prove our author­ship from colour laser prints, even with­out print­ing proper cred­its. I’m pretty sure that the ser­ial num­ber in columns 11 – 14 will have been encoded correctly.