[NewsPoster]

A handful of Xerox devices have been found to randomly substitute characters while performing a copy action, but not an optical character recognition (OCR) analysis. Confirmed by experiment, both the Xerox WorkCentre 7535 and 7556 perform the swap, with a possible eight other devices also by Xerox manifesting the issue. The researcher who found the problem discovered that "patches of the pixel data are randomly replaced in a very subtle and dangerous way: The scanned images look correct at first glance, even though numbers may actually be incorrect."

According to reseacher David Kriesel, "the error does not occur if PDFs are scanned with OCR, or TIFs are scanned (the latter seems plausible, as the pure image data should be saved into the TIF). Additionally, there seems to be a correlation between font size and scan dpi used. I was able to reliably reproduce the error for 200 DPI PDF scans without OCR, of sheets with Arial 7pt and 8pt numbers."

Since original discovery, the error has been linked to overzealous compression within the scanner and printer combination. The JBIG2 algorithm, when used in "normal" mode (but not higher levels) has been found to make the substitution during copy or document saving operations when OCR is not being used.

The error is beyond a simple "8 for 6" exchange as seen in the third image below, as the JBIG2 routine "creates a dictionary of image patches it finds 'similar.' Those patches then get reused instead of the original image data, as long as the error generated by them is not 'too high'." Xerox confirmed the problem with the researcher in a conference call a few days after the discovery, and the substitution effect is seen in the second image below.

Responding to the issue, Xerox has said that the default print quality of "higher" does in fact prevent this issue from manifesting itself. In a statement, the company claims that "for data integrity purposes, we recommend the use of the factory defaults with a quality level set to 'higher.' In cases where lower quality/higher compression is desired for smaller file sizes, we provide the following message to our customers next to the quality settings within the device web user interface: 'The normal quality option produces small file sizes by using advanced compression techniques. Image quality is generally acceptable, however, text quality degradation and character substitution errors may occur with some originals.'" If the resolution is set at the printer at the time of the scan, the alert is not given, however.

The eight other models reportedly having the issue are the Xerox WorkCentre models 7530, 7328, 7346, 7546, 7535, and the 7556. The Xerox ColorQube 9203 and 9201 are also allegedly manifesting the problem, according to reader reports.

[SierraDragon]

Wow. Billion-dollar problem and Xerox is just ho-hum about it. Wrong attitude. Way wrong attitude.

They should be in emergency recall mode on affected machines, like a car with brakes that may fail unless users push them just right.

[pottymouth]

"...the error does not occur if PDFs are scanned with OCR, or TIFs are scanned..."

What? Is he perhaps talking not about what is BEING scanned, but the format that the scan is being saved to? Because as it is, that quote makes absolutely no sense. Perhaps "...the error does not occur if the scan is saved to PDF with OCR, or saved to TIFs..."? THAT might make sense.

A low quality original scanned with a low quality scanner and then compressed is going to yield low quality results. And now you want that to work with 7pt text? Let me guess: you printed that 7 pt text from an inkjet printer on generic copy paper?

Whatever. This is a non-story.

[SunSeeker]

Yeah. It's a non story until we hear of a major disaster due to an incorrect number entered from one of these photocopies.

[Makosuke]

Sorry, I can't call that a non-story since it's doing it in copy mode. When you put something on the platen of a copier and push the button, you have come to expect the same thing to come out the other end so long as it's readable at all.

The swapped bits of small text are at least tiny; the transposed 6 and 8 characters are large and clearly readable, so there'd be no reason to suspect that something was off if you were handed a copied document with those on it.

[Sebastien]

Wondering what this has to do with Macs/Apple (given that it's showing up on MacNN)

[Charles Martin]

Sebastien: are you seriously suggesting that stories about printers have no connection to Macs or Apple? Really?

[bjojade]

It would be a non issue if the resulting copy was just an illegible result, but if the actual DIGITS in a number are being changed, that's a pretty big deal, even if it's small print.

Major flaw in their compression formula if it's letting that happen.

[sessamoid]

Imagine if bridges, skyscrapers, or planes are being built on specifications copied on these copiers. Sound more important now?