Xerox scanning bug update

Xerox scanning bug update

Summary: The company has identified the exact bug and the products that are affected and is working on a patch. Contrary to earlier reports, all configurations of affected products may exhibit the bug.

SHARE:
TOPICS: Software
1

Since my initial report on August 6 of a scanning bug in certain Xerox products the company and outsiders have been working to clarify the problem and provide guidance to customers.

The bug caused scans of documents with small numerals in them to have the numbers changed in the scanned image. For example, the paper version might have a '6' but the scanned document would show an '8'. 

The problem was initially found by German computer scientist D. Kriesel, who reports that Xerox has worked hard on the problem and kept him in the loop. According to Kriesel '…the issue indeed affects all compression modes across a whole lot of devices, all are able to mangle numbers'.  He adds that they have identified the bug, that a patch is being developed, and that they want him to test it with his originals that identified the bug in the first place. Smart move.

Xerox has produced documents for customers on the matter. In this entry of their Real Business at Xerox blog, Rick Dastin, corporate vice president and president, Office and Solutions Business Group, Xerox reveals that the problem may manifest in all settings of the products. They are working on a patch to fix the bug.

The company's initial responses indicated that the problem only happened with quality settings lower than the defaults for the product. The problem is indeed most likely to occur with 'stress documents' ('documents with small fonts, those scanned multiple times and hard to read').

The company says that this is a complete list of affected products:

  • ColorQube: 87XX, 89XX, 92XX, 93XX
  • WorkCentre: 5030, 5050, 51XX, 56XX, 57XX, 58XX, 6400, 7220, 7225, 75XX, 76XX, 77XX, 78XX
  • WorkCentrePro: 2XX
  • BookMark: 40, 55

Xerox produced a FAQ document entitled Xerox Scanning Update: What You Need To Know which adds several other useful data points:

  • Only scanning is affected, not copying or faxing
  • The 'character substitution' bug is unlikely to occur when the products are configured with default settings
  • The affected component in the scanning software is the JBIG2 compressor. JBIG2 is an industry standard for image compression (click here for the specification). Character recognition is part of the spec and one of the reasons for the high compression rates.
  • The patch will be available 'within a few weeks'.

Click here for instructions on how to reset your Xerox device to factory defaults.

Topic: Software

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

1 comment
Log in or register to join the discussion
  • Something Seriously, Seriously Wrong Here

    You expect photocopiers to be dumb image replicators. OK, they may do some simple contrast enhancement to maintain readability of text, but that's about it. When they start using compression encodings that depend on correctly recognizing the shape of those text characters, then you have to start to worry. With a dumb copier, errors in replication manifest in obvious degradation of the image (noise, contrast etc), which is a hard-to-ignore warning sign that the copy is perhaps not entirely trustworthy. But when a more sophisticated compressor makes an encoding error, you end up with an image that is perfectly clear, legible--and wrong.
    ldo17