At last the ISO is beginning to hammer out standards for calculating printer specifications. Eventually, print speed and toner life figures will actually have a bearing on real-world performance.

The full spectrum of human deceit can be split into three levels of gravity: lies, damned lies and printer specifications. The last-mentioned is a familiar bugbear when testing printing products because the real-world performance of laser, inkjet, thermal, ribbon and dot-matrix printers never matches the official specification. Often it doesn't even come close. An office laser printer rated at 20ppm is unlikely ever to exceed 18ppm in real life, for example. Worse, a home inkjet rated at 6ppm will probably struggle to achieve half that speed.

Given that these official specifications are clearly nonsense, what are they for, where do they come from, and what industrial standards of measurement are they based on? First answer: to mislead. Second answer: the printer manufacturers. Third answer: none whatsoever.

The numbers game
If you speak to a technical product manager, it will be explained to you that the official speed rating of a printer refers to raw engine speed. That is, it's the maximum speed at which the printer can pass one sheet of paper through the machinery from input tray to output tray. The figure does not include any other process, such as sending the print job from computer to printer or even the couple of seconds it might take to image the drum in a laser printer. In other words, it's the speed at which a bunch of rollers and cogs turn round; it has nothing to do with the speed at which something is actually printed.

Ironically, it was an HP inventor who coined the expression 'the numbers game' for dubious product specifications like that. HP's digital-imaging specialist Bob Gann has often expressed his frustration that competitors quote misleading figures for their scanners' resolution, quality and speed. Manufacturers know that consumers like to compare products using numbers, so it's important for them to fill product specifications with bigger numbers than everyone else's. As a result, manufacturers feel compelled to spend money increasing the 'dpi' measurement of a scanner to ludicrous levels at the expense of the hardware that matters. In most cases, a 1200dpi scanner with good optics will produce superior images to a 3200dpi scanner with a poor lens-and-mirror apparatus, but heck, 3200dpi sounds so much better than 1200dpi, doesn't it?

The numbers game is a favourite with marketing departments at printer manufacturers. Not only is the engine speed of a printer less than useful (because it doesn't tell you the speed of printing, remember), it can be deliberately misleading too: sometimes the per-page handling speed is measured in US Letter sheets instead of A4. For a reasonably fast network laser, this can be enough to add an additional 2ppm to the figure. Also, some manufacturers quote speed for draft mode instead of normal printing mode.

Now consider toner or ink cartridge life. Who calculates how many copies you can print from a cartridge before it is depleted? What does '5% coverage' refer to, and does it bear any relationship to real-world printouts? Every manufacturer uses different testing criteria, so there's no way of making a fair comparison between products. What buyers need is for the manufacturers to adhere to...

A new standard
Enter the International Standards Organisation (ISO). At last, this body has decided to replace the arbitrary and sometimes dodgy printer specifications with ones that will be respected across the computing industry. The first fruit of its endeavours is the not-so-catchily named ISO/IEC 19752:2004(E). This international standard describes a strict process by which you can determine the page yield for toner cartridges in mono laser printers. It has three elements:

  • a standard test
  • a standard method for calculating page yield values from the test results
  • a standard method of describing the page yield to consumers

The test involves printing a specially designed document over and over again until the toner runs out. The document is not in the public domain - it must be purchased from the ISO - but it looks like this...

As you can see, it's a mock-up of a business letter with some line-art and greyscale elements thrown in. Already the nickname 'LSA chart' has been coined for the test document, based on the fictitious company name on the letter.

What really matters, however, is the test method. There are strict guidelines to help the tester determine when a toner cartridge is depleted, including instructions on how to use the 100th printout as a 'fade reference' and explanations of how to detect 'fade' itself. The tester must run the toner depletion routine on at least three identical printers using at least three consecutive cartridges per printer, making a minimum of nine tested cartridges. The ISO standard explains what to do in the case of defective cartridges and printer errors, and even covers issues such as cartridge shaking and dealing with obstructive 'toner low' warnings.

The page yield value which manufacturers are allowed to quote is calculated by combining the sample average with a sample standard deviation, producing a result that is apparently offered 'with 90% confidence'. If all laser printer manufacturers quote a page yield figure according to ISO/IEC 19752, you know you're comparing like for like.

Long time coming
Now you may be cynical and wonder why it has taken the ISO so long to come up with such a fundamental standard. After all, laser printing has been around 20 years, and all the ISO has done in this time is invent a business letter that you print out lots of times. But remember that the ISO is not really an independent organisation; rather, it is a large group of interested parties, including many printer manufacturers. So the very fact that they have finally seen the need for standardisation in product specs should be seen as a positive step. The ISO is now working on other versions of the test for colour printers, and is said to be looking at an appropriate standard for defining printer speeds.

HP, for one, says it will begin quoting the ISO/IEC 19752 standard page yields for its next generation of printers and hopes its competitors will do the same. That said, it will continue to quote a second figure based on its old, proprietary method of measuring page yield according to that mysterious '5% coverage'. We are told that maintaining both sets of specifications will help avoid confusion, although this may have been a joke.

What will be interesting to see is whether the ISO/IEC 19752 figure ends up higher or lower than the old page yield rating. If the ISO standard produces a higher number, you can be sure the old test figure will be dropped from spec sheets pretty quickly. If lower numbers are the result, which may well be the case for colour output and print speeds after the introduction of the next round of standards, we imagine the old, misleading but higher figures will be around for a long time to come.