rex@PMGVAX.WR.USGS.GOV (Rex Sanders) (06/20/90)
> I tried to print a perfect square on an Apple LaserWriter II NT using ... (several lines of reasonable PostScript code deleted) > But the square that I got was distorted. It is approximately 0.5 mm longer > in the y direction (11" direction) and 0.5 mm shorter in the x direction > (8.5" direction). It is essential for us to get an undistorted printout. Various brands of laser printers I've played with for the last 6 years have some kind of distortion like you describe. You must remember that most laser printers are basically copying machines with some fancy electronics and lasers added. Have you seen *any* copier that did not distort your original? The printing process used by laser printers & copiers is essentially analog, with lots of belts, rollers, etc., that are subject to slippage. Try printing a 20 mm grid across the page, and measuring each cell. You will find the variance from cell to cell is as great as the variance from "perfection"! We do lots of maps on laser printers, and have warned users not to use the maps for accurate measurements. You can also measure the average distortion across the page, and put fudge factors in your plotting software. However, the fudge factors will vary from printer to printer. Laser printers are wonderful devices, but don't depend on them for high-accuracy graphics. -- Rex Sanders, US Geological Survey rex@pmgvax.wr.usgs.gov
BNB@MATH.AMS.COM (bbeeton) (06/23/90)
on 30 may, andy sun anu-guest asked about distortion suffered on an apple laserwriter ii nt when he tried to print a square. i'm neither a postscript programmer nor a regular laserwriter user, but recognizing distortion is important to us at the american math society when we are implementing document designs in tex. we discovered that the old laserwriter guarantees no more than 2% distortion; that's about 3pt in 2in, and the greater the distance, the more noticeable the discrepancy. to determine the exact behavior of the particular machines we're using, we developed a test in which several rules run horizontally and a similar group of rules vertically. (the pattern we chose is 4 parallel rules of nominal 3, 4, 5 and 6in.) we then print the same file in both portrait and landscape mode, and lay them one on top of the other on a light box to see what's happening. this doesn't solve the problem, but it does very quickly let us know the degree of variation. and if it's more than 2%, we can yell at the hardware maintenance folks. the only device we've had come in right on the money is our typesetter (an aps micro-5). a machine based on the xerox xp-24 engine can be tuned so that it's under 1 pixel off in 6in; unfortunately, it's a write-white engine and has worse problems on that account. a couple of ricoh engines do reasonably well. sad to say, the laserwriter is the worst of the lot, and 2% is really an unsatisfactory degree of variation for graphic arts purposes. but maybe, after you find out what the reality of your machine is, you can find a hardware competent who's willing to tune it tighter than the promised 2%. good luck. -- bb -------