jlo@elan.UUCP (Jeff Lo) (05/19/88)
I have the following piece of code that I sent to our LaserWriter Plus:
%!
0 setlinecap
.24 setlinewidth
newpath 72 72 moveto 360 0 rlineto stroke
showpage
From this I would expect to get a 5 inch long horizontal line. However, it
is actually about 1/32 inch too long. Is it too much to expect better than
1/32 inch (~10 pixels or ~0.5%) accuracy, or is this known to be what happens
in a LaserWriter Plus and perhaps other printers as well?
--
Jeff Lo
..!{ames,hplabs}!elan!jlo
Elan Computer Group, Inc.
(415) 322-2450cdl@mplvax.nosc.MIL (Carl Lowenstein) (05/20/88)
In article <313@elan.UUCP> jlo@elan.UUCP (Jeff Lo) writes: . . . postscript code deleted . . . >From this I would expect to get a 5 inch long horizontal line. However, it >is actually about 1/32 inch too long. Is it too much to expect better than >1/32 inch (~10 pixels or ~0.5%) accuracy, or is this known to be what happens >in a LaserWriter Plus and perhaps other printers as well? It seems that the average Canon printer engine is only nominally 300 dots per inch. All the ones I have seen with Imagen labels have been more like 297 x 302. The ones with Apple labels have been something else, I forget what. But only within +/- 1% or so. One would like to believe that this is adjustable, somewhere in the hardware. -- carl lowenstein marine physical lab u.c. san diego {ihnp4|decvax|ucbvax} !ucsd!mplvax!cdl
greid@ondine.COM (Glenn Reid) (05/21/88)
In article <313@elan.UUCP> jlo@elan.UUCP (Jeff Lo) writes: . . . postscript code deleted . . . >>From this I would expect to get a 5 inch long horizontal line. However, it >>is actually about 1/32 inch too long. Is it too much to expect better than >>1/32 inch (~10 pixels or ~0.5%) accuracy, or is this known to be what happens >>in a LaserWriter Plus and perhaps other printers as well? > >It seems that the average Canon printer engine is only nominally 300 dots >per inch. All the ones I have seen with Imagen labels have been more >like 297 x 302. The ones with Apple labels have been something else, I >forget what. But only within +/- 1% or so. One would like to believe >that this is adjustable, somewhere in the hardware. I believe that this is actually due simply to the mechanical difficulty of feeding paper at a precise rate. Even different thicknesses of paper can make minute differences as the paper is being fed through the marking engine. It is not that the "resolution" is 297 dots per inch. It is the process of transfer, I believe. The PostScript interpreter has a 300 dpi frame buffer which it happily feeds to the marking engine. The bits in the frame buffer are, indeed, 5 inches long. Beyond that, it is probably just mechanical "noise". Glenn Reid Adobe Systems
rcd@ico.ISC.COM (Dick Dunn) (05/21/88)
> >From this I would expect to get a 5 inch long horizontal line. However, it > >is actually about 1/32 inch too long. Is it too much to expect better than > >1/32 inch (~10 pixels or ~0.5%) accuracy,... ... > It seems that the average Canon printer engine is only nominally 300 dots > per inch. All the ones I have seen with Imagen labels have been more > like 297 x 302... Even if the print engine could be absolutely precise, you have to take account of expansion and contraction of the paper--mostly due to humidity. Normal "copier" paper can easily vary by 0.1" or so in length, which is about 1%. With that much variation, (a) what's the point in trying to make the print engine more precise, and (b) how can you tell, anyway? In particular, consider that the printer heats up the paper considerably just fusing the image--that can shrink the paper a few hundredths (by drying it). -- Dick Dunn UUCP: {ncar,cbosgd,nbires}!ico!rcd (303)449-2870 ...Never attribute to malice what can be adequately explained by stupidity.
jad@dayton.UUCP (John A. Deters) (05/24/88)
In article <313@elan.UUCP> jlo@elan.UUCP (Jeff Lo) writes: . . . postscript code deleted . . . >From this I would expect to get a 5 inch long horizontal line. However, it >is actually about 1/32 inch too long. Is it too much to expect better than >1/32 inch (~10 pixels or ~0.5%) accuracy, or is this known to be what happens >in a LaserWriter Plus and perhaps other printers as well? Talk to a printer. You'll find out that one point is not exactly* 1/72 of one inch (pretty darn close, tho.) One point is 1/72.27 of an inch. (See the red book for all the truly gory details.) -- -john deters Dayton Hudson Department Store Company uucp: rutgers!dayton!jad MIS 1060/700 on the Mall/Mpls, MN 55402 So many pedestrians ... WARNING: Do not leave these opinions ... so little time ... lying around -- they attract flies.
mark@hpcvlx.HP.COM (Mark Rowe) (05/26/88)
>>From this I would expect to get a 5 inch long horizontal line. However, it >>is actually about 1/32 inch too long. Is it too much to expect better than >>1/32 inch (~10 pixels or ~0.5%) accuracy, or is this known to be what happens >>in a LaserWriter Plus and perhaps other printers as well? > >Talk to a printer. You'll find out that one point is not >exactly* 1/72 of one inch (pretty darn close, tho.) One >point is 1/72.27 of an inch. (See the red book for all >the truly gory details.) The fact that one point is actually 1/72.27 of an inch will not explain any inaccuracy in output of a PostScript file. In PostScript the basic unit is exactly 1/72 of an inch independent of rather or not this is actually one point. Mark Rowe Hewlett Packard Corvallis, Oregon mark%hpcvlx@hplabs.HP.COM