km@mathcs.emory.edu (Ken Mandelberg) (07/07/90)
Does anyone have a really fast way to print full page bitmaps on a
postscript printer?
I've done a little experimentation with the two obvious methods. Doing
a direct "readhexstring" into "image", seems to run essentially at the
baudrate. On the other hand, using run length encoding with
readhexstring drastically decreases the communication cost, but seems
to greatly increase the computation time in the printer. (I may be
fooling myself here, since the communication time and computation time
may be largely overlapping in the unencoded case).
For example, on a full page 200 dpi image containing a lot of white
space, I sent 475K of "postscript" to the printer, while only 33K with
encoding. On a 19.2K baud NTX it took 266 seconds to send and print the
unencoded image, and 108 seconds for the encoded image. 248 of the 266
is communication time, while 17 of the 108 secs was communication time
in the encoded case.
Clearly the encoding helps, but it would really be nice to get the
speed down a lot closer to the communication speed. Run length
encoding is very simple, and wouldn't really seem to add much to the
computation.
The encoding I used was out of the pgmtops filter with the pbmplus
package, and is only a few lines of postscript.
--
Ken Mandelberg | km@mathcs.emory.edu PREFERRED
Emory University | {rutgers,gatech}!emory!km UUCP
Dept of Math and CS | km@emory.bitnet NON-DOMAIN BITNET
Atlanta, GA 30322 | Phone: (404) 727-7963