greyscale , it dosen't look good as b/w.
I scan at 300 dpi monochrome and convert to .pdf files and I usually
get a very readable datasheet that only takes about 50k-100k bytes
a page. I use the Xsane driver scanner program under Linux, outputting
Postscript and then distill the postscript to .pdf using the Ghostscript
ps2pdf script. In tough cases, (Like old National databooks that are
old, yellow and have a lot of print-through from the other side of the
page) I have to scan as grey scale and convert to a monochrome using
threshold function in The GIMP image manipulation program, manually
finding the brightness range of black or white pixels.
The key is to use an "indexed" graphics format (gif, postscript, png)
where there is a small pallette of colors/brightness, and have as
few colors as possible. These formats are run length encoded where,
if adjacent pixels are the same color, they just keep a count of them
and the size of the data file depends on the number of changes in color.
Having a greyscale picture with a lot of little variations in brightness
balloons the file size right up there.
The .pdf format seems to yield the smallest monochrome files. .Png
are a bit bigger. Then .gif, by a factor of about two or so.
The programs for viewing .pdf files seem to have the best image
scaling in them. Viewing a GIF on anything but some integer scaling
factor can look real ugly if the people who write the viewer software
didn't spend much effort on that.
Examples at ftp://ftp.eskimo.com/u/m/mzenier/
Mark Zenier
[email protected] Washington State resident