My HP C6270A works with scan-hp driver.
However, there are a couple of problems I noticed:
1. The first time I ran the "xscanimage" it took five minutes
each time I scanned a page. I decided that the software
was junk, since the scanner is suppose to scan a page in
15 seconds. However, when I went to show someone how badly
it worked, it scanned in the 15 seconds.
Later I figured out why: My home directory is NFS mounted,
and unlike the Windows version which scans and then saves,
xscanimage waits for the data to be written. Solution:
Documenting this "feature" should be sufficient.
2. The quality I get is random. The first time I used it
I was getting all sorts of gray "dirt" like I had smired
ink on the page. The next time (after a reboot) I kept
getting rainbows across the page, or dark bands in grey
mode. After a third reboot, I began getting nearly perfect
scans.
Solution: If this is a bug in the software, then the
best solution is to fix it. If this is a problem with
the scanner, then there needs to be a way to reset the
scanner other than rebooting my machine. (My Linux box
acts as a server, so I can't reboot regularly...) I
tried things like "rmmod gs.o" but that doesn't reset
the scanner... I also tried just unplugging the scanner,
but then I can't access the scanner at all until I reboot.
3. If there is paper in the auto document feeder when I start
the "xscanimage" program for the first time, it decides my
pages have zero lines. (This makes for very fast scans of
zero size files...)
4. Is there a way to specify an output program? e.g.
|cjpeg -q 100 > foo.jpg
as the output program? Since the scanning seems to be
slowed by IO time, this might be even faster than just
writing a huge PNM file.
Or perhaps I might want to try something like scanning
at 1200 DPI and using
|pnmscale 0.25 > foo.pnm
to see I it improves the quality.
5. Is there a way to tell the program to scan all the pages
in the document feeder?
Currently I'm involved in a project designing a "scanning house" for
making a catalogs.att.com website that will hosts catalogs for very
low rates. This means they'll need to be able to scan approximately
1.5 million pages per year. Clearly the 6270 isn't going to be up
to this task, considering I'd dread just scanning in a few catalogs
for test purposes. However, it is the scanner I'm using to develop
the post-scanning filter software, and if we are lucky, the "sane"
drivers will be used under Linux, instead of some Windows NT program
we have no control over. To do this, I'll probably have to code
things to use the sane.so files directly. Is there any documentation
for the API interface, or will I just have to study the source
code?
Thanks,
Bill
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com