Probable Bugs & Suggestions

A. Leidenfrost (
Thu, 18 Dec 1997 16:08:23 +0100 (CET)


Hi everybody,

we recently bought a scanner in my group and are delighted to be able to use
such neat stuff like SANE to access it under Linux. Thanks a lot!

To make it even better here are some probable bugs:

o the scan resolution is not saved correctly, i.e. when setting it to 200 dpi
it is set to 198 after a new start of xscanimage

o setting the value for Exposure to -13 or less results in an error (besides 0
is 0 when coming from positive values and -0 when coming from negative ones)

o trying to change any parameter during a preview or a scan results in both
xscanimage and the scanner to go nuts. There is a error popup and the
following error message is repeated eternally on the terminal where
xscanimage was started:

Error: Failed to set value of option brightness: Device busy.

o when using xscanimage as a gimp plugin the transfer of the image from
xscanimage to gimp sometimes doesn't seem to work. In fact, nothing seems to
happen and monitoring the process list with top reveals that xscanimage is
growing and growing and xscanimage and gimp use nearly the whole CPU. (When
xscanimage reached 23 MB I killed it, and that was for a fairly small
image of about 5 MB.)

and suggestions:

o what about using separate tab folders in the scan dialog for each area
surrounded by a line. This would allow for a smaller scan window plus the
user could better focus on the kind of parameters he actually wants to
manipulate. For instance you probably often only need to change resolution
and mode, so all other parametes need not to be visible.

o Automatic Color Calibration: Ok, although I know nearly nothing about
scanning and image processing, here is my thoughts about color calibration
for a certain printer. Suppose you have an image with three linear gradient
bars for the three primary colors red, green and blue from full color to
black. Print this image on the desired printer. Take the printout and scan
it in 'Calibration Mode'. Calibration Mode compares the reference with the
scanned gradient bar separately for each color and computes three gamma
corretion curves one for each primary color. The set of gamma curves can
then be saved under the name of the printer.
What do you think about this? Would it work? What about an
additional greyscale bar for brightness correction? Would it be hard to
I got this thought because the colors of our scans are pretty wrong and we
really have problems correction them by hand.

Finally, the environment:

o Pentium 133Mhz, 64 MB Ram running Debian-1.3.1r6
o Microtek ScanMaker E6
o GTK 0.99.0
o SANE 0.68
o Microtek-0.4
o Gimp 0.99.16

Thanks a lot for your time.



- ---
Andree Leidenfrost, GEOPHYSICIST

Institute of Geophysics, Hamburg University
Bundesstrasse 55, 20146 Hamburg, Germany
phone +49 40 4123 2984, fax +49 40 4123 5441

finger for PGP public key
PGP fingerprint = 5F FC 44 FF 84 71 C4 46 72 64 9A F0 EF 99 89 7C

Version: 2.6.3ia
Charset: noconv


Source code, list archive, and docs:
To unsubscribe: echo unsubscribe sane-devel | mail