Jochen Eisinger wrote:
>
> I've send Eddy De Greef some tools, if someone wants to contribute
> contact him. If there are any results, I'd be glad to include them into
> the mustek_pp backend...
>
The tools are really helpful. I did some experiments during the weekend.
After staring at the hexadecimal codes sent by the Windows driver and
the SANE driver for a few hours, I started to see some correlation.
The scanner seems to respond more or less correctly to the SANE driver,
but the returned images were all white, so I concentrated on the
read_line_101X function.
First I disabled a few statements, and plugged in the byte sequences
sent by the Windows driver. This had no effect (the scanner reacted in the
same way and the images remained white), which makes me believe that the
driver is sending the right commands to the scanner (there are a few
differences between the two driver sequences, but they seem to have
little or no effect).
After a while, I began to realize that the SANE driver was reading some
data, but that they got "normalized" by the driver in the wrong way.
It seems that the black reference value (RefBlack) is very high (251),
and possibly also the calibration values are not correct. After forcing
the black reference value to 0, something became visible: when I put a
small white piece of paper on the scanner (the inside of the cover is black),
I got a "stretched inverted outline" of the piece of paper: the paper appeared
as a black area, and next to the paper, everything was white.
I didn't continue my experiments (it was getting late), but I looks like the
problem might be the calibration of the scanner.
From what I understand from the code, the driver tries to find the black upper
and left edge. However, I wonder whether there are really any black edges.
From the outside, it looks like the edges might as well be lightgray (the
color of the plastic case). I'll open the case this evening to have a look at
it. Moreover, the fact that the inside of the cover is black (all other
scanners that I've seen have a white cover) makes me wonder whether looking
for black edges is the right thing to do. Maybe a black cover is something
specific for CIS-scanners ?
Anyway, I'll do some more experiments during the next few days.
Of course, any help is welcome.
Regards,
Eddy
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Mon Oct 16 2000 - 01:33:42 PDT