[sane-devel] Epson 1250u vs plustek-45-TEST5

Gene Heskett gene_heskett@iolinc.net
Tue, 7 Jan 2003 17:45:08 -0500

On Tuesday 07 January 2003 12:56, Jaeger, Gerhard wrote:
>On Monday, 6. January 2003 20:37, Gene Heskett wrote:
>> Post a notice here when it is please :-)
>> >> I don't think this is quite "Ready for prime time" :-(  We
>> >> still have that white vertical line artifact, and the
>> >> segfault at the end of the scan seems to be somewhat of a
>> >> coin toss.  Phase of the moon, odd/even minute, whatever.
>> >
>> >- phase of the moon! I didn't get any segfaults here after a
>> > scan.
>> Now wait a minute, I don't want any MORE poltergeists in my
>> system! With kwin crashing if I'm playing a game and actually
>> win, I don't need any more ghosts.  I've also downloaded the
>> latest bugzilla, and installed 20 megs worth of stuff from CPAN
>> that it says it needs, but I still can't file a bugreport
>> against kwin and there does not appear to be any humans at
>> bugs.kde.org to override things.
>> Thats O.T. here, so back to sane/plustek:
>> Whats the appropriate debugging line I should use to turn it on
>> & see if it can be trapped/pinpointed?
>> >> Suggestion/feature request:  Can this thing get a focus
>> >> adjustment made available in the gui, and saved in the prefs?
>> >> We no doubt have some users who aren't quite comfortable
>> >> shuffling around in the source code with vi just to get the
>> >> proper focus/registration for their individual scanner, and
>> >> obviously from this, hard coding it isn't going to be optimum
>> >> for everybody.  Mine was pretty bad when set for 16, but I
>> >> remembered 8 was pretty good from previous testing, and it
>> >> still is.
>> >
>> >No need to have this adjustment, as it should be okay for all
>> >devices...
>> How did it get reset to 16 then?  Possibly for the TPA maybe? 
>> But that seems counterintuitive to me, or do positive numbers
>> raise the scanhead, in which case getting into focus for a
>> mounted slide, that setting may be correct.  Somehow I had it
>> fixed in my mind that greater numbers meant farther from the
>> glass, not closer.
>> In any event, the TPA is surely going to need a different
>> setting if for no other reason than the thickness of the slide
>> mount frame raising the film off the glass at least far enough
>> to get rid of the newtons rings.  Also in that event, the array
>> isn't capable of holding a seperate value for each mode.  Fixing
>> that will be a major hack too.  :(  Which is half the reason I
>> asked for the prefs setting instead, which is probably about the
>> same amount of coding & caffiene in the long run.
>I think there's something wrong here;-) Sensordistance has nothing
> to do with the focus...

Ok, then where is the 'focus' variable set in plustek-devs.c?

>For the Plustek backend it is used in
> conjuction with the optical resolution to determine the min
> available x resolution: 600dpi / 8 = 75dpi...
>That's all here, for other backends it sometimes means the
> "distance" between red, green and blue lines.

In this case, it *does do* the 'distance' between red/green/blue, 
allowing perfect registration of all 3 colors to be achieved in a 
vertical, direction of the scan carraige merge.  Set it for 4, or 
12 and the image will be visibly out of color registration in the 
vertical dimension.

So in that sense, it works just liike the others, and apparently has 
little or nothing to do with the resolution available.  If it does, 
I haven't noticed it.

In any event, thats a minor quibble, fixable with vim and a 
recompile.  What I really need to do is figure out why it segfaults 
90% of the time at about the point where the scan is done and it 
turns the carriage around and heads for home.  The whole thing goes 
away, and the carriage continues on its merry way to home base, 
stopping just as if it knew where to stop.  But it leaves the lamp 
on forever, or until I restart it just to shut the lamp off.  This 
seems to be true for both a preview scan, and a real one at say 600 

On those rare occasions when it works, it will work repeatedly, and 
the preview image is just as pretty as the image the iscan 
libraries (gt-7200 output).  And its stable there now, whereas the 
second and succeeding scans of the same page were always way out of 
color balance before.

Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.21% setiathome rank, not too shabby for a WV hillbilly