plustek: Canon LiDE 20. Partial Success


Previous by date: 5 Feb 2003 11:45:17 -0000 Re: Canon LiDE 20. Partial Success, Alain Wenmaekers
Next by date: 5 Feb 2003 11:45:17 -0000 gamma correction, LarsTäuber
Previous in thread: 5 Feb 2003 11:45:17 -0000 Re: Canon LiDE 20. Partial Success, Alain Wenmaekers
Next in thread: 5 Feb 2003 11:45:17 -0000 Re: Canon LiDE 20. Partial Success, Alain Wenmaekers

Subject: Re: Canon LiDE 20. Partial Success
From: Gerhard Jaeger ####@####.####
Date: 5 Feb 2003 11:45:17 -0000
Message-Id: <200302051234.19561.gerhard@gjaeger.de>

Hi,

regarding your pictures, it really makes me wonder that the colours
are that bad, I've tested here with a color-calibration foto
and I got at least correct colors...

Anyway I'll crosscheck...

Gain means the gain-factor on the analog-frontend, for more info
checkout the download  download-section on the web-site for documentation upon
the LM983x and the Software-Reference guide..
I've put them this morning onto the web...

On difference between vuescan and the plustek backend is the illumination
setting (regs 0x2c-0x37), my calibration tries to find some reasonably
good values for theses registers (although it doesn't work correctly ;-)
and vuescan uses always 1 for the on point and 16383 for the off
point of each color. While the white fine calibration is turned off in
the Plustek backend, I think vuescan only uses this kind of calibration
(see also the Software-Reference guide for further explanations) I
began to think over a rewritten calibration routine for CIS devices here...

In your case, you might also get better results, when playing around with
the y-offset of the white calibration strip. To find the correct value, you
can adjust the startvalue 100 down to 0, then you'll see the calibration
area Play around with this value to have only the white strip in the lower
part of your picture, the new found value should replace the currently set
35...
structure in line 645, file plustek-devs.c (Lide20)
{{ 0, 100}, 35, -1, {2550, 3508}, {75, 75}, COLOR_GRAY16 },

If you have more questions - ask
Good luck
  Gerhard


On Dienstag, 4. Februar 2003 20:40, Alain Wenmaekers wrote:
> > I will play a bit further with these options, trying to find correct
> > levels for my scanner (and hopefully for most other LiDE 20 owners).
>
> I have fiddled a few hours with these settings. So far, these are the
> optimal settings that I found (caution, a better combination may exist):
> --------------------------
> option red_gain   1
> option green_gain 37
> option blue_gain  32
> ---------------------------
> With these options, I'm able to see the difference between the colors
>
> I have a testscan with some magazine pieces with colors in it:
>
> The original, with sane set to the default settings:
> http://users.skynet.be/bluedigit/sanedefault.jpg
>
> The one with the changed settings (still ugly, but the different colors are
> visible):
> http://users.skynet.be/bluedigit/corrected.jpg
>
> The reference scan with "vuescan" (colors are much softer):
> http://users.skynet.be/bluedigit/vuescan.jpg
>
> Scans under Windows/canon software look OK, but they are far worse than the
> ones from vuescan.
>
>
> What do the "gain" options mean actually? And how do they work? (so it's
> easier to find an optimal setting :-)
>
>
> Grtz
>
> Alain


Previous by date: 5 Feb 2003 11:45:17 -0000 Re: Canon LiDE 20. Partial Success, Alain Wenmaekers
Next by date: 5 Feb 2003 11:45:17 -0000 gamma correction, LarsTäuber
Previous in thread: 5 Feb 2003 11:45:17 -0000 Re: Canon LiDE 20. Partial Success, Alain Wenmaekers
Next in thread: 5 Feb 2003 11:45:17 -0000 Re: Canon LiDE 20. Partial Success, Alain Wenmaekers


Powered by ezmlm-browse 0.20.