plustek: Re: [sane-devel] Epson Perfection 1250
Subject:
Re: [sane-devel] Epson Perfection 1250
From:
Gene Heskett ####@####.####
Date:
7 Mar 2002 14:47:32 -0000
Message-Id: <20020307143548313.AAA109@mail.iolinc.net@there>
On Thursday 07 March 2002 02:21 am, Reinhard Max wrote:
>Hi,
>
>On Wed, 6 Mar 2002 at 21:11, Gene Heskett wrote:
>> and I think at the moment its a combination of a poor
>> calibration strip
>
>I don't think that the calibration strip is to blame. I've had a
> look at mine. It's just plain white.
>
>> and a semi-cold lamp that didn't have time to adequately warm
>> up, something that would effect the color of the ends of the
>> lamp much more than the center 90% of it..
>
>I am also not sure that a variation in the lamp's intensity
> between center and ends is the main reason.
>
>To me it looks like the color corruption on the borders is due
> to the optical system in front of the sensor. The sensor itself
> is just about 4cm wide and the light is being projected to it
> via a system of mirrors and lenses. AFAIK the light throughput
> of a lens degrades towards it's border. This gets even worse if
> the sensor is not adjusted exactly at the center of the lense.
> Of course the impact of this effect reduces as the lamp becomes
> brighter during warm up. When my lamp is warmed up for an
> adequately long time, the color errors on the borders become
> much lower, but then the center of the page starts looking a
> bit overexposed.
That seems to sum it up. The warmup time allowed (I tried some
hardcoded in the code) however didn't seem to have near as much
effect as the difference between the first scan of the day, and
all subsequent scans. There, there is a positively huge
difference! And its even greater if its had a 2 or 3 day rest.
The first scan is a throwaway, always.
--
Cheers Reinhard, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz 512M
98.6+% setiathome rank, not too shabby for a hillbilly
