[sane-devel] EPSON 3490, calibration and raw data

[Levente Novák]

On Sun, 2006-02-19 at 21:09 +0100, Krzysztof Halasa wrote:

> I will probably end my experiments with a patch.
> 
> > Do you know whether this scanner sets a different focal plane when
> > scanning slides/film compared to the flatbed (as it should)?
> 
> Haven't checked that yet. But I don't know if the optics aren't fixed,
> that's ~ $100 consumer scanner after all.

OK, but it is important to know. If it is fixed (and set to the plane of
the flatbed), I'm better to not use the film holder provided, instead I
must put the film directly on the glass plate.

Once, I was planning to perform some experiments by putting a high
resolution film at different heights from the glass plane in order to
determine the exact focal point. I think it's time to do it. But this
will only work if there is not a kind of "autofocus" built into the
hardware (or done by software).

Levente

PS: By several years of experience with optical instruments (microscope,
telescope, etc.) I must say that achieving quality is not always by
buying something expensive. The best optical setup is often attained by
very simple means, and even the best refocussing algorithm on a blurred
image is worse than a less capable on a sharply focussed one. In the
world of mass-produced equipment, there are factory tolerances, you can
get a perfectly adjusted component as well as one which is clearly out
of focus.