[sane-devel] Snapscan + Epson PF 3490

Oliver Schwartz Oliver.Schwartz at gmx.de
Wed Nov 9 22:25:47 UTC 2005


> I own an Epson Perfection 3490 and I am beginning to use it with
> the CVS version of SANE (snapscan backend). While it seems to work
> well at resolutions up to 1200 dpi (but I did not yet perform
> really thorough testing), I noticed a kind of interlacing behaviour
> when scanning negatives/diapositives with the TPU at 3200 dpi. The
> scanlines are periodically offset from each other. Maybe it is
> alsoi present for the flatbed part, but I don't have such high
> resolution reflective stuff to test it out.
> I would like to help the snapscand backend maintainer to fix the
> problem, as I would like to scan slides and negatives besides the
> usual flatbed content. What is needed in order to help as much as
> possible? I am thinking about scanning e.g. a small area with
> different resolutions and bit depths, then visually compare the
> results. But what should I do with these test images and the
> backend logs? I don't want to post binary data directly to the list
> for bandwidth reasons. Should I use the bug page at alioth?

The effect is known from other scanners such as the Epson perfection 
2480. For this model the correction is already done in the Snapscan 
backend. If you can provide me with a high resolution scan that shows 
the effect (i.e. an image that allows to count the pixel offset 
between two adjacent lines) the changes for the 3490 should be easy 
to implement.


More information about the sane-devel mailing list