[sane-devel] HP ScanJet 5590 issues

[Ilia Sotnikov]

On Feb 14, 2008 11:26 PM, Paul C. Bryan <email at pbryan.net> wrote:
...
> 1. "Overscanning" pages (simplex/duplex ADF)
>
> If the ADF-scanned page is smaller than the geometry set in the -y
> parameter value, content from the following page is included on the
> previous page. It would seem that the backend is not detecting where a
> page is actually physically stopping, and is simply continuing onto the
> next page to get the content it wants. Where there is no content (no
> more pages to feed from the ADF), the remaining "expected" page area is
> filled with a quasi-crosshatch pattern).

You're right - the scanner doesn't have the paper sensor which could
be accessed programmatically (at least original Windows doesn't show
such a possibility).
Thus pages are distinguished only using image size calculations.
However, I havent' seen such a strange behaviour. I'll need to check
with one of the scanners I could access.

> 2. Duplex pages mirrored and upside-down pages (duplex ADF)
>
> The backs of pages are being scanned mirrored and upside-down. I have
> managed to automate a solution to this in a script (attached), which
> seems to work.
>

The ADF duplex scanning was discovered occasionaly and implemented in
most straight way - return double-height pages to scanning
application, otherwise an user
should mention that he needs 2 x number of pages.

> 3. Color picture scans are streaked (flatbed)
>
> High-contrast areas in the tops of photos seem to streak down below into
> the rest of scanned photos, causing significant degradation of image
> quality. This kind of streaking does not seem to occur with lineart
> scans.

Also haven't seen such quality degradation effects, at least our users
didn't report such ones. Will also need to check.

Could you please open the bug on SANE bug tracker (as mentioned on
http://www.sane-project.org/bugs.html) and attach some example
pictures which show such effects?

Thanks in advance,
-- 
 Ilia Sotnikov