[sane-devel] ANNOUNCEMENT: dust removal: please test

[Major A]

> and it worked (despite of the graphic-distiller). duster is compiled.

Ah, yeah, if you don't have GTK2 installed, you can skip that one by
using "make -k".

> I have tested duster with different more ore less dusty (and hairy, 
> because of the cats here around ;-) ) slides, but I see no big difference
> in the output of duster and the original rgb-scan. This ist contradictionary
> to the log of duster:

This depends on the type of film you have. Take a look at the
beginning of sanei_irclean.c and modify the IR_BASE and R_SLOPE
values. The values give a threshold for the infrared channel according
to

  ir-threshold = IR_BASE + R_SLOPE * red-value

where everything is normalized to 16-bit values (0..65535).

> Reading IR file... done.
> Reading RGB file... done.
> Combining files into one array... done.
> Cleaning image... done.
> Replaced 5728 of 6747 obscured pixels.
> Writing output file... done.

6747 is not very much, in full-frame scans with the LS-30 I usually
get about 200000-500000 obscured pixels (OK, that includes the border
which is covered by the film feeder).

> btw: The output couldn't  be opened with gimp (1.2.3). 
> Message of the gimp: x-resolution is wrong.

Ah, thanks for telling me. I always use 16-bit images which GIMP can't
use anyway, so I've never tried opening the result with GIMP.

  Andras

===========================================================================
Major Andras
    e-mail: andras@users.sourceforge.net
    www:    http://andras.webhop.org/
===========================================================================