Hi,
> > However, the RGB offsets as I understand them are not an offset in value but a
> > spatial offset. E. g. if you scan a black square on white background in
> > color mode it doesn't have sharp edges but maybe a blue edge on one side and an
> > orange edge on the other. Just try it.
>
> This is interesting. Is there a solution/fix for this? I'm asking because I'm
> in the process of implementing a correction for the same "feature" in the
> EPSON Perfection 610 scanner.
Obviously some effort has been made in the current snapscan backend, but for
some reason the code is now marked as obsolete and #ifdef'ed. Most of it can
be found in snapscan-310.c. If I understand the solution correctly it works as
follows:
The offset values are retrieved from the scanner using the inquiry command.
Then every scanned line is stored in a ring buffer (with r, g and b value).
Once there are enough lines in the ring buffer to synchronize r, g and b the
lines can be read out again.
There are some comments in the code that suggest that there is a RGBRouter
somewhere that should replace the old sync routines but I haven't found any
implementation for it.
Another problem is that the inquiry command says that all offsets are 0 for my
scanner...
Oliver
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Fri May 05 2000 - 07:33:14 PDT