[sane-devel] Current TODO list

Henning Meier-Geinitz henning@meier-geinitz.de
Sat, 26 Jul 2003 14:19:13 +0200


On Fri, Jul 25, 2003 at 11:23:37PM +0200, Stefan Schlörholz wrote:
> > However, I'm not really sure what is being asked here.  Can you
> > provide more details?  Is it just moving cdjpeg.h  and djpeg.c into
> Since I do not know much about programming, header files, object files 
> and the works I do not know what you are talking here.

Well, we are talking about the TODO list and one of its entries was to
move some internal files. I'm not sure why you don't respond to the
email that covers the icm topic but to this one.

> Let me explain my suggestion:
> The software for my scanner under windows offers a dialog where you scan 
> a reference target, tell the programm which target standard you used 
> and mark the positions of the reference marks.

So it does a preview scan and you must select where the target is
positioned on the glass? Or does the software find out the position
automatically ?

> The program then knows at which position to expect what color, say pure 
> red. It also knows which color it actually got from the scan. It then 
> calculates a profile file (suffix .icm) with information how to correct 
> colors. Any scan is then corrected using this scheme automatically.

Ok. From your description I don't see any need to add support for icm
files to backends or the SANE API. For me it looks like it can be
included in the frontends without the help of the backends. Or a
special "frontend" could be created that just does the calibration and
other frontends could just use the .icm files.

If I understand the method correctly, the frontends then use the .icm
files to adjust the gamma tables. What I don't really understand is
how that corresponds to the other scanner settings like exposure,
gain and offset which have influence on the brightness/gamma of the
image, too. So the .icm profile is only valid for one scanner with one
setting of these values?

> The quality of a scan under linux differs extremely from a scan under 
> windows. The windows scan is by far closer to what you expect. The 
> histogram, gamma correction and the works can not compensate for it.

If you can't correct the color by using a gamma table, calibrating
with the standard target won't help. Maybe in your case the widnows
driver uses other settings (e.g. exposure, gain etc)? Generally
speaking, the images scanned by SANE are NOT always worse than with
Windows, so this may be a backend/scanner-specific problem.

> There are some projects for this color management under linux but they 
> end up applying the profile after scanning and reworking it in a third 
> application applying histogram or other funktions.
> This could be done in one go if sane could handle the color management. 
> It may be splitted in frontend and backend part. Each scanner has its 
> own profile. The aquire and processing the .icm file should always be 
> the same.

But the profile muste be created with the help of the frontend anyway.
So I don't see what role the backend plays here. Keep in mind that
quite a lot of scanners don't have gamma support at all, that's all
been done in software.

> I believe that this correction procedure is very important to scanners. 
> Having such a color management would greatly improve acceptance at the 
> professional users.

Ok, then these users should propose the details on how to do it and
send patches :-)

I'll add a pointer to thie discussion to the TODO list.