[sane-devel] Epson V500 Gray negative scan problem

[Drasko DRASKOVIC]

Hi Olaf,
thanks a lot for clarifications.

On Tue, Apr 24, 2012 at 4:55 AM, Olaf Meeuwissen
<olaf.meeuwissen at avasys.jp> wrote:
> I've since gotten a lot more
> familiar with iscan and can now confidently say that almost all image
> processing with iscan/epkowa is done in the iscan frontend.  Conversion
> of negatives is definitely done in the frontend.

Yes, I saw that it was a long time ago, but I guess that issue is very
similar... I was afraid that most of image processing is done in the
closed source front-end, but I was more surprised with bad results I
obtained from XSane... I guessed that my setting were wrong.


>> Does anybody have idea how this can be resolved ?
>> How autoexposure/enhancement can be set-up to be correct and resemble
>> to what iscan is doing ?
>
> Apart from using iscan, you cannot.
>
> The image processing module is non-free (with an effectively
> undocumented API) so most frontends will not be able to use it.

Really bad news... I was afraid of this.


>> Can this be used :
>> http://www.xsane.org/doc/sane-xsane-setup-enhancement-doc.html
>> or these setting should be used only for lineart (i.e. commic scans)
>> and not for film negative scans ?
>
> That would give you a black and white (without any gray) image.  I don't
> think that's what you're after.

Yes, I supposed that, but I was not really sure. As I see, these
settings are more for comic strip scan, for example. I was just
searching setting for negative scan, to see if I can improve scan
quality.

However, here is what I discovered yesterday evening during few hours
of experimentation :

1) Scanning film negatives in Gray gives bad quality (like bad
resolution/small bit depth). Looks bad in the Preview and in the Scan.
2) Turning on "Media definition"
(http://www.xsane.org/doc/sane-xsane-medium-definition-doc.html) to
change the setting give orange-reddish preview. Why? Why not just
gray?
3) Pressing on "Enhancement default" button (while having orange
preview in "Media definition") turns this image finally in the Gray.
Good. Now I can edit the Gamma, Brightness and Contrast by looking at
B&W image in Preview window which changes appropriately.
4) When I am satisfy I press directly Scan button with "Media
definition" still enabled. The Scan of what I have set-up comes out.
If I turn off "Media definition" and come back to simple Negative Gray
scan, all the settings that I set up will be changed and I will end up
with a bad quality scan. I do not want to save new media every time,
because although I am dealing with the same film, I'd like to correct
exposure/contrast that came from a different lenses and measurements,
and I want to do it per photo, not per film role.


Based on all of this, as it is a bit unusual process, I have a few
additional questions :

1) Why are the negatives in Preview window turning in the orange color
when "Media definition" is selected and into Gray only when
"Enhancement default" button is pressed. Moving sliders from there
keeps image in Preview window Gray, which is OK, but why do we always
have to begin by pressing this button?

2) Why when selecting any other film type (Agfa film, Kodak film...)
image in Preview windows becomes orange? Then I take note of Gamma,
Brightness and Contrast, press  "Enhancement default" button for image
to become Gray, and then apply this values by hand. This is kind of
buggy behavior...

3) Why by turning off  "Media definition" all the settings are messed
up and why although I change this setting by hand to resemble to what
I had in  "Media definition" mode the Preview scan is done with other
(obviously pre-defined) settings and they are set-up automatically
after Preview scan (maybe because I keep media to Standard Negative
option, so it keeps turning back to it) ?


Finally, with the method I described I am capable to obtain decent
quality. It is not maybe as good as iscan, but I guess that it can be
set. I noticed that it iscan also seems to be doing some king of
better focusing (like sharpening the image), although I might be
wrong.

As a final question, I noted here :
http://www.xsane.org/doc/sane-xsane-medium-definition-doc.html
that it is written in bold letters :
"Do not use this mode to scan images, xsane will not behave like you
expect when you try to scan images in this mode."
However, as you noted from my process, this is exactly what I am
using. This is because I want to avoid saving one setting per image. I
know that this should be done per film, but I am trying to obtain best
scan per image, by correcting wrong exposure time and lack of lenses
contrast and try to get the best scan per image. What are the
differences, and is it mandatory to save these setting every time? Are
these kept in the separate files, or all is film types are kept in one
.mdf file ?

BR,
Drasko