[sane-devel] newbie

[m. allan noah]

Regarding the colors of the iX500, they are not strictly raw data. The
scanner self-calibrates, which removes most of the high frequency
noise. But, the default brightness and contrast values may still need
some manual tweaking. It is also possible that the windows software
does some further enhancement, which is not done in the sane backend.

allan

On Mon, Jan 25, 2021 at 11:36 AM Andy Bennett <andyjpb at ashurst.eu.org> wrote:
>
> Hi,
>
> > I'm looking to get into scanning in the GNU/Linux world - I do
> > not yet own a scanner. I know the SANE compatibility matrix
> > exists [0], but I am hoping for some anecdotal recommendations.
>
> Welcome!
>
>
> > What scanners are people using and how happy are you with your
> > scanner and your SANE front-end?
>
> I have a Fujitsu ScanSnap iX500 and I'm very please with it.
>
> However...
>
>
> > Also, I read a short blog post [1] about the author's scanner
> > yielding different results under Windows and GNU/Linux. The
> > author claims that the Windows driver can have better color
> > accuracy. Does anyone have any commentary about such notions?
>
> > [1] https://www.digitaltidying.com/dont-use-linux-to-scan-photographs/
>
> I can't speak to this issue particularly, but the iX500 has colour issues
> of its own.
>
> I scan to pnm files with the included "scanimage" backend.
>
> I then have my own toolchain that does the things I want to these files.
>
>
> AFAICT, the data that ends up in the pnm files is the raw data from the
> scanning head. i.e. its. plain RGB intensity values.
>
> On the one hand, this is great because getting the raw data leaves me with
> the most options.
>
> On the other hand, it's a pain.
>
> These files look terrible on Linux. Weirdly, the exact same files look fine
> under Windows.
>
>
> If I correct the files with pnmgamma, applying the sRGB or CIE profiles
> then they look fine.
>
>
> Also,
>
> The scanner is a duplex unit and therefore has a scanning head and a bright
> light on both sides of the page. This means that the exact colour
> reproduction varies greatly with the thickness of the paper.
>
> I have a stock set of colour curves that I use to correct things to how I
> like them.
>
>
> ...but in general, all colour is lies, and faithful or accurate conversion
> between CMYK (used for print) and RGB (used for screen) is particularly
> difficult.
>
>
> Perhaps if you use one of the more featureful SANE frontends then it does a
> lot of this stuff for you, but I was keen to make my own toolchain.
>
>
> When I first got the scanner I read this guide:
> http://www.samhallas.co.uk/repository/scanning.htm and learnt a lot.
>
>
>
> Getting passable scans from any toolkit (Windows or Linux) is fairly
> straightforward. Getting "stunning" scans is always an art, similar to
> digital photo manipulation.
>
>
>
> What you'll need to go alongside SANE will depend a lot on what you want to
> use the scanner for.
> For example, documents might benefit from OCR (optical character
> recognition). Photos might benefit from a full Photoshop style package.
>
>
>
> Let us know some more of your requirements, such as use case, budget, etc,
> and I'll try to offer some more anecdotes (if I have any).
>
>
>
>
> Best wishes,
> @ndy
>
> --
> andyjpb at ashurst.eu.org
> http://www.ashurst.eu.org/
> 0x7EBA75FF
>


-- 
"well, I stand up next to a mountain- and I chop it down with the edge
of my hand"