[sane-devel] HP scanjet 2300c, xsane, xscanimage, somme comparisons

[stef]

Le 22/04/2012 13:53, Gauthier Vandemoortele a écrit :
> Hello,
>
> I've done some tests with my scanner (Hewlett-Packard ScanJet 2300c)
> and xsane or xscanimage.
>
> I found xsane more user-friendly, but now I note that with some
> images, results from xsane are not usable, while xscanimage can
> handle such tricky cases.
>
> Particularly, I can'nt get soft-toned images at 1200 dpi with xsane:
> all that i get is a sharp black and white-only image.
>
> In case it can be interesting for developpers, I've uploaded some
> image at
>
> http://www.k1ka.be/pics/scanner/
>
> example_xscanimage_300dpi.pnm   with default gamma
> example_xscanimage_600dpi.pnm   idem
> example_xscanimage_600dpi_b.pnm with custom gamma, better
> example_xscanimage_1200dpi.pnm  with default gamma
> example_xscanimage_1200dpi_b.pnm with custom gamma
>
> I was surprised that a correct gamma at 300dpi becomes too dark at 600
> and more, and it was necessary to lighten the loud keys.
>
> example_xsane_1200dpi.pnm  (all with same setting, see gamma_xsane.png)
> example_xsane_300dpi.pnm
> example_xsane_600dpi.pnm
>
     Hello,

     this scanner has 2 ranges of work, 75-300 which has an exposure and 
600 which has another. Due to these different exposure, I'm not 
surprised that you need a different gamma curve since you may have 
different gain and offset calibration. One way of achieving this would 
be to use the same exposure for all resolutions. Anyway, could you send 
me the log of a scan session where there is a 300 and 600 dpi color 
scan. I want to have a look at calibration results.
      1200 dpi is in fact 600 dpi for sensor and 1200 for motor. So 
sensor's data are expanded by software, which doesn't work with 
soft-toned scan (ie B/W scans) as it is currently written.

Regards,
     Stef