[sane-devel] Benq 5150C / 5250C users: please test

[Oliver Schwartz]

Hi,

> Ok, I did some testing with my BenQ 5150C.
>
> My first try was in 300dpi color mode (because this was default).
> There are no color stripes anymore. However, the scanned area was
> not very wide. Although the full available range
> ((0,0)x(216.0,297.0)mm^2) was selected in xscanimage. The resulting
> image had only 109 mm in width and approx. 145 mm in height
> (measured with millimetre graph paper)
>
> This is independent of selected resolution and colour mode.

Looks like the measurements are off by a factor of 2. That should be 
easy to fix.

> 'Gray' doesn't work. It still results in very distorted Images.
> When doing 'gray' with xsane there is a message box saying "Backend
> sends more image data than it defined in parameters", after
> scanning in preview mode.

Hm, ok. The backend currently always sends some static calibration 
data for color mode. I'll try to get some gray scale mode calibration 
data as well.

> 'Halftone' sends wrong colours, and sometimes the same patterns as
> gray. In xscanimage, preview crashes in this mode.

This may be related. However, a lot of modern scanners don't support 
halftoning at all. Does the windows driver offer something similar?

> > At present the calibration algorithm for the 5150 / 5250 is
> > unknown. The calibration data sent to the scanner is currently
> > some static data which may or may not work for all scanners.
>
> I'm not a hardware developer. But if you need any data, be free to
> ask. But as my attempt to sniff on the USB-bus in Windows fired up
> the whole system and I could not scan anymore until the next
> reinstall of W2K, I won't do that anymore.

I have some data samples, but the hard part is to make an educated 
guess on the used algorithm.

/Oliver