[sane-devel] This 1200UB mystery

Henning Meier-Geinitz henning@meier-geinitz.de
Sat, 5 Apr 2003 12:00:30 +0200


On Fri, Apr 04, 2003 at 08:50:52PM -0600, jhaynes@alumni.uark.edu wrote:
> I just set up the 2.4GHz machine and the 66 MHz machine side by side,
> scanned the same picture with the same scanner with both machines, and
> copied the file from each to the other.  The result is that the scan
> done on the 2.4 GHz machine has severe moire effect, while the scan
> done on the slow machine looks good, when viewed on either machine.
> The raw bits files from scanimage are of identical size on the scan from
> either machine.  (But the files differ, starting at byte 38, line 5)

Please send me both files. But not as jpeg but in a lossless format
like pnm. Please also sned me the command line switches/options used
for both images for scanimage.

> Does anybody have a clue why the speed of the CPU seems to make a
> difference?  Or what else it might be?

Theoretically that could be a timing problem, i.e. your computer is
too fast :-) Maybe the scanner expects some pauses at a certain point
of the protocoll and with faster computers the commands are just to
fast. The protocol of these scanners is broken/not conforming to the
USB spec anyway, so I wouldn't be surprised.

On the other hand, I'd expect I/O errors and not moire artifacts in
such a case. And nobody else reported such problems until now. 
With a 1.4 MHz Athlon it doesn't happen and I don't have a faster one

Is the scanning time different significantly? Try "time scanimage
>image.pnm" on both computers.