[sane-devel] genesys-backend and ML5345 test results with 1.0.17

Stéphane VOLTZ stefdev at modulonet.fr
Thu Dec 29 07:48:43 UTC 2005


Le Mardi 27 Décembre 2005 15:20, Martin Jacobs a écrit :
> Hello,
>
> after downloading libusb-0.1.10a and 1.0.17 I repeated my tests
> with USB-Scanner Medion MD5345 and got the following results:
>
>
> Test environment:
>
> Linux box based on SuSE 8.1, dual P3 933MHz, 768MB, m/b MSI
> 694D-Pro (MS-6321).
>
> Resolution 100     200   300    600
> Color      ok      ok    ok 3)  ok 3)
> Gray       ok 1,2) ok    ok     ok 3)
> Lineart    ok 1,2) ok 4) ok 4)  ok 3,5)
>
> Test conditions: each scan was done at least two times. Tests have
> been done using scanimage.
>
> Footnotes
> 1) Reset of scanner (power off) does not help, hotplug service has
> to be restarted with unloading of module usbcore.
>
	You can generate debug logs by doing:
export SANE_DEBUG_GENESYS=255
export SANE_DEBUG_GENESYS_GL646=255
scanimage 2>scan.log >scan.pnm

	Send me log of a failed scan so that I can search what's going wrong. Having 
USB logs of the same sort of scan under windows, recorded with 
'usbsnoop' (http://benoit.papillault.free.fr/usbsnoop/index.php.en) would be 
great.

> 2) Only first scan is successful, any other fails.
>
	How does it fails ? Does it hang or do you get black/garbage image ?

> 3) Head is moving back and forth.
>
	This is normal. At thes resolutions the amount of data to send is higher than 
the USB bandwidth. The data are read slower than they are produced. So the 
scanner as to pause, or even go backward to the place for which data have 
been sent. This is commonly called 'backtracking' and is not specific to this 
backend. By specifying a narrower width, needed bandwidth decrease and this 
goes away.

> 4) Second scan out of 5 was scrambeled.
>
> 5) Third scan out of 5 was scrambeled.
>
>
	4 and 5 are surely related, and I got one other report of this. I did try it, 
and obviously this doesn't happen here. However I did seen such with cvs 
version, but got it fixed by disabling power saving helpers. You could try to 
compare with SANE 1.0.16 to see if you have also this 'scrambling'.
	I will try it on another PC, it may be related to USB subsystem (or the way 
it is used by the backend). Maybe you could try this also. Is your mother 
board USB 1.1 or 2.0 ?

>
> Compared with sane-backends there is a great improvement!
> Congratulation.  But why do the "simple" test modes Gray and
> Lineart fail?
>
> Regards
>
> Martin Jacobs

	Regards,
		Stef





More information about the sane-devel mailing list