[sane-devel] Interesting image problem when scanning color ...

[Gerhard Jaeger]

Hi Kenneth,

On Wednesday 13 October 2004 23:16, Kenneth H. Peters wrote:
[SNIPSNAP]
> I download, compiled and installed the latest 1.0.14 CVS version (this
> version also exhibited the imaging anomalies with plustek.) I wanted to=
 try
> another backend, so I purchased the Epson 2480 (need latest CVS to supp=
ort
> this) and used the "snapscan" backend.
>
> I was able to scan using this scanner which indicates to me that any
> big-endian issues associated with running on a IXP425 in BE mode is not=
 in
> the frontend (i.e. scanimage) but maybe in the backend.
>
> Attached is a ZIP file that contains output using the "script" command.
>
> The "gentoo_xxx" (Works fine) are files generated on the following:
>
> =09- Gentoo 2.6 kernel
> =09- x86
> =09- libusb
> =09- Canon LiDE 30
> =09- plustek backend
> =09- scanimage frontend
>
> The "ixp425_xxx: (demonstrates "color-bar" problem) are files generated=
 on
> the following:
>
> =09- MV 2.4.18
> =09- IXP425 in big endian mode
> =09- Canon LiDE 30
> =09- plustek backend
> =09- scanimage frontend
>
> These all used the "scanimage -x 10 -y 10"
>
> There is additional output in the function "scan_it" from scanimage.c. =
This
> diag output is from "if ((output_format =3D=3D OUTPUT_TIFF) || (parm.de=
pth !=3D
> 16)) and dumps "buffer".
>
> Please let me know what additional information I can provide.

When scanning with the debug options enabled, there will be a file=20
called plustek-pic.raw

Can you send me the files from your gentoo-box and from the IXP?
Please scan something which contains some colors and or stripes to
see the mismatch.
=46rom the logs you've currently sent I can't see any problems. The regis=
ter
settings seem to be okay, even the calibration steps seem to work on the
two boxes.
The only difference I see is, that on the gentoo box, scanimage seems to
get serveral lines in one transfer, while on the IPX each line is transfe=
rred.
The next odd thing is, that on the IXP you'll get some 0 data, but why?
You'll probably send me all the raw files you get from one scan...

Gerhard