[sane-devel] genesys Canon Lide 210 aspect-ratio-problem at 600 dpi

[Stef]

On 25/04/2014 02:55, Andrew Kanaber wrote:
> I hit the same problem with a Canon LiDE 110. I tried measuring the distortion
> by scanning a sheet of graph paper in various modes.
>
> With sane 1.0.24 in 600dpi lineart and gray modes the scan is stretched
> vertically by about 30%. Color mode scans are actually also stretched at
> 600dpi, but much less noticeably by about 2%. The stretching is still visible
> though if you compare the scanned images side-by-side. With version 1.0.22 the
> lineart and gray scans are correct at 600dpi but the 2%-ish stretching is
> still present in 600dpi color mode.
>
> In case it helps I've put some of the scans up at
> http://www.chiark.greenend.org.uk/~akanaber/lide110/
>
> Let me know if there's something else I can do that might help,
>
> Andrew
>
> M G Berberich wrote:
>> Hello,
>>
>> Canon LiDE 210 with sane-backends 1.0.24 (libsane:amd64 1.0.24-1.1+b1
>> from debian testing) has problems with 600 dpi gray and lineart.
>>
>> In 600 dpi gray and lineart the scan is partial and aspect-ratio is
>> not correct. Vertically only about ? of the area is scanned and then
>> scaled to the area, horizontally all is fine. (So probably the driver
>> sets the wrong scanline-skip (to small)).
>>
>> This worked in older versions, at least 1.0.22.
>>
>> 600 dpi Color-scanning works fine, so do preview, 300dpi and 1200 dpi
>> in all modes.
>>
>> 2400 dpi and 4800 dpi do not work at all, but never did.
>>
>> 	MfG
>> 	bmg
     Hello,

     the vertical resolution is given by the motor speed which is 
eventually an int. So rounding errors may give ~2% distortion:
     target=((exposure * dpi) / base_dpi)>>step_type;
     Maybe by decreasing or increasing slightly exposure time rounding 
errors could be mitigated, but changing exposure may raise other 
problems. If the windows driver can do it, I'll manage to find something.

     I'll check how the gray scan beahves. The change between 1.022 and 
1.0.24 is to use the LINESEL function when we try to scan below the 
minimal scanning speed of the motor. This should affect scans below 600 
dpi, but not 600 dpi which is the minimal scanning speed.

     2400 and 4800 dpi scans works on my LiDE 210. So I'd need some logs 
or better information on what is not working for you.

     I'm currently working on shading calibration problems for gl646 
based scanners (HP2400 and HP3670). I'll look into this issues when I 
have finished it. If you have time for it, you can open a bug in the bug 
tracker (https://alioth.debian.org/tracker/?group_id=30186) and assigned 
it to me.

Regards,
     Stef