[sane-devel] CanoScan 5600F support

[Stef]

On 03/06/2013 22:15, Carsten Jensen wrote:
> On 06/03/2013 09:35 PM, Stef wrote:
>> On 31/05/2013 17:04, Carsten Jensen wrote:
>>> Hello.
>>>
>>> I found this:
>>> http://sane.10972.n7.nabble.com/CanoScan-5600F-Genesys-backend-td16745.html 
>>>
>>>
>>> and got curious if there is any news on this ?
>>>
>>> the thread seems a little vague to what needs to be done.
>>>
>>> I'm new to sane, so please treat me as such :-)
>>>
>>> cheers
>>> Carsten
>>>
>>     Hello,
>>
>>     I couldn't work on it since I'ma lacking such a device. To have 
>> full support for it  there are device descriptions to add in 
>> genesys_devices.c and genesys_gl847.h . Currently the gl847 handles 
>> only CIS sensors, and not CCD yet like your scanner. Calibration 
>> would have to be adapted to this sensor (look at the genesys_gl841.c 
>> code to have an idea of what to create).
>>     Since we don't have datasheet of each components inside the 
>> scanner, the only way to work is do reverse engineering on the 
>> scanner, using an USB snooping program to record windows driver 
>> activity.
>>
>> Regards,
>>     Stef
>
> C/C++ isn't one of my strongest sides. but I can do a lot of USB 
> sniffing. That is, if you have time/energy to work on this. I'm 
> getting a Mustek 1200s USB A3 too which also isn't supported. I can do 
> USB sniffing on that too.
>
> cheers
> Carsten
>
>
     Hello,

     the C code involved isn't difficult. The hardest thing for these 
scanners are the tests and tuning. After looking at the code, there 
doesn't seem to have significant differences for CIS/CCD sensors. The 
biggest change will be the analog frontend. From an old usb snoop log I 
already have,  it is a "wolfson I" type, so we'd have to make offset and 
gain calibration work. Code is already there, but it has never been tuned.

Regards,
     Stef