[sane-devel] GT68XX Plustek OpticSlim 2600 Support

[Michal Waligora]

Hi,

Since I took so long to create the issue and gather the code... I wanted to link the issue here, just in case someone would want to find it in the future: https://gitlab.com/sane-project/backends/-/issues/805

I'm still working on figuring out the color calibration issues.

Cheers,

Michal
On Monday, December 30th, 2024 at 8:31 PM, Ralph Little <skelband at gmail.com> wrote:

> Hi,
>
> On Mon, Dec 30, 2024 at 11:25 AM Michal Waligora <michalwaligora62 at proton.me> wrote:
>
>> Hi,
>>
>> I’m extremely pleased to say that I got the Plustek OpticSlim 2600 working just fine!
>
> That's great to hear!
>
>> Sadly, calibration failed completely, even with the GT68XX_FLAG_OFFSET_INV flag. I traced back the “setting AFE reached limit” error message to the relevant source code and saw that the GT68XX_FLAG_NO_CALIBRATE flag, which isn’t mentioned on the adding page, disabled a block of code responsible for calibrating. The scanner works with the flag only.
>>
>> I am unsure why the calibration is broken. Is the existing GT68XX driver somehow slightly incompatible with the OpticSlim 2600, or could it be caused by hardware damage (although the scanner looks fine)?
>
> It may be that calibration is highly scanner specific or at least paramaterized in ways that the backend cannot yet handle.
> Unfortunately IFAICS, these chips do not have public programming instructions so I suspect that the backend has been composed largely of replicating captures and inference.
> I do now have 1 or 2 of these scanners and started to look at the code recently. It is probably low on my list of stuff to do though. :(
>
>> Next, multiple manuals I found online suggest that the output color depth should be 48-bit, but compiling with anything more than {12, 8, 0} and {12, 8, 0} gave an error about invalid arguments (I think) at runtime. I tested DPIs from 300 to 2400, and they all worked.
>>
>> Please let me know if you would like this upstreamed. I’m more than happy to help contribute it on Gitlab.
>
> Yes, certainly. Let us not lose the work you have done.
> Probably the best thing would be to create an issue on gitlab here: https://gitlab.com/sane-project/backends/-/issues and add what you have found with any code fragments that you think pertinent.
>
> Thanks again for your efforts.
>
> Cheers,
> Ralph
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/sane-devel/attachments/20250820/e15ec12e/attachment.htm>