iSRD has Problems with Reflecta DigitDia 5000

[hoges]

SilverFast 6.5.5r2 seems to have a problem with using iSRD on a DigitDia 5000. When I use it, the result is like that:



And when I switch it off, the result is as expected from the Prescan:



I did the following:
- I installed SilverFast
- I started it via "SF Launcher"
- In the Options, I unchecked "Prescan schneller" and "HiRePP"
- After leaving Options with "OK", I declined to do a Prescan
- I set the DPI to 3600, enabled iSRD and set "Mehrfachscan" to 2
- I saved the frame settings as "Setting_1"
- I did a Prescan
- I called the function "Bild-Rahmen suchen"
- I recalled "Setting_1"
- I called the function "Bild-Automatik"
- Finally I scanned to a "TIFF" file (with JPEG2000 I got the same results)

During the scan, the progress window always showed the correctly cropped picture, only sometimes turned by 90 degree.

The result with iSRD switched on has less dust, so the function itself is working.

Did I something wrong or is this really a bug of the software?

[LSI_Heidorn]

Dear Hoges,
i just deleted my previous posting where i stated that this bug is fixed.
I now retried and must have done something different --> I can see & verify your Bug !

The problems seems to be the combination of:
iSRD + rotation of prescan + fine ( degree-wise ) rotation.

If you eliminate one of those factors it works.

We will fix that one :-)

Greetings,

Nils Heidorn

[hoges]

Hello Nils,

it's good to hear that you can also reproduce the error. As a software developer I know the problems with it.

I found also out, that I have no problems with portrait scans and with iSRD switched off.

But is there a way to switch off the fine rotation during batch scan with "Rahmen suchen" while using the functionality of an automatic frame? Maybe this could help me for the moment.

Regards
Horst

[LSI_Heidorn]

Dear Horst,

you cannot realy switch it off as it is not meant to be buggy :-)

You can adjust the rotation and hold SHIFT, then the rotation will snap to 45 degree steps so you can easily set it to 0 degrees so you switched it off !

Greetings,

Nils Heidorn

[hoges]

Hello Nils,

I remember that it was working sometimes. So in these rare cases the rotation was probably 0 degree.

It's too bad that it's not possible to switch off the rotation while being in batch mode! So I have to wait for the fix.

Do you have any estimation when this will be done?

Regards
Horst

[LSI_Heidorn]

Dear Horst,
as fast as possible with testing afterwards !

I hope to have a customer upload ready end of next week ?!

Lets try :-)

Greetings,

Nils Heidorn

[marekv]

Hi, I have a Reflecta DigitDia 4000, and with 6.5.5r3 I experienced the exact same problem. Going back to 6.5.1r1a seemed to "fix" it, though I have not tested whether that is because, for instance, the fine rotation feature is missing or what. The really interesting part was that even with auto-frame OFF it would still happen in 6.5.5r3. Possibly my saved frames were not at 0 degrees rotation, I did not know of this problem when I was trying it out.

Best regards, Marek (Oslo, Norway)

[LSI_Heidorn]

Dear Marek,
we fixed some "fine rotation" bugs from r3 to r4, could you verify if that also solves your Problems ( using the r4 from the web ) ?

Greetings,

Nils Heidorn

[marekv]

Hello Heidorn, I just tried using the latest (r4) release, and the issue is still there.

My parameters: Frames found using Auto frame, that were slightly rotated (correct, my scanner feeds a little bit skewed). Frames then Saved in both Landscape and Portrait mode, and these saved ones were the current selection on the Frame tab - not Auto Frame.
I used iSRD with IR strength 8 detection, no other filters of any kind (AACO, GANE etc) active. Auto contrast and Colour cast removal were active.

Using batch mode with two images, one portrait and one landscape, the progress dialog thumbnails showed the correct orientation, but the final images were both cut to landscape, which was the current red frame outline shown in the Prescan window.

I repeated the procedure, this time snapping the rotation to 0 degrees; otherwise everything was the same. This gave the correct results.

I can send you log files if you wish.

Best regards,
Marek