Page 2 of 4

Posted: Mon Apr 02, 2007 9:35 am
by BOBRAPP
I have two scanners - Minolta's 5400 Mark 1 and Microtek's 120 tf with FW 1.8. The Minolta works fine with M-E (although the original did not) and occasionally showed alignment problems when extended to ME 4).

M-E does not work with any source wether 35mm or 6X9. I am truly waiting for a version the works. The results from the Minolta continue to validate my selection of scanner and software - which is far better than making projection prints with the best lenses!!

Kind Regards,

Bob Rapp

Posted: Wed Apr 04, 2007 4:38 pm
by LSI_Heidorn
Dear all,

i treid out several scans on the ArtixScan 120tf Firmware and everything was fine...

UNTIL i switched from 48 Bit HDR Color to 48 Bit Colour !

After the switch i experienced similar Problems !

Could you ( all ?! ) verify if things go better if you scan HDR instead of 48 Bit Colour ?!

Its hard to say if it depends on my Image Selection, so maybe together we can find out whats going on.

If you also get more stable results in HDR, then i got a starting point for my bug-hunting !

Greetings:

Nils Heidorn, R & D, LaserSoft Imaging AG

SprintScan 120 with same issue

Posted: Wed Apr 04, 2007 5:11 pm
by Rocky
Nils Heidorn,

Thank you for the update. My first tests on the Polaroid SS120 were with HDR. After I found serious image degradation in these scans, I tried 48 Bit Colour. The results of both were the same. My scans were made from 35mm and 2 1/4 x 2 1/4 transparencies by way of both the Film Holder and Glass Holder.

I well be interested to hear about the results that others have had when scanning with Multi-Exposure.

Rocky

Posted: Thu Apr 05, 2007 9:52 am
by LSI_Heidorn
Dear Rocky,
i'm also eager to hear from the others..

When you talk of "image degradation", are you talking of that "color sparkeling edge" effect or of other quality Problems ?! Please note that HDR Scans are meant to look dark & dull, just apply a Gamma Curve of 2.0 to them to make them look a *bit* better ...

Greetings:

Nils Heidorn, R & D, LaserSoft Imaging AG

Posted: Thu Apr 05, 2007 1:36 pm
by Rocky
Dear Nils Heidorn,

I suppose the term "image degradation" is vague. What I find are that the scans appear similar to the one shown by deldridg's example posted on the first page of this thread, which reveals a sense of "color sparkling edges." Of course, loss of image information in some small regions can also be noticed at high magnification.

Rocky

Posted: Fri Apr 06, 2007 8:56 am
by BOBRAPP
I retested my Microtek with 1.8 FW and the results are the same. There is alignment problems with both 48-bit and HDR. The degree of mis-alignment is the same for both scans. This is a result of 4 different images.

Cheers,

Bob Rapp

Posted: Tue Apr 10, 2007 2:29 pm
by LSI_Heidorn
Dear BOBRAPP,
thank you for your Testing !

As you got different results than me, i fear we have to retest the whole thing, we might contact you via PM to ask you about the images you used and maybe some scan samples, okay ?

Greetings:

Nils Heidorn, R & D, LaserSoft Imaging AG

Posted: Wed Apr 11, 2007 12:45 pm
by BOBRAPP
Hello Nels.

Please do. I will be back with my scanners tomorrow.

Bob

Posted: Sun Apr 29, 2007 10:14 am
by BOBRAPP
An updated version was posted on 20 April. This version still does not correct the M-E problem. Is there any hope?

Regards,
Bob Rapp

Posted: Wed May 02, 2007 7:30 am
by LSI_Heidorn
Dear Bob,
a colleague of mine is working hard on this, the source of the problem is located but i am not sure about a solution, i'll ask him to comment this thread later today !

Greetings:

Nils Heidorn, R & D, LaserSoft Imaging AG

Posted: Fri May 04, 2007 4:37 pm
by Johannes
Good evening Bob and Others!

This is my first posting in Lasersoft-Forum.

My question: How can i reproduce the error?

I am using a Microtek Artixscan 120tf and downloaded Silverfast 6.5.0r5 in Demomode. I had to switch to "48 bit color" to select "MEx2". Then i scanned a negative (Agfa Optima) with MEx2 and without. The MEx2-Selection took longer as expected.

But: I do NOT see any differenz between the scans using WindowsXP build-in Picture Viewer opening the scans saved as TIF (135MB each, 6x7 format, 2000dpi). Windows-Viewer refused opening the JPG, because its a new 12-bit JPG as long as 48 bit Color is selected.

Neither an error nor an degrading/improvemet!!! What did i wrong? Must do it again...

Posted: Sun May 06, 2007 10:16 pm
by Johannes
OK, i did now a test with a Kodak Ektachrome Slide in 48-Bit-Color and got alignment problems too. Has some pixels offset and red edges here and there. Looks not good.

Posted: Mon May 07, 2007 7:58 am
by LSI_Muenier
Dear Johannes,

I have made ME tests at different resolutions and found out that scans larger that 2000 dpi (where a scanner resolution of 4000 dpi is used) lead to a slight vertical stretch (!) when exposure is increased.
With 2000 dpi and below I did not observe such a strange effect (which I couldn't believe at first).
The result in the ME scan is a vertical offset between the image data from the partial scans which is continuously *changing* in the vertical direction.
This is really unique to the SS120/120tf, I did not see anything like this with all the other scanners for which we have implemented our Multi-Exposure!
As a work around, please use scans with resolutions up to 2000 dpi, for the moment.

Best regards
Martin

Posted: Mon May 07, 2007 4:16 pm
by Johannes
thanks a lot for your answer. i will try this. gives me at least the chance to compare image quality with/without Multi Exposure.

i am not sure, can it be, that 120tf sometimes has this vertical stretch also between two independent scans at different dpi-resolutions??? i mean, i saw this also in the past using Microtecs own scanner-software instead of Silverfast. scans with very early versions of silverfast also had some blurr unsharpness in lower dpi-resolutions. only the max resolution of 4000dpi delivered fine results. but this problem seems to be solved with up to date Silverfast-versions for the 120tf.

i like this scanner and would like to see a solution. time will tell...

Posted: Tue May 08, 2007 4:20 pm
by Johannes
did a try with 1000dpi. the scan with Mex2 looks almost the same like the one without. no huge pixel offset! but quality is even a bit worse than scanned without MEx2. So i asume it is not working even in 1000dpi resolution setting.