halos around high contrast edges scanning E6 film no ICE

[gharhoff]

On quite a number of my scans there are pronounced halos around high contrast edges (noticeable at 100%), rendering the scans virtually useless. At first I thought this is because I switched on Digital ICE (standard) on my last round of scans. But I rescanned one of these slides with and without ICE and it does not make a difference at all. I rescanned using Nikon Scan with ICE enabled and the halos are gone! So I think this is not a problem of the scanner hardware. I use a Nikon Coolscan 9000, Win XP prof., Silverfast Ai Studio 6.5.1r3, normally as a standalone application, no sharpening applied. Film is Fuji Provia 100F or 400F. I know the very same problem has been discussed on this forum with regard to Kodachrome using ICE and has been solved with a Silverfast update. But since this issue occurs with my setup using the lastest SF version and E6 material regardless whether ICE is enabled or not, I don't have a clue what's going on. My halos look exactly like those on the pictures posted on the thread concerning Kodachrome film. Does anybody have an idea?

[degrub]

try manually setting exposure ( wp,bp) and see if that changes it.

any filters applied for the exposure processing in SF ?

just some ideas.

[gharhoff]

Hi Degrub, thanks for your help! Actually I always manually set white and blackpoint for each RGB-histogram individually, then I adjust curves trying to get as much shadow and highlight detail as possible. I don't apply filters, at least not that I am aware of. Anyway your suggestion set me on the right track: The cause of the problem seems to be that I tweaked curves too aggressively on some scans. I had thought of this, too, therefore I did not apply any adjustments at all when I rescanned one of my problem images yesterday. What I did not realize was that SF had saved my last correction and applied it to this image again. Now I scanned once more and reset the curve before - now the halo is gone. Thanks again, I would not have tried this again had it not been for your suggestion. My next question is: Is there a way in prescan mode to find out how much adjustment you can apply safely? From what I have understood it is better to make these adjustment in SF rather than later in Photoshop.

[degrub]

Well, if you are scanning in HDR mode (nominal 16 bit) then all i do is set B and W and pass it to HDR to work on it ( it could be PS just as well with CS2 or CS3 ) . IF you are scanning with an 8 -12 bit scanner, there is some value in making as many adjustments in the scan as possible so you avoid posterization later. However, you are working with a small portion of the scanned image so it will be a little off on exposure, but usually not to much. But i find if the image was reasonably exposed to begin with for moderate enlargement i don't see any issues.

i don't do much curve tweaking at scan time any more unless the image is poorly exposed or i am combining a high range and low range scan of the same image.

Watch out setting B and W for each channel - you can miss a sunset or two

i do use multiscanning to get the last 1-2 ( 4x - 16x) bits of random CCD noise removed.

Yes, i have been caught by that feature occasionally as well. :-?

[gharhoff]

Degrub, thanks so much for your tips! I guess I still have to learn quite a bit about scanning. I was batch scanning a large number of slides and applied my routine adjustments, but now I see that at the end of the day it saves more time to do your homework first rather than having to rescan a whole number of images.

[degrub]

THen i won't tell you about the first 1000 or so i re-scanned :D :D :D

[gharhoff]

Wow! I had believed rescanning would be obsolete once I replaced my old scanner with the Coolscan 9000. But I can see your point. Can you recommend any particular learning source or did you just struggle along? I did read Taz Tallys book and another one in German but they didn't alert me to the problems I was facing last week.

[degrub]

Well, there is

www.luminous-landscape.com
for imaging techniques, equipment and learning how to see an image - from the scan and when taking the shot

www.computer-darkroom.com
for silverfast and imaging in general

Then i have learned a lot from

the Yahoo group - under scanners - ScanHi-End where i learned all sorts of technical information about scanning and what is really going on from the drum scanning folks.

FLAAR is another interesting source.

also newsgroup - comp.periphs.scanners

search for any posts by Kennedy McEwen on Nikons - he knows his technical imaging.

You can also look at some of the astronomy photo websites - since it is CCD imaging as well , for more techniques and some software to boot !

multiscanning to reduce measurement noise came directly from the astro guys.

and then for learning how to read an image -

any of the books by Margulis - he is from the old school printing techniques where color reproduction and image quality depended on the skill of the scanner and printer ( and still does - CMS only get you closer usually ;-)) )

Katrin Eismann on retouching is excellent.

and then practice with difficult slides to learn the techniques.

[gharhoff]

Thanks so much! I do know Michael Reichman's and Ian Lyons' site, but the others are new to me. I own Eismanns book on masking and compositing and I've been thinking about getting one of the Margulis books for some time, I know his LAB method is being discussed quite controversially. So much to learn... Thanks again for the great tips, next time I'm in Texas I'll have a Shiner on you!