PDA

View Full Version : Locking Exposure with SilverFast



sanking
23-Nov-2005, 06:47
I am scanning B&W negatives in Grayscale with the Microtek 1600Xl using Silverfast. The 12X20 negatives must be scanned in two parts and later stitched together in Photoshop. The problem I am having is that there is a small density difference between the two scans which makes seamless stitching difficult, and I can not figure how to lock down the exposure with the Silverfast software so that the two parts have the same density. Both Vuescan and the Epson software that I used in the past with an Epson 836Xl permitted locking down the exposure, and I figure that Silverfast must also, but am not able to figure out how to do it.

Anyone know how to lock down the exposure?

John_4185
23-Nov-2005, 07:40
What version of Silverfast? If an older version will work, then perhaps you can get it from Silverfast - or write me privately. I'm sure it is okay for me to send you a full copy of an older version if you have a licensed current version. I'll check to see what I have at the lab today.

Scott Schroeder
23-Nov-2005, 08:48
Sandy,
For me I noticed silverfast pulls up the same histogram and curve settings from the last scan. I believe that is what you meant by locking. It is probably under one of the settings. Maybe on the general tab-> option parameter...I have mine set to <save>
I also have on the main screen: Settings set to save.
hope that helps

sanking
23-Nov-2005, 09:18
I have the latest version of Silverfast Ai, or 6.

Yes, Silverfast does pull up the same histogram for the second scan and I don't change anything at all between the scans, not the histogram nor the curve adjustments. Neverthess I am getting a slight difference in density between the two scans, and this happens whether the scanner is cold when I began or whether it has been turned on for several hours.

And since I am scanning in Grayscale I figured color would not have anything to do with it, though my scans are of pyro stained negatives.

However, there appers to be something that is allowing the density to drift somewhat between scans.

Kirk Gittings
23-Nov-2005, 09:41
As Scott said you should be able to save the settings. I never do this but at the end of each scan when I go to close out SF a dialogue box comes up asking me if I want to save my settings. They can be saved there. Also try....once all your settings have been applied save the settings as a new Frame. I think when you go to use that frame again your settings will reaply.

I am doing some scanning this afternoon. I will think about that issue.

Kirk Gittings
23-Nov-2005, 09:49
The other thing would be to not!!!!!! Prescan the second image. Set up the first as a batch scan capture it for the batch scan with all the adjustments and then move the negative over and recapture it as next batch scan without reinterpolating it with a new prescan.

Kirk Keyes
23-Nov-2005, 09:55
"And since I am scanning in Grayscale I figured color would not have anything to do with it"

You may be scanning in greyscale, but your scanner is scanning in color and then converting or tossing some of the color channels.

You mention it being worse when cold - I have that issue too (a now vintage UMax Powerlook III). I suspect the lamp temperature in my case. Because of the lamp variability, I run several preview scans before I ever try to actually scan something, or make adjustments to the scanner settings.

If you have the latest version of Silverfast, prehaps you could try setting the multipass option to do a few scans automatically, and then discard that scan, make your adjustments right afterwards, and then do your final scan.

You may also want to test the eveness of the illumination. Are your matching up parts of the image that are scanned from opposite sides of the scanner? Your scanner may not have the same illumination when comparing one side to the other. Maybe rotate the film 180 degrees and try to position the spliced area over the same area on the scanner surface...

Kirk

Kirk Gittings
23-Nov-2005, 10:02
As per what Kirk said, it would be best to that as a batch scan with multipasses with dummy scans first to heat everything up before the real scan. I do that on my 4990 to stretch the negative so that I can get good registration on the MPs.

sanking
23-Nov-2005, 10:59
Actually, what I meant to say was that the problem is the same whether I start with the scanner cold or if it is warmed up for several hours. And BTW, I don't do a prescan of the second image. I just import Silverfast and click on Scan. I do check the historgram and levels adjustment to make sure they have not changed from the first scan and that has never happened.

Also, I don't think the multipass solution is practical. Even with one pass it takes about an hour at 1600 dpi to scan the two parts. With multipass option it might take me a whole day to get through one of these negatives.

The part about scanning in RGB and converting to grayscale makes me think that the problem must be in one of the control controls since I have the histograms and levels locked down.

Scott Schroeder
23-Nov-2005, 11:10
Well you can just stop shooting 12X20 and send the camera to me!
Okay, since that's not gonna happen :) I thought of one other thing. Do you always scan the same side first.....the left then the right? If so you might try switching that around or even possibly scannig with the same side off the scanner...i.e. one scan will be upside down.

Kirk Keyes
23-Nov-2005, 13:23
If it is already warmed warmed up, then it's probably not the lamp output changing and you should not need to do any warmup passes.

Have you done an "Eveness" type scan? Easy to do with reflected light - set a sheet of white paper on the bed and do a scan and bump up the contrast to see where there is light fall-off and where the scanner "sweet spot" is. Kind of like testing your enlarge/enlarger lens for fall-off on the easel by printing onto high contrast paper. Not sure what you could put in the bed to do this? A piece of ND acetate? I suspect a sheet of Lee theatrical filter would work well for this.

Are you trying to butt the two scans together end to end? What about having some overlap and then feathering in the two scan? That could help hide any minor differences in density...

sanking
23-Nov-2005, 14:01
Kirk,

I have not done a test for eveneness. Will think about how to do that.

In stitching I don't butt end to end. I overlap one layer over the other and then smooth the edges with the eraser tool set with foreground to white. This makes the meeting area absolutley seamless, unless there is significant difference in density between the two layers.

Actually I must have clicked on something today that corrected the problem this morning when I was looking at the various tools trying to figure out what was going. All of my scans today have been almost perfectly even with the first and second scan. But I can not figure what I did to correct the problem, other than changing back to color matrix from the tubes and setting tolerance back to the mid-range.

I appreciate all the suggestions. Some good ideas in there for consistency.