PDA

View Full Version : black & white workflow-Silverfast Ai-Microtek 1800f



Eric Brody
3-Aug-2006, 21:49
I am now the proud owner of what Jim at Midwest Photo told me is one of the last 100 Microtek 1800f scanners at the Microtek importer in California. It arrived today, Thursday; it was ordered on Monday, pretty quick service. It's quite a hunk of a machine compared to the ancient Epson 2450 it replaces.

I purchased the 1800f for scanning 4x5 black and white negatives. I did lots of reading and consultation before making the decision that it's likely to be superior even to the new 7xx Epsons for 4x5.

Paul Butzi's article on his website makes it pretty clear that Silverfast Ai software is superior to Microtek ScanWizard. The scanner comes with Silverfast Ai, so I would like to to learn to use it. However, Silverfast is, at least at first view, not the most intuitive software I have ever used. Silverfast seems to be another order of magnitude of complexity compared to the Nikon scanner software I used with my 35mm LS-40.

Can someone make some specific recommendations for workflow for Silverfast Ai on the 1800f for black and white negatives? I have searched this site and done some googling as well but still have questions. I'm considering attending one of Ted Harris's and Michael Mutmansky's workshops, but in the meantime...

I've read that it is best to scan as a RGB file, do people use 48 bit, positive or negative? What about the negafix component, does one set it to the black and white film being scanned, TMAX 100 in my case, or does one pick a color film setting?

I have read Scott Rosenberg's article on choosing a channel, the sharpest, and using that. Is it done with channel mixer in PS CS2 or by somehow tossing the unused channels, and if so, how does one do that? I've seen some tutorials on the web and looked at the quicktime movies in the program but they're all dealing with color.

I would be most grateful for any assistance.

Eric

Ken Lee
4-Aug-2006, 04:04
I have a Microtek 2500F, and have the latest of both the Silverfast and Microtek drivers. I use VueScan, available at http://www.hamrick.com.

The microtek drivers give sharper results, but the extreme values get clipped. The silverfast drivers give a longer tonal range, but softer images. VueScan does not suffer from either of these problems.

It is a valuable step to calibrate the scanner. Microtek and VueScan make this fairly easy to do.

Michael Mutmansky
4-Aug-2006, 06:01
Eric,

You need to do some trials on the process to determine what method is giving you the best results for your needs.

You shoud test to determine which channel comes out sharpest when doing a RGB scan. Green will probably be the sharpest. Definately use 48 bit for B&W, as you end up with a 16 bit file once you have rejected the two channels down to the sharpest one. If you use 24 bit, you will only have an 8 bit file when all done, which can cause banding problems at times. Even with 16 bit, you have to be careful of banding.

You can reject the channels in the channel mixer, or by copying the channel you want and then erasing the others. There are probably other ways to do it as well, but those are the most direct that I know of.

Whether to scan in positive or negative mode is somewhat dependant on the operator; most people prefer to see the scan preview correctly, so they scan in negative mode. However, there may be a subtle (or not so subtle) tonality difference between two scans that are the same except for where the image was tonally inverted (PS or in the scanner software), so I recommend trying that variable as well to see which way you prefer.

Negafix applies curves to the scan, and it is another case where you will have to try it and see if you get results you like. I generally don't like to use adjustments that I am not clear on the details of, so I generally don't like, to apply curves using Negafix without careful study of what is going on. If, however, you run some tests and find that the curve for the film you use produces a good result, then by all means use it.

I have used Vuescan as well as the ScanWizard, in addition to Silverfast, and in many cases there may be no functional difference among then in terms of results. However, Vuescan seems to have a real problem with selecting the proper highlight point on some images, and so does ScanWizard. In almost all cases, when this occurs, Silverfast seems to andle the situation the best.

Even though I prefer the interface on ScanWizard, the results with Silverfast are probably equal to or superior in most cases. That's enough to convince me that it is best to learn Silverfast and use that as the primary software for the scanner.


---Michael

xvvvz
4-Aug-2006, 06:39
Since you are new to Silverfast, you might like to read these tutorials:

http://luminous-landscape.com/tutorials/silverfast-scanning.shtml

http://www.computer-darkroom.com/sf5-negafix/sf5_negafix.htm

Also, don't forget about the video tutorials provided by Silverfast. You can download them from their website. They also sell a book if books are more your style.

Doug
---
www.BetterScanning.com

Ed Richards
4-Aug-2006, 07:10
I use both Vuescan and Silverfast, when one cannot handle a negative well, the other often can. Vuescan improved a lot last fall, so I recommend that anyone who tried it before that might want to try again. I tend to use Vuescan as the default because I do not try to fine tune the image in the scanner. I am more interested in making sure nothing gets clipped, then I can do the rest in PWP or Photoshop. As we discussed to the beating a dead horse stage some time ago, as long as you are getting 16 bit output, it really does not matter whether you are doing the adjusting in the scanner or the editor - as long as nothing is clipped and nothing is too compressed at either end of the curve. Go for fairly flat out of the scanner. Silverfast's updated multi-scanning feature seems a great improvement over the previous version, at least on the few negatives I have tried since it was updated recently. If your negatives are too dense or thin, it looks like it will help.

Kirk Gittings
4-Aug-2006, 08:02
"as long as you are getting 16 bit output, it really does not matter whether you are doing the adjusting in the scanner or the editor - as long as nothing is clipped and nothing is too compressed at either end of the curve.

This is only true if do not manipulate tone very much as Ed does. If however, you alter contrast much, burn and dodge allot etc. it is better to make your major global contrast, density, curve etc. adjustments in Silverfast. By doing it in the scan stage you build up allot less noise than doing the same later in PS.

The "alignment" feature in SF multipass scanning works only if the passes are pretty close in size. Your first scan of the day on a cold negaqtive may give a scan different in size enough that it cannot be aligned. If this occurs do a "warm-up" scan first to heat up and expaned your negative. then immediately do a real scan. I set them up as a batch scan with the first simply a one pass warm up which i discard followed by the real scan. I find a 4x usually sufficient to significantly reduce noise.

Brian Ellis
5-Aug-2006, 23:16
Sorry to be so late, I missed this thread when it was first posted. If you use Silverfast Ai (and I used it along with Vuescan for quite a while, finally decided on Silverfast Ai exclusively) I'd suggest spending about $30 for the book "Silverfast - The Official Guide" by Taz Tally. It does a much better job IMHO than anything else available in explaining how to use Silverfast to its maximum advantage.