Using Silverfast To Extend Scanning Dynamic Range

[neil poulsen]

I was on the Lasersoft website recently, and I noticed they have a version of Silverfast that takes a normal scan and a slow scan. The slow scan enables the scanner to gather greater detail in the shadows. The two scans are combined to give a resulting scan that has increased dynamic range. It's called their multi-exposure technology. It's available in both their Studio version and their Archive Suite version. It's also available in a third configuration, althought I can't remember what it's called.

Has anyone used this approach? If so, how well does it work in practice? It sure sounds interesting.

[neil poulsen]

Here's the link to the Archive Suite information:

http://www.silverfast.com/show/silverfast-archive-suite/en.html

By the way, these guys are great to deal with. How refreshing. After losing my Silverfast Ai disk for about a four or five year old Epson 4870 Pro scanner, they're allowing me to upgrade to their highest end software for a reasonable price. They would have replaced the needed Ai software for free. But, I opted to upgrade instead.

What a class act.

[Chris Strobel]

I have it, and a 4990.On my PMK 4x5 negs I see no difference, but instead zooming in from 100% to 400% see a slight smearing of details, possibly from the neg heating and moving?I don't know.I have Silverfast Studio, Epson Scan, Vuescan Pro, and a Scanscience wetmount kit for 4x5 and 8x10.Even wet mounted doing multipass scans just seems to smear details on the zoomed in level, and makes no difference in my prints.I spent two full days testing, wet mounted, dry mounted, various shims for focus, platen direct mount, multi pass 2,4,8,16, on and on.My conclusion with b&w negatives developed with PMK in the print on a Epson 4800, ABW mode, Harman Gloss FB AI, no perceptible difference.

[neil poulsen]

Interesting.

I'm wondering if it's possible to control the speed of the scan oneself? This might give the operator greater flexibility in bringing out shadow detail. One could manually conduct both scans and blend them in Photoshop.

Out of curiosity, with all the testing that you conducted, what did you conclude? What's the best way that you found to scan?

[Chris Strobel]

Well I'm back to just using the Epson 4x5 holder and Epson Scan software for 4x5, and for 8x10 Epson Scan, and using the Scanscience 8x10 matboard holder and fluid to avoid the newton rings I would sometimes get scanning using the Epson 8x10 area guide and laying the film on the platen.By the way the Lumina fluid sold by Scanscience DOES leave behind some residue contrary to their claims, but it does clean up pretty easily, and is basically oderless which is why I like it over Kami fluid.

What really drove me nuts was this whole plane of optimum focus thing.Scanscience gives you these shims to vary the height of the film above the platen glass.I experimented going from the film directly on the platen all the way to stacking all the shims together, and then zooming in 400%.It was playing mind tricks on me, my wife too, only when all the shims were stacked did we think 'MAYBE' it wasn't as sharp, but when printed out at 12x15 absolutely nada difference.Man did I waste alot of expensive paper and ink.

With the software it comes down to work flow preferences for me.Of course I 'want' to like Silverfast and Vuescan better, I paid for them, Epson Scan was free, but I just can't with tons of tweaking make anyone produce a better scan/print than the other.Mind you this is all with Ilford fp4+ PMK processed negatives in both 4x5 and 8x10.I don't do the color thing at all 'yet', so maybe its a whole nother ball game scanning chromes and color negs, I don't know.Hopefully the color guys will chime in.

[neil poulsen]

The nice thing about Silverfast for color negs are the presets that one can specify for each brand and type of film. (e.g. Portra, 160S, etc.) I found that this really helped me scan color negatives.

Using Epson Scan, I have more trouble getting decent color balance from color negatives.

[Keith S. Walklet]

Neil,

For my best results (color transparency), I configured the scanner as follows:

I set my gamma to 3.0 and created a profile with Monaco EZ color, then configure the scanner to use that profile, and scan into prophoto colorspace. I set the highlight and shadow threshholds to 2 levels, just in case something actually bumps either extreme.

No sharpening. 48bit HDR, multi exposure to at least 2 (check results) to minimize noise in really dense shadows. I don't both with ME if the scan is not challenging.

The idea of gamma 3.0 is to push the image data to the right where there are more discreet levels to describe the image.

The result are high-key, low contrast scans. I set the white point and black point in PS when I open the image for editing. As is the case with RAW capture files, I find it much easier to darken tones than to try and lighten them, which usually introduces noise.

All this, plus wet mounting.

But, the long and short of it is, often I couldn't get the entire dynamic range with EpsonScan, and I CAN get it with Silverfast.

[neil poulsen]

Neil, For my best results (color transparency), I configured the scanner as follows: . . .

Interesting. I just received my Archive Suite (Ai and HDR Studios) from Lasersoft. I'll have to give this a try.

[urs0polar]

Never thought I'd write it as it confused me at first, but I'm liking Vuescan now. I have a Nikon 9000 and a V700. On the 9000, I've been messing around with some Efke 25 6x7 B&W negs. Vuescan is sort of a pain, but now that I am locking exposure for the blacks and locking film base color and so on, I feel like I am getting a very accurate scan. This is the first time I have been able to beat Nikonscan.

I have Silverfast for the V700, and still use it for color negs, but I think with Vuescan's ability to lock the film base color, I can profile the *exact* base color so as to get the truest representation (this is for color neg film obviously, but I do it for B&W too since I scan in RGB).

So, I was dreading the $400+ I would have to spend for Silverfast for the Nikon, and now I'm liking Vuescan Pro for $80. So, definitely try Vuescan and try locking exposure and so forth before shelling out for Silverfast.

As for the Multi-Exposure, I think the scanner head has to scan the whole film a few times (unless the scanner supports this in-place, which I think few scanners do), so it is very difficult to line everything up exactly the same for each pass, hence the smearing. Myself, I found that it helped on some negs (I've done it in Silverfast).

Here is a Silverfast scan I did from a 6x9 Panf 50 developed in HC110 negative from an old voigtlander folder. This was scanned on the V700 with the betterscanning dry mount station. I played around with multi-exposure to a hint of the leaves in the bottom of the frame to come out. This negative had so much dynamic range in it, I had like 5 different pictures to choose from -- and I'm not sure panf is the highest DR film out there. This is scanned at -3 EV (from what Silverfast thought was a normal auto exposure -- whatever that means) with multi exposure applied.

http://www.flickr.com/photos/urs0polar/3623440941/in/set-72157615694598067/

Now that I'm a little better with Photoshop, I think that I could do around the same thing with curves/burn/dodge and only one scan, especially with the latitude of B&W film on 120 and 4x5. If you don't do multi-pass, you can still do multi-sample, which leaves the scan head in the same place and averages values to get rid of noise. I think the above image is 4 samples.

Also, see this thread, just found it; Vuescan also does multi-exposure:

http://www.largeformatphotography.info/forum/showthread.php?t=31847

-Mark

[ronald moravec]

You can scan twice, once for shadows, once for highlights and combine in photoshop with a luminosity mask.

Nice tutorial on the lights right website., digital darkroom tab, video tutorials

You may have to allign the two scans by putting the top in difference mode and use the arrow keys.

[Peter De Smidt]

Some scanner/software combinations allow multiple line reading. Instead of doing a couple of scans of the entire negative and combining them, MLR takes multiple readings of each line during one pass of the scanner head and averages them. This can help eliminate noise, and it tends to have less registration issues than multi-pass scanning.

[pocketfulladoubles]

Here's the link to the Archive Suite information:

http://www.silverfast.com/show/silverfast-archive-suite/en.html

By the way, these guys are great to deal with. How refreshing. After losing my Silverfast Ai disk for about a four or five year old Epson 4870 Pro scanner, they're allowing me to upgrade to their highest end software for a reasonable price. They would have replaced the needed Ai software for free. But, I opted to upgrade instead.

What a class act.

+1 same experience here. You don't always get customer service like this anymore.

[mrladewig]

One of the irritating things about EpsonScan is that you don't really have direct control of the exposure as you would in say NikonScan. But I have noticed that the scan times vary with the selection made in the levels control. If your selection is made largely to the left side of the histogram, you'll get a longer exposure/slower scan. If the selection is encompassing a larger range, the scan times and presumably exposure are shorter. I would expect that you could trick EpsonScan into giving you a slow scan and a fast scan by manipulating the levels configuration. You'd have to blend it all together manually in photoshop, but this should work.