PDA

View Full Version : Improving Scanner DMax - Scan both a negative and Duplicated Positive?



Michael Heald
30-Aug-2006, 09:29
Hello! Since I don't need to perform a lot of scans, I am wondering about techniques that may improve scanner performance, even though they may be time intensive.

From what I've read, scanners have the most difficulty in the densest part of a negative. For simplicity, in B&W, that would correspond to the highlights of a scene. Hence, a scanner with high DMax is able to pull more information out of the densest part of a negative (scene highlights) that anothe scanner that has a lower DMax

If I copy a negative in the darkroom to a positive, the densest areas of the negative become inverted to the least dense areas. The scanner should be able to scan these areas of the new positive much easier than in the original negative. Is this correct?

Also, copying a transparency generally results in an increased contrast for the scene. I would imagine that this effect might be ueful in scanning a scene's highlights as well (the densest part of the original negative).

As a result, I was wondering if scanning both a negative and a positive of that negative that had been made in the darkroom and then combining the two scans in photoshop would improve the final digital scan. Best regards.

Mike

Kirk Gittings
30-Aug-2006, 09:41
Michael,
It seems like a long route to fix a fairly common problem. You would have all the problems associated with copy negatives plus registration issues with the scans etc. The solution as I see it is to tailor the exposure and dev. of the negative to fit the scanners capabilities. If you are talking about existing negatives dev. for traditional processes with dense highlights, a first rate drum scan by a competent tech can do amazing things or you might consider reducing the negative before scanning.

Michael Mutmansky
30-Aug-2006, 10:32
Michael,

I see this as a non-issue for the most part, as long as you are developing your negatives to an appropriate contrast (approximately to that which produces good silver prints).

Let me explain. I develop my large negatives for palladium printing, which requires a much more contrasty negative than silver. I aim for a DR of 2.0 approximately for those negatives. For silver, I aim for about 1.0 to 1.1 DR.

The palladium negatives arenot 2x as dense in the highlights, they are almost 8x as dense, so we are talking about a considerble difference here.

The palladium negatives will cause a flatbed scanner fits because the density of the highlights are getting up into the range where the scanner can no longer distinguish density differences. However, with good scanning skills, it is possible to make a good scan from them even.

So back down in the land of silver negatives, unless the negative was seriously overdeveloped or overexposed, the density range on the negative will be well within the capabilities of a flatbed scanner.

A far greater problem is the selection of the white point within the scanner software. This is critical, and is where the majority of scanning problems occur, IMO. Black point isn't as important, because the black and near-black seems to be much more open to intrepretation for many people, but the white point seems to have much less room for error before people will object.

While specular highlights on a negative may go up pretty high in density, for the most part this is irrelevant because specular highlights are meant to be paper white, and the scan does not need to retain these highlights to maintain a good tonal range in an image.

One last thing that I find heavily impacts the quality of the scan; the software. I've seen negatives that looked terrible on one software program, but as soon as the same image was scanned in Silverfast, all the problems went away. This happens with at least one image at every workshop that I'd been part of, and it has convinced me that for B&W scanning, I don't think there is an equal to Silverfast AI. Despite that fact that I don't like the interface too much, the results make me recommend this program above all others.


---Michael

Bruce Watson
30-Aug-2006, 11:12
...I was wondering if scanning both a negative and a positive of that negative that had been made in the darkroom and then combining the two scans in photoshop would improve the final digital scan.
I don't see how it could. The dup, no matter how good it is, is still going to loose some information. I would think that the dup would also show some increased fog as well (fb+f for both sheets). That, and the image size on each sheet is going to be very slightly different due to the vagaries of emulsions and processing. Scanning both (and you'll never get the two films to line up exactly the same relative to the CCD array) and adding them together in Photoshop is more math on the pixels; you'll loose some information and add some blurring.

As the others have said, if the negative is a normal darkroom negative, any scanner in good working condition can read it. If it is an alt. process negative, most any scanner in good working condition can read it. A Dmax of 2.0 should be within reach of all the current crop of consumer flatbed scanners, and certainly anything more capable can read those too.

B&W negatives use metallic silver for image formation. The silver tends to scatter light (Callier Effect). I find that negatives that are less dense than a typical darkroom negative tend to scan a bit better, but if the negative is optimized for the darkroom then it scans just fine in my experience. I find this to be especially true of drum scans, but it is also true of my old Epson 2450 consumer flatbed.

So, if you are having problems scanning B&W negatives, I would look to the scanner and in particular, look to the software. I found the Epson software very difficult to use to get good negative scans, while VueScan (http://www.hamrick.com/) was much better. I've heard that SilverFast (http://www.silverfast.com/) is excellent with negatives, but it has a seemingly universally reviled user interface.

Michael Heald
30-Aug-2006, 11:15
Hello! Now I am confused!

Reading past posts, there seems to be two quality issues of scanners - sharpness and DMax. While I can get a hand on how to compare sharpness between scanners, now I am confused by the importance of DMax.

For a negative developed for silver printing (or perhaps with a little less contrast range) say with the highlights in Zone 7, from these comments, it would seem to me that a scanner with high DMax will pull just as much printable detail from the highlights as a scanner with a lower DMax, because the negative was developed to a lower contrast range. For higher contrast negatives, a scanner with a higher DMax would perform better. Is my understanding correct?

Best regards.

Mike

Brian Ellis
30-Aug-2006, 11:19
My guess is that whatever improvement in dmax you might see would be more than offset by the loss in quality you'd experience when working from a second generation "negative." Perhaps it was me but my interpositives never retained all the detail and tonal gradations of the negatives from which they were made. And I can't tell you what a pain it was to make the interpositives. The only thing worse was making the internegatives from the interpositives but of course you wouldn't have that pleasure.

Bruce Watson
30-Aug-2006, 11:27
For a negative developed for silver printing (or perhaps with a little less contrast range) say with the highlights in Zone 7, from these comments, it would seem to me that a scanner with high DMax will pull just as much printable detail from the highlights as a scanner with a lower DMax, because the negative was developed to a lower contrast range. For higher contrast negatives, a scanner with a higher DMax would perform better. Is my understanding correct?
How about an analogy. Let's say you have a car that can run at 200 kph. You use it to run around town at 60 kph. You also use it to run up and down the M roads (Interstate, Autoban, whatever) at 120 kph. Your brother has a car that can run at 160 Kph.

Which car does a better job of running at 60 Kph? 120Kph? Exactly. Both do equally well. Neither is straining at either speed.

So if you have a negative that has a Dmax of 1.2, and one that has a Dmax of 2.0, and a scanner that can reach a Dmax of 2.5, it should scan either negative just fine. If the scanner could reach a Dmax of 3.0, it would still scan either negative just fine. Just because a scanner can reach a higher Dmax doesn't mean that it will give a better scan...

keithwms
31-Aug-2006, 05:01
Why not just scan your negative as a positive, if you think you are losing detail in the shadows? I don't see any reason to go to the bother to make an analogue positive. After you scan as a positive you can invert in postprocessing. Easier, no?

But of course the better strategy is to learn how to expose properly for scanning. It is much easier to get a scanner close to Dmin 0 than Dmax 4.

Keith

buze
12-Sep-2006, 06:44
Scanners /always/ scan as positive. When you scan as "negative" it inverts it for you afterward.
The only way to improve DMax is to do two passes; one for the shadows (overexposing on the scanner) then the highlights (underexposing on the scanner) then mix the two.
VueScan does that, but it does it badly (it fails to remove part of the blown highlights "bleeding" generated by the scanner).