Registered: Oct 08, 2010
Total Posts: 1634
Country: United States
The NEF files from the D800 are much easier to deal with than the .tif's from the Nikon scan software.
what exactly do you mean with the part of dealing with these files? I am in the middle of scanning about 15,000 slildes to TIF, opening in Camera Raw as if they were a NEF file and the result is simply amazing compared to what the same scanner did 13 years ago. Disk space is virtually free these days (just added another external 2TB backup unit for $90), so file size has really no drawback. I decided to keep the master TIFF files of most images, in case there's better software to process them one of these days. I also use a stack feeder, which lets me scan images unattended.
How do you deal with dust, scratches and any other imperfection? Can't run ICE software without the infrared scan layer. I would spend 10 minutes to one hour cleaning up a single image if the software didn't handle that. I stopped scanning black and white negatives, because it is simply mind numbingly impossible to clean up all the scratches and spots and other flaws by hand when scanning something the software can't clean.
Well, I did not like the original scans with the 5000ED because I had done them with the slide feeder and auto exposure, hence it did not get the exposure nor the white balance correct on many of them. Plus, these slides were in excellent condition with plastic mounts that did not shed copious amounts of dust like the cardboard ones do. No scratches, fingerprints, stains, and minimal dust made it possible.
Now with the D800 I was able to check the exposure, etc., in live view to get it just how I wanted it and then run things in post as necessary. Sometimes NX2, other times PS6 using bridge for particularly tricky images. I also could make perfect duplicate images with different exposures to do a bit of HDR if I really wanted to get all there is in the slide - something more difficult to do with the 5000ED. It also takes less time per slide to image it, though it take more of MY time because of the lack of a auto feeder for the slide holder. If some enterprising person could make an auto feeder that can be hooked up to at D800 + macro lens and automate the imaging process would be fantastic.
Also, the flare inherent in the 5000ED is really troublesome with snow images that have very dark objects in them (like people standing in the snow with black pants on). The 55/3.5 + D800 gives a much more flare-free image in such situations.
I would not try this with some of my older slides as they were far too scratched and dusty to make it worth it. I may try again with a more collimated light source to try to get the sharpness up a bit at the expense of additional grain.
Some time ago I did a comparison between the D800 + 55/3.5 vs. 5000ED. The DR was about the same - much better than my D700 - but the flare was absent with the D800 + 55/3.5 macro.
My slides were in archival storage, pro lab developed and most of them never out of the archival page since put in there 13 years ago. when done scanning, Viewscan shows what it had to fix in the ICE process and it is horrifying to see all the crud that's on there. Most of it can hardly be seen, but I doubt I'd have the patience to correct any image in that level of detail. There are a few hundred cardboard mounts - had to use whatever local labs offered us when they let us process film late at night on a weekend to get stuff online, but for the most part, I ran my film through my local pro lab.
White balance, and everything else I leave up to the software post scan. Each roll is different, some were push processed and look all muddy and blue, etc, so I batch scan at a pretty generic setting, but all files are manually tweaked in Camera Raw, as if they were NEF files. Gets me the dynamic range of the scan and full control over white balance, which even with the very brightly colored subjects I used to shoot is pretty close for the most part. Being always the same subject helps me there to correct manually. Also, the slides differ so much between batches, there's no way I would have the time to correct it in the scanning process - each slide takes 20 minutes to go through (dual exposure, 12 passes), and I don't have the time to sit there and do that for each when it comes around to scan the next one.
Vuescan always does multiple exposures, which for slide scans is the only way to really get some detail out of the shadows. My 48 bit files feel like they have an extra 2-3 stops of dynamic range, especially in the shadows, which in almost every case I have to pull up, while highlights are pretty close to maximized right from the scanner. They aren't D600 files, but then there's also not that much detail in the shadows on film to begin with
The auto feeder on the Coolscan is the only reason I am even willing to scan my old slides. It is just a daunting task to manually feed each slide, which I did back in the 90s using a SprintScan 35. That was seriously tedious, not even counting the lack of ICE cleanup .
Flare is a problem I also have - it only happens at certain angles of high contrast lines, and it was much much worse before I cleaned the mirror in the scanner (I do this every 1000 scans now, seems to keep it as good as it gets). Still, some images need some manual local adjustment brush work to control the bleeding of light into shadows. Maybe 1 in 20 images displays a noticeable halo effect or whatever you want to describe this as, and maybe 1 in 100 needs some work to fix it.
well, flare only comes up visible in my scans when I really lift shadows and go for some HDR-like recovery of the darks next to a very bright harsh edge that goes vertical across the 35mm frame with the dark on the left side of the bright edge. Take a look at my recent Flickr uploads to those sets from 2000 and 20001 and I doubt you'll see much of that in there. All those are perfectly preserved slides, some of them straight in the archival page from the lab and never pulled before I scanned them, and all of them have scratches right form the developing machine or mounting machine, as well as thousands of tiny specs the IR light picks up and cleans up.
Anyway, I don't have macro lenses, and my volume needs are much higher than yours. I have years of pro work to digitize and only a bulk feeder can get me through that part.
The flare really isn't an issue big enough for me to worry about, at least after cleaning it seems to only show in extreme cases such as here (not corrected, see halo around roll bar), but I could correct this in a few minutes of post if I really cared, or just mount this slide with a wet scan kit if it really mattered to get it as good as I can get it. These scans I am doing right now are almost as good as what I got from a $30,000 drum scanner in the year 2000, and definitely better than a professional flatbed Scitex scanner my local lab used at the time.
The difference in what I got out of the same slide back in 2000 versus today, mostly based on better software and I guess more understanding of what needs to be done to get the most out of a scan, is simply stunning and I'm quite happy with what I get. The next step up in quality would really require to manually wet mount every single slide, and that's just not worth it to me.
This scanned in 2000 on Polaroid Sprintscan with Photoshop 6.0 (I think), single file scan manually tweaking things
This scanned last week on 4000 ED, VueScan, Camera RAW in CS6, in a batch, generic scanner settings, then adjusting everything in post.
Sadly, none of these linked images were created with Nikon gear