Archiving Images Part 1: Digitizing Slides

I began shooting slides for stock photography purposes in the early 1980s.  I freelanced as a motorsports photographer and did some portrait photography back then, but whenever I had the chance I was out "shooting for stock," I called it, hoping that I'd be able to market my images and make some money off them that way.  Over the years, as a result of this practice, I accumulated thousands of slides on a wide variety of subjects.  By the late 1980s I recognized the need for an improved method of retrieval so I came up with a simple date code using six digits: the first two were the year and the last four were the sequence in which they were taken during that year. I figured I wouldn't have a year where I shot more than 9,999 images. Heh. That was before digital, obviously. So except for my most recent slides, I have thousands cataloged in archival sleeves by number.  As it turns out, as time consuming as that process was, assigning numbers and putting them in sleeves was the easy part of the process.  Digitizing the slides and then building a database for them, as well as for all my other images, are the true time consuming parts of this process.

The process I will be outlining in this and the next few articles is where the images will be archived to DVD.  Since these will be archives, and since all this process is quite time consuming, it will behoove you to buy good, archival quality DVDs.  Very short lifetimes have been well documented with cheap DVDs. Archival quality ones should last for decades.

Modern digital photographers talk a lot about "workflow" these days and now that I have a DSLR and have become fairly active with it, I can see why. Because of the deluge of digital images that can be produced one needs to have a workflow design such that ones time is used most efficiently. Otherwise many hours can be spent needlessly spinning ones wheels. So I've spent a fair amount of time just thinking about this whole "workflow" business, and what I'll need to do to archive my images. Even more important than archival process is how my images will be organized, and how easy it will be to retrieve any given image. It is this latter point that will cause the initial setup of images to be a lengthy process. But once it's done, then the workflow becomes much more efficient in its operation.

So first off, I decided to start in on my slides.  While it is true that they are archived in the sense that I have them stored in archival sleeves, they were not archived digitally, so not easily retrievable and transmittable.  Assembling a quality digital archive of my slides has ended up being a long and laborious process, mostly because when I started out digitizing them I knew much less about the whole process of digitizing slides than I do now. I started out by scanning slides on my Epson 3170 flat bed scanner at 3200 pixels per inch (ppi). The 3170 began showing some color streaks in some of the images, thus degrading image quality, and I felt that the scans could have been sharper, so I upgraded to an Epson 4990. The 4990 is definitely a better scanner, but only by an incremental amount. Then a friend gave me an HP S-20 dedicated film scanner, so I gave it a try and found it to be on par with, or perhaps just a bit sharper than, the Epson 4990. But the Epson has a better Dmax than that HP so I felt it to be a toss-up between the two.

 Epson Perfection 4990

Still, I wasn't happy with the results I was getting because I could examine my slides with a decent loupe and clearly see more detail than any of my scanners were capturing. And then I found out about "digital" slide copiers. A digital slide copier is designed to mount to the front of a DSLR's lens, kinda like this:

In order for the lens to focus properly, the copier has an interior element, which can also be used as a close-up diopter lens for macro photography.  The element is reasonably well corrected, and the setup takes decently sharp photos.  Clearly sharper than those I was getting with my Epson or HP scanners.  But it was something of a pain in the butt to use, which I was willing to put up with, but I found upon close inspection with a loupe that there was still detail in the original slide that wasn't being transferred. 

I liked the slide duplicator idea, though, but in order to do it right, I began to suspect one needed to do things the old school way, like with one of these jobs:

The above is a proper bellows with a macro lens attached to the front and with a slide duplicator accessory attached to the front of the macro lens.  However, unless one has a full-frame DSLR (and most folks don't), slides will be cropped because of the magnification that occurs when using a "crop body" camera with 35mm-format lenses.  So that's out for me, at least, with my APS-C sensored Canon EOS DSLR.

So I started playing around with what I had, which was my digital slide duplicator and some other close-up stuff.  I wound up with a combination of the slide duplicator tube, a Nikkor 55mm f/3.5 macro lens, and a 10mm extension tube, plus a Nikon-EOS adapter so I could fit this contraption to my camera.  The setup in the following picture shows both the slide carrier and the roll film stage at the bottom.  I scavenged both of these off a 1980s-vintage zoom slide duplicator.

Finally I had an outfit that was capable of recording images that were critically sharp with my 10 megapixel DSLR.  This rig is clearly superior to any other device I've tried for digitizing slides. 

So now that I had the hardware straightened out, then I had to turn my attention to the method of capturing the images.  At first, I just took my slides outdoors and used a piece of poster board as a reflector and pointed my duplicator outfit at the poster board.  But I found that I was restricted to doing this at mid-day only, or else the color temperature drifted and then I had to try and color correct the slides, which wasn't always as easy as one might think.  I found that on overcast days the time of day wasn't as critical and that I could just point the rig up at the clouds.  On balance, these methods worked well, but since I was restricted to mid-day hours, I decided to set up an off-camera flash rig and give this a go.  I don't have any of the fancy TTL cables and off-camera adapters, or TTL flashes for that matter, that will provide TTL flash with my EOS DSLR.  So I just used another flash I had that has variable power settings, set it to its lowest setting, and just used standard PC-sync cabling and a single-contact off-camera hot-shoe adapter.  I was able to find out the optimum distance between the duplicator and the flash after only a few tries, and found that with some slides in which the exposure may have been a bit off, I could even adjust it somewhat by moving the camera either closer or farther away from the flash to correct for this somemwhat.  This was the easiest method by far, and provided me with images that had perfect color balance.

So then it was a matter of staving off feelings of drudgery, really.  Periodically, I would check the focus of the lens with a slide of known sharp focus using the camera's Live View feature.  I had to pay attention to horizon lines in the images and see if I could correct for them if they were somewhat off, if at all possible, by rotating the front of the duplicator tube.  I needed to insure that the slides were each pushed into the duplicator stage the proper amount, all the while trying not to bump the focus.  And repeat the process several thousand times!

Of course, I had my camera set to record the images in raw mode, which provides me with about 800 ten megapixel images with an 8 gb memory card.  Once the card was full, I downloaded the raw images to my computer's hard drive as temporary storage, emptied the card, and filled it back up again with more images.  Lather, rinse, repeat.

During this process, it might be worth your while taking some notes.  For example, I have my sleeved slides sorted according to subject, and when I shoot digital duplicates of them, I keep to that order.  I don't change the digital file names -- it would be way too time consuming to do so.  Instead, I record a raw image's file name at the beginning of each subject, which makes it easier to keep track of the various subjects when it comes time to transfer the images to DVD.  What I do is create a separate folder on the DVD for each subject and then dump all the files pertaining to that subject into its folder.  But before doing so, I will still double check that I've started the file count at the right file name.

After burning the DVDs, I use a Sharpie Marker and record the subjects of the folders the DVDs contain and any other information I consider pertinent.

In my next article I'll discuss archiving negatives and how they offer an entirely different set of problems.