Writing > ASMP

I was feeling pretty smug. After many hours of inspired work, the image on my monitor looked just the way I wanted it. A real eye catcher! I hit Save, ejected the high capacity disk from its drive and sent it off to my digital imaging service bureau (ust'a be photo lab) for output to film. Or so I thought.

The film came back looking pretty darned red compared to what I thought I'd created. Maybe another outfit would be more accurate. It turns out that they don't have a drive that's compatible with the type of disk I'm using. With a little help from some friends, we get that problem licked and I send the image over on another type of cartridge. This time though, the image on film is about 200% darker than expected.

Is it me, I wondered? Or is it them? Or my monitor? How in the world is anyone supposed to get reliable results if they vary so widely? And how much do the different kinds of systems being used for the scans and output influence the outcome?

I persuaded three local photo labs to participate in some tests and comparisons. We agreed that it was not a contest to establish winners or losers among them, but rather a chance to work together to determine how much variance exists and possibly establish some guidelines to help their customers get the best results with the least frustration.

To compare the systems, we agreed on a two part test.

First each of them would work from the same original photograph. Each would do their own scan and output to film, doing their best to match the original image without resorting to localized changes.
With a second image, we'd have yet another party do the scan, which would be delivered to each lab solely as a digital file for each to do an as-is output to film. There was to be no correction or the benefit of any reference image as a guide.

Both images were chosen on the basis that they had lots of reds, contrast, highlight detail and skin tones. The results were even more varied than anyone had expected.

The group that assembled to review the results, agreed the first image was acceptable from all three labs when viewed independently of the original. Since the whole process was tightly under each lab's control, they all produced satisfactory and usable images. Although one came very close, none matched the original exactly. When the red matched, the snow was off and so on. One was too warm, which isn't a cool way to depict snow. The opinions as to which was best depended on the potential usage. The image that stood out as considerably punchier in color and contrast was preferred for separations and print reproduction. A critical examination revealed that it wasn't quite as sharp as the other two though.

The second image was another story altogether, with the results running a much wider gamut. Since the labs had no control over the scan and simply did a straight uncorrected output to film, it showed just how widely the calibration of their systems ranged from one to another. The one that came closest overall was too warm; the one with the most accurate color was too dark. The third…well, it was in a class of its own.

What would account for such differences? All three used the same kind of scanner, but that was the only commonality. Each had a different film recorder and each used a different kind of film. The person operating the equipment has to be considered a factor too. As Paul Peregrine put it, "there's so much human involvement; so much human judgement. These are what someone's eyes view as good or bad. It's all still very subjective." How well the technicians do or don't see color undoubtedly affects how accurately their equipment is going to be calibrated and what corrections they will make to any image. And then there's the chemistry and equipment involved in processing the film. For good reason, results are likely to vary from one day to the next at the same facility.

Educating oneself before jumping in can do a lot to reduce frustration and costs. Before you start any digital project, talk to the technicians at the service bureau with which you plan to be working. Make sure that you feel comfortable in your ability to communicate with them or try again somewhere else. Discuss the whole project from start to finish and get their suggestions.

Make sure that any hardware you'll personally be using is compatible with theirs too. Especially ask about disk types and capacities. That's your means of porting images back and forth between your computer and theirs. Most labs can handle 88 megabyte SyQuest cartridges, but the 3 1/2 inch magneto optical and SyQuest disks are seeing an increasing popularity and most of the labs are now acquiring them.

The quality of the scan is critical to get the process off to a good start. Just as a photograph taken with a poor lens has little chance of being improved, a digital image needs the best origins possible. Imaging software can be applied to enhance the image, but it can only go so far. Add the costs and time invested for corrections and a poor scan gets expensive fast.

Ideally you should ask for a proof at the same time as you get any scan that's heading for alterations within your computer. Although it adds to the cost of the project, it gives you a chance to see how their film recorder handles the information contained in the scan. You may even discover that the scan itself is unsatisfactory and ask for a re-do. With a proof in hand, you can calibrate your monitor to match the output of their system and be much more confident in the end product.

In a perfect digital world, perhaps all the local service bureaus would regularly run co-operative tests and work their calibrations towards one another. Seeing our tests results emphasized the need. "It really behooves them to cooperate in something like this. It's quite a benchmark, Greg Shriener remarked. "How many opportunities do you get to go against your competition head to head like this." With everyone calibrated to the same tolerances you could confidently take your digital file just about anywhere for output. But it's not yet as reliable as getting film processed.

Looking at the great diversity Schriener lamented, "this defeats all the open architecture that everyone is striving for. We should be able to exchange files and get usable results."

In this less than perfect world, your best output results are likely to come from the same place that did the scan, since their scanner and film recorder are calibrated to the same loop. Getting the output elsewhere may require a few extra tries.

Doing your own desktop scans is less likely to yield high-end results, since what's typically considered to be high-end equipment when used in photo and design studios is barely entry level equipment for the service bureaus. A film quality output from a scanned 35mm image tends to come in at about 60 megabytes of data. Most desktop scanners are only capable of about half of that. If you do your own scans, invest some time in getting your system calibrated by outputting some test images through the lab with which you'll be working. Walking in cold with your file may put a chill your enthusiasm when you see the results.

On your own system, don't be switching the image back and forth between RGB and CMYK. In most cases you're losing valuable information with every switch and gradually degrading your image. The labs employ specialized software for just this purpose.

When submitting a file for output, provide a color guide for the lab to refer to if it's at all possible. David Sowers noted, "it's obvious that everyone needs an original or a guide. You can't just take your SyQuest around and say, 'here, output it'."

Once you're zeroed in, the excitement and potential of the digital era is yours to enjoy. There are no bounds, save your own. Hey, wanna kick? Try taking an image for some off road driving the next time you're moving down the information highway.