One of the great things about an all digital creative workflow, is that – due to all that binary malarkey – once an image has been captured, there is little reason for that image to degrade in any way.
But with an analogue workflow, there are a few extra stages to go through. Each of which can have unpredictable results on a photographic image.
Take scanning, for example. Once upon a time, the only way to get a photographic image into a computer, was to scan it. Ideally, this would be a transparency (or slide). But otherwise, it would was a printed photograph or even a negative.
When I worked in design, we wouldn’t let a digital image near a professional design workflow, unless it had been scanned with an extremely expensive drum scanner. Although, eventually, a few flatbed scanners improved enough to allow some of them to be good enough for non-critical commercial print work.
For those of us still dabbling with film photography (and who don’t use wet printing techniques), there isn’t a lot of choice of affordable scanning options. Drum scanning is almost non-existent – and it’s always been very expensive anyway.
The various photo bureaus will offer scanning as part of their film development service, but the quality is sometimes not very good (often over sharpened, high contrast and low resolution) and expensive.
So the use of ‘home’ scanners is often the best option. There are flatbed scanners, which can scan film using adapters. But I ended up opting for a dedicated film scanner.
I like to print fairly large – so I went for a model that could scan optically to 7200dpi. And, as many of us know, optical resolution is the key term, when buying scanners. But one thing that I had never hear of, until I came across this post, was measured resolution.
I do remember being told, years ago, to always scan at the highest resolution, as reducing a scan can have the effect of ‘sharpening’ it. But I have also read, recently, that it might be best to not scan an image at the highest setting that a scanner is capable of. Which prompted me to carry out a little test. Now, I’ve no idea if the results that follow have anything at all to do with measured resolution and this is far from a scientific experiment, (your mileage may vary and so on and so forth). But I’m always interested in seeing how I might be able to improve the scan quality of my, comparitively, cheap home scanner.
So, I scanned an image at the 7200dpi maximum of my scanner and reduced it in Photoshop by 50%. I also scanned the same image at 3600 and then compared the two. The film used was Agfa APX 100 and the images have been scanned with VueScan, with filters off. No other adjustments were made in Photoshop, (other than resizing them again for the web).
The most surprising, and noticeable, thing was that the image scanned at 3600 had quite obvious ‘blown’ highlights, and perhaps even banding in the gradient transitions?
The image which had been scanned at 7200 and then reduced in Photoshop by 50% (using ‘bicubic sharper’ setting), had a much more gradual transition in the highlights.
It’s hard to see in a fuzzy web optimised JPEG, but the 7200 scan, was also slightly sharper (although possibly also a bit noisier).
So, I’m not sure to what degree this proves anything scientifically. The results here could simply be unique to my particular scanner sample (although, I did run the test twice). Also, if I had more time, I might have experimented with a few other resolutions (and other images). There may even be different results with different films – I have noticed that some film types seem to scan more successfully than others.
But this test does seem to fit in with my subjective experience of using this scanner. Scan as large as I can and reduce, as required.