Monday, 18 February 2013

Scanning Old 4 x 4 Super Slides — Part 2

Following on from the last post on 4 x 4 cm superslides, I scanned those in glassless plastic mounts in the Epson V500’s frame for 5 x 5 m (2” x 2”). For the transparencies I removed from edge-bound glass mounts, I left the thin paper or foil mask around them. For those I removed from the glazed plastic mounts, I simple dropped them onto the bed of the scanner, fully expecting to see Newton rings. But I didn’t and so carried on.

One problem I did have early on with the V500 was sometimes getting purplish-pink vertical lines on the scans. A quick Google search showed this problem is caused by dirt on the first few inches of the scanner bed. A good clean of this area was the answer; cleaning the area to be scanned was not sufficient.

At the outset I set out what I was trying to achieve. I did not want to have to scan these transparencies ever again. Therefore, I decided to scan at a resolution that would yield a reasonable-sized print in the unlikely event that I would want a print. Most of my old photographs are viewed either on the iMac screen, iPad, or on the television either as HD video or DVD.

I decided to scan at 3200 dpi. Printing at 300 dpi would give me a 15” x 15” (38 x 38 cm) print from these superslides.

Incidentally, I have found a lot of utter rubbish written about the scanning capabilities of the Epson V500 on internet fora. Some people have claimed (without providing any evidence, visual or historic) that the maximum ‘optical’ or ‘hardware’ (i.e. non-interpolated) resolution is about 2400 dpi and that anything above that is interpolated. However, the Epson website clearly states that the optical resolution of the V500 is 6400 dpi.

I did actually check that I was getting better resolution at higher dpis by enlarging in Photoshop the scans at 1200, 2400 and 3200 dpi to the same size (in cm) as that reached at 4800 dpi, with all set to 300 dpi. I then cropped a small area of detail from each, copied them to an A4 sheet and printed them on glossy paper. Resampling for the enlargements was kept at bicubic throughout. My argument for doing this check was that if 2400 were the maximum optical resolution, then scans at 3200 and 4800 should appear no better since they would have been resampled in scanner. In fact, at this enormous final size (nearly 2 feet or 60 cm square) there was a discernible difference on the printed output. 1200 dpi, not surprisingly, gave the poorest resolution since interpolation to the final size was greatest. Both 3200 and 4800 dpi were sharper than 2400 dpi but I really could not separate 3200 and 4800. Clearly though, 2400 is not the maximum optical resolution in this scanner.

I should stress that these comparisons were made on a print at 300 dpi. Unfortunately, at screen resolution the small differences cannot be seen and so it is not worth trying to show the print in the space available in a blog post.

I scanned using Epson Scan software in ‘Professional’ mode. I compared in trials 24- and 48-bit but found no difference in output. So I settled on 24-bit, with a considerable saving in file size. I also tried various settings for unsharp mask, grain reduction, dust removal or digital ICE. Here is one comparison:

Slices of scans of the same slide scanned in the Epson V500 (Epson Scan software)
A. No additions
B. + Unsharp Mask - Medium; Grain Reduction - Medium; Dust Removal
C. + Unsharp Mask - Medium; Grain Reduction - Medium; Digital ICE
D. + Unsharp Mask - Medium; Grain Reduction - Medium; Digital ICE; Colour Restoration

For these old slides, I eventually settled on: Unsharp Mask - Medium; Grain Reduction - Medium; Colour Restoration and Digital ICE (Quality or Speed, depending on condition).

The .tif files generated (63+ MB) were imported into Aperture for final tweaking.

I said in the earlier post that my Superslides were either Agfacolor CT18 or Ektrachrome X. I shall return to CT18 in another post but there is no doubt that the Ektachrome X slides have been the better survivors overall. However, CT18 slides survived either extremely well or fared badly in relation to the development of a purple cast. There was nothing in between. By contrast, the Ektachrome X slides seemed to have survived with little or no colour degradation. I was not that keen on the colour rendition of Ektachrome X in the 1960s (a cold, blue film). The irony is that they look better after scanning and with some colour correction than they ever did in a viewer or projected on a screen.

I have been able to improve on the originals by using gentle manipulations in Aperture or Photoshop. Details have appeared from shadows, for example, while highlights can (sometimes) be held in check.

Finally, two scans of superslides taken with a Rolleiflex 4 x 4 in Macau in March 1966.
Macau. March 1966. Agfacolor CT18. Rolleiflex 4 x 4
A photograph extending from deep shadow to direct, sub-tropical sunlight
on the wall

Macau. March 1966. Agfacolor CT18. Rolleiflex 4 x 4