6 posts in this topic
My Epson V700 Photo has scanned its last neg. I had bought it after my Nikon Coolscan 8000ED stopped working and Nikon couldn't supply parts to fix it. As expensive as the Nikon was, and as affordable the Epson was, neither ever managed to consistently deliver edge-to-edge grain-focused scans - the only way to get that with the Nikon was by sandwiching the neg between glass in a glassed carrier, but even anti-Newton ring glass still produced those image-destroying interference artefacts.
Lately I became aware that the graphic arts people are using LED light panels (like the illumination panels of monitor screens) as a cold backlight source for tracing and the like. They're cheap (A4 size around $20) and the light is consistent and even, and with the panels also being perfectly flat it was a no-brainer to pop one onto my copy stand and use the Sigma sd Quattro-H with 70/2.8 macro lens as the "scanner". I know this technique is nothing new, but previous contraptions of earlier times generally used globes in a lightbox arrangement which were never perfectly even, and the heat the lights generated would encourage film to warp and curl uncontrollably if not under glass (and therefore subject to Newton rings). I'm sure that others have done what I did here, but I thought it might be worthy of mentioning anyway.
Anyhow, the setting up of my new "scanner" couldn't have been simpler - just mounting the camera on the copy stand as usual and plugging in the LED panel was all that it took, and I was away and scanning negs far quicker than I have ever done. I used the glassless plastic neg carriers that came with the Epson to hold the film, but should I ever get film that won't lie flat I still have the solid metal neg carriers from my Durst 1200 to keep things flat.
What no neg scanner I have used has ever provided has been an ability to manually focus on the neg, or had an adjustable aperture to stop down and increase depth of field should the neg still be a bit bowed. This setup fixes those two things perfectly.
What also came to light was something I had suspected for a long time - that the Epson certainly wouldn't have any lenses that could be described as first class, and the barrel distortion evident when flicking back and forth between the two clearly confirms that suspicion.
As for working - well what a treat it was to get scanned negs that had crisp, sharp grain uniformly over the whole image. Recently I had taken a shot with my now defunct Pentacon Six and was disappointed that the scanned neg just didn't look sharp - I blamed my incorrect focusing or maybe slightly shaky triggering - but I wasn't at all pleased with the shot because of that softness. So I decided to compare the Epson scan with the new method, and below are the whole images in reduced size, along with two 100% sections. The results speak for themselves, I hardly need to indicate which was the Epson (first) and which the Sigma (second). Note the evenness of the grain in the Sigma scan as compared to the mush of the Epson, which clearly missed focus in its attempt. I have now copied over 60 negs and absolutely every one has been perfectly sharp corner to corner. I've never had that reliability from a scanner.
...and the setup:
These 6x6cm negs deliver cropped square files of around 17MP/100MB in size, which of course is pixel-for-pixel with the Sigma, no softening interpolation as from any Bayer or X-Trans sensor being necessary. Blowing them up they easily make 24"x24" with no pixelation visible, and still hold together well at 30"x30".
A while a go I asked about scanning old slides with a flatbed vs. using a DSLR for duplication as I had sold my slide scanner a few years ago. Silly me had forgotten to scan the 1999 South Africa trip before selling it. All answers pointed to using the DSLR so here are some observations:
Input is from Fuji Sensia slides in Reflecta frames (In contrast to the US, it was very economical to shoot slides in Europe in the '90s, I could get a roll of Sensia with development and framing in proper frames (not cardboard!) for less than $5). They had been stored in boxes in magazines holding 100 each. The boxes are not airtight, so some dust was to be expected. The sturdy frames made it easy to insert them into the copy adapter. I obtained a used ES-1 slide copy adapter and BR-5 step down ring from Mike Gorman (thanks Mike!). The step down ring is needed to mount the copy adapter onto the AFS 60mm macro lens. Even with the ES-1 in the closest position, the slide will not fill the whole frame, so I get 20MP or less (too lazy to really calculate it). If I remember correctly, the adapter was desinged for a 50mm or 55mm macro, not a 60mm. I initially wanted to use an LED panel as a light source, but it was too weak to provide illumination for F11 at safe shutter speeds, so I only used them for focusing and the key light source was an SB800. F11 at ISO 100 with the flash near the lowest power setting. WB set to flash. On very dark slides (sunsets) I increased the ISO to 200 (too lazy to change the flash output, I could set ISO with a mouse click). I fired the flash with a radio trigger (Pocket Wizard). I used qDSLRDashboard to tether the D750 to my PC and set Capture One to monitor the incoming folder.
I used a rocket bulb blower to clean the slides before putting them into the holder. Initially I used live view on with AF all the time, but that turned out to be a huge battery drain. With F11, the DOF is sufficient to fix the AF once and be done with it. So I ran this blind. In contrast to using a slide scanner or the Epson flatbed, the setup kept me busy at all times, constantly exchanging slides and then pressing the shutter (via mouse click on computer).
With a scanner there is always a significant wait time between the scans (it was several minutes with the Canon FS4000), especially if you use multi-pass scanning with an additional dust removal scan. In the end, the total time spent to get all slides scanned is significantly less with the adapter than with the scanners. I used exiftool in batch mode to change the capture date in the resulting NEFs to approximate the date the slides where shot. The flatbed Epson V550 Photo is not much worse than the Canon FS4000 slide scanner I owned previously, but faster and does not require a SCSI connection. The difference between 3200ppi and 4000ppi is pretty much irrelevant, both show the film grain.
So what's the verdict on using the DSLR with the copy adapter?
Vervet Monkey in Krüger Park, 1999:
100% screen shot of DSLR copy on the left and Epson scan on the right (the scan would need sharpening).
Color: Much easier to get accurate colors with the DSLR than with the scanner, even when using IT8 calibration targets.
Accurate is still subjective of course, you get the exact color of the slide ;-) Sharpness: The DSLR wins, but not as definite as with color, the scans need more sharpening than the NEFs, but sharpen ok. Highlights: With the DSL there is much more headroom to fix highlights than with scans. Exposure was set so that there where no blown highlights in the copies. Noise/Grain: Both methods show the film grain, but depending on the scanner the scan can be noisier.
I have no noise with the DSLR, only film grain. And still no perfect tool to remove it ;-(
I guess I need reprofile my old copy of Noise Ninja. So far it was too drastic. Dust: Well, without ICE (the infrared dust scan) there is dust even after fastidiously using the blower. But it is only noticeable in relatively bright areas like the sky and quickly dispatched with the spot remover of Capture One.
F11 makes dust bunnies on the sensor easily visible, so this lead to a sensor cleaning session... Cost: If you get the copy adapter used, the cost is negligible. Film scanners are quite expensive used and one needs to sell it after use, way too much hassle.