Hi guys - I recently decided to try DSLR scan (well, technically in my case and in most people’s case nowadays I guess it should be called mirrorless camera scan).
I have a Sony A7III but I don’t have a macro lens. I have a 55mm f1.8 and an old 55-210mm lens I kept from my NEX5 days. I bought a set of 10mm and 16mm extension tubes from Amazon but still struggle to get close enough to the film (either 35mm or 120) to fill the frame/sensor.
What lens do you use with your Sony mirrorless? And if you are given the lenses I have, other than extension tubes are there any other ways to get true “macro” scans?
The lens equation (find it on this page) says, that a thin (model) lens with a FL of 55mm would need an extension of 55mm to get to a 1:1 imaging ratio. Modern lenses might need other extensions, because they are nowhere near to be a “model, thin” lens. It’s best to not stack a lot of thin rings, but one single long tube instead. Your lens manual might say, which extension you need for 1:1 imaging (or whatever you need).
Whether your 55mm lens will deliver good macro images is hard to say. You’d need to be more specific about what the lens is exactly. The full model name could help.
I’d advise to NOT invest in scanning with a zoom lens, but if you want to test with longer extension tubes that you’d need anyway, you can always do it.
I am capturing negatives with a Sony a7R4, full frame 60MP camera. You can see my full write-up about the detailed considerations and solutions I adopted in my article on PhotoPXL.com. (Digitizing Negatives with a Camera – Revisited – PhotoPXL). What’s most important is the design of the lens you use. Ideally it should be a macro lens designed for flat-field copy work (like a high quality enlarger lens for example). Bellows is better than rings because it gives you much greater flexibility for accurate scaling.
I use the a7iii as well and started with extension tubes for my Tamron 24-70. I struggled with consistently of focus, too. I decided to buy a macro lens and picked up a Sigma 70mm macro lens. It works perfectly with my copy stand and Kaiser LED light box.
While the tubes were inexpensive, I just didn’t think they were able to give adequate sharpness. The macro lens wasn’t overly expensive and I can use it away from the scanning job.
When you say the Sigma Macro lens works perfectly, what exactly do you mean? Have you tested for whether it delivers corner to corner sharpness focusing on film grain or using a fine resolution target?
I first determined the optimal focal length for the lens (f/13) and then used my magnetic base (which has various markings and several items both printed and photographed. I took samples of each at f/13 and compared them in Lightroom.
For my purposes, the Sigma macro lens produced acceptable sharpness even into the corners. It isn’t perfect but the results, for me, were acceptable.
Perhaps my use of the word “perfectly” is what caused your response. My meaning was that the copy stand, camera, lens, and lighting setup produced, let me use a different word, good results.
I am not printing all of my images and they are primarily done to share old images with family. We are spread out across the country now and the only way to share these images is electronically. In that environment, the camera/lens combination produces more than acceptable results.
If I was planning to make large prints of my old images, I might need a better scan. However, large prints are inherently limited by the print mechanism so I don’t know how much one actually gains by producing ever more precise scans.
By they way, let me thank you for your posts. They are always concise, precise, and interesting.
You have satisfied my curiosity, and all of that makes perfectly good sense. One needn’t take things beyond what suits the purpose.
I should mention - I routinely make large prints (17" width by lengths from 22 inches to 15 feet) using an Epson SC-P5000 printer - still the machine with the widest colour gamut and highest resolution of any machine in this class of professional printers. The inconvenience of it is that sometimes one sees things in the prints that one misses on the display. These kind of prints do challenge the quality of scanning systems - that is, the binding constraint on output quality is not the printer (if one has a really good one), but the material being fed into it.
And glad you are enjoying my posts - makes the time and effort worthwhile. I also enjoy and learn from these discussions.
I don’t use Sony but if I did I think I’d be looking at a 70mm Sigma Macro as well, there are two versions but both are recommended widely on here, the ART version is probably slightly better. Robert O’Toole seems to like them both also:
I was equally surprised. However, I did a series of shots at every aperture on a variety of targets and f/13 was the consistent winner. I read some books and checked some YouTube videos on the subject and the consensus was to actually test your lens versus just picking something like f/8.
Perhaps it is just my lens (individual variations are always possible) or my old eyeballs! Regardless, I always suggest doing the sharpness test first.
The products made by the companies you mention are undoubtedly of very high quality. They are very high resolution targets etched on glass, with a scientific level of consistency. They are absolutely what you need for critically measuring the actual resolution of a setup, although usually only at the centre of frame.
However I would argue that most of us don’t need to know this and that Vlads Test Target is actually more akin to our needs. The fact that it is on film (albeit very fine grain) and shot with a camera rules it out for purely scientific measurements but that is a positive advantage for us as it closely mimics what we will be photographing ourselves, namely negatives & transparencies. Vlad is up front about this, clearly not all the bars can be resolved but it goes considerably beyond the capabilities of any lens that I’ve seen the results from. Crucially also the patterns are repeated at the corners and centre edges, there may be etched glass targets that do this but I haven’t seen one.
So, in short, it is ideal for determining the best aperture to use for a particular lens, for comparing the results from different lenses, and also for confirming absolute alignment. As I’ve mentioned before Vlad has an Airtable where people can upload their results so it would be great to see some from the very high end systems that some, including yourself, have shown on here.
I’d quite like it to be extended to include the results from actual scanners as well in fact, it would be a useful resource to compare, say, an Epson flatbed or a Minolta 5400, as it’s on film I can even put it through my Imacon Precision II, but that would be something for Vlad himself to endorse.
I have nothing to do with the company but I do admire the product.
I was responding to your statement that Vlad’s target is very good for “confirming sharpness”. Well, OK, there’s “sharpness” and there’s “sharpness”. When I think of sharpness I think of an objectively measurable photographic property and the tools allowing us to measure the greatest extent of its ability to be reproduced. When I hear “confirm” it denotes conclusiveness. Insofar as your post above alludes to a more subjective concept focusing (pardon the pun) on what is sufficient for an individual’s needs or taste, that is conceptually different and I would always defer to the common sense of not buying stuff beyond one’s self-perceived needs or preferences.
Some years ago, on the launch of the Epson B850 scanner, I wrote a fairly comprehensive article published on Luminous-Landscape that provided test results from all the scanners you mention. That article is available here: https://luminous-landscape.com/wp-content/uploads/2015/02/The-New-Epson-V850-Pro-Scanner-Final.pdf. Done back in 2015, the target used for the measurements was a LaserSoft Imaging USAF1951.
You can run through an Imacon scanner any target that can bend without breaking.
I have also not found a glass-etched target that repeats the high-end resolution bars at the four corners. That would be a good, handy, product to have, but the workaround is to use the current ones by moving them around the image frame and making repeat exposures. This allows one to determine alignment issues and make the necessary adjustments. It takes a bit of time, but it’s workable and quite accurate. I just completed such a re-alignment exercise this week; quite amazing, when viewed under high magnification, how small the adjustments can be in order to restore an alignment. But again, different people will have different needs and standards with respect to the extent of magnification they work with.
I can imagine that if you are wanting to write articles about it you might want to quantify your results but it’s wrong to imply that anyone that doesn’t want to do that is somehow simply accepting lower standards, happy to bumble along. As I said Vlad’s target is perfect for comparing lenses, optimum apertures and corner-to-corner performance and those results might be stellar or poor, that’s really got nothing to do with what I was saying. If you were to tell me that your undoubtedly excellent setup revealed everything that Vlad’s Test Target was able to throw at it that would be interesting in itself.
These high-end targets are etched on glass I think, not that easy to bend to go around the ‘Virtual drum’ of an Imacon. I shall read your article with interest.
I’m not personalizing any of this. I’m just saying - to put in the simplest possible terms - that if you have two targets, one capable of showing lower resolution than another and you are satisfied with the lower resolution, you are therefore OK to be satisfied with that target, and that’s fine. I’m not passing judgment - everyone should buy what works best for themselves. My original intent was simply to point out that there is better out there, and depending on what people want or need, that can be useful for them to know if they don’t already.
I have a copy of Vlad’s target; at 1x magnification ratio it is not capable of demonstrating resolution beyond 62 lp/mm on my system, whereas the Thorlabs returns 112 at that magnification, or between 166 and 180 lp/mm at about 2x magnification ratio, simply because it is far more finely manufactured. I need to be able to assess more resolution than I found Vlad’s target capable of telling me about, because I am digitizing 35mm colour negatives for being able to print them with fine detail at sizes ranging between 13 x19 to 17x22 inches. I recognize readily that many people do not place these kinds of demands on their film digitizing, but for those who do they should know about the analytic tools that will be more useful to them.
And right - you cannot bend a Thorlabs target in an Imacon set-up - it’s incompatible.
That’s what I am trying to get at, if I understand you correctly you are saying in effect that your high-end copying system out resolves Vlad’s Test Target so that were you to upload a full resolution scan to the Airtable it would eclipse any results uploaded so far. As I also said that would be interesting in itself. Nothing uploaded there so far gets anywhere near doing that in the centre, let alone at the edges and corners.
I have only had time to skim quickly through your article, it looks like just the sort of comprehensive review that I enjoy reading. In the UK we used to get very good reviews in the weekly British Journal of Photography, I saved quite a few, but then that went monthly and transformed itself beyond recognition.