Lou Jost wrote: ↑
Wed Aug 05, 2020 12:43 pm
You say you can't think of a scenario where a longer tube lens would be useful, but at the same time you note that a 200mm tube lens often gives poor corners on FF. The problem you note about edge quality is exactl ythe scenario that longer focal length tube lenses will fix for you.
As you know, if you are happy with the results of a 200mm tube lens on an APS sensor, you will get the same image, with the same corner quality, by using a 300mm tube lens on a FF sensor. The FF image will be better than the APS image because you are putting more pixels under each feature. Yes, some of this will be empty magnification, so the improvement won't be linear with the increase in pixels per feature, but as Rik and others have shown, there are still advantages to having more pixels.
It is a misunderstanding to think there is anything special about a 200mm tube lens, independent of sensor size. The size of the sensor matters. A 500mm-600mm tube lens would be appropriate for medium format. For FF, the natural focal length for a tube lens would be 300mm (given the experience of most of our community that a 200mm tube lens is a good fit to an APS sensor).
Sure, some objectives will not be able to take much advantage of the extra pixels under each object feature, but some will. The 20x 0.75 apo Nikon is an example. There are 10x 0.45 and even 10x 0.50 objectives.
The 200mm ITL200/CMH-200 yields poor corners on FF because it's not designed for FF, it's a crop sensor lens. I didn't say 200mm will often give poor results, I said the ITL200 will always give poor results because it's a crop sensor lens. I had the opportunity to test a fullframe optimised tube lens and it's better in every aspect, corners or centre. It would have been available if some kind of malware didn't hit the world.
In theory, a longer tube lens cuts out the nasty corners, that's based on a huge assumption. Assuming the 300mm TL performs the same or better than the 200mm counterpart. This is often not the case, from personal experience of course. More pixels is always theoretically better too, but in practice, it kind of doesn't matter. The competition is tough here, we have the ITL200 and the clone CMH-200 that's a tad better. Which 300mm lens out there can outperform this couple? Moreover, is spending at least $250 to try out say, a 300mm ai-s truly worth it when one can simply make a small panorama? Sure it will take more time... still less than trying out a myriad of different 300mm lenses to find one or two gems (which I respect).
An objective that couples well with a 12MP FF monochrome sensor will work nicely on a 24-36MP colour sensor, and yield more information on a 50MP colour sensor. This is always true, more pixels more information; however the amount you'll get depends on the NA and other aspects. A 10x NA0.28 will benefit from that 50MP, you'll get a load more information. A 20x NA0.42 however won't. In theory, there's more stuff; in practice, you'll see minimal improvements. This is something I think we agree on. A 20x at an astounding NA of 0.75 couples with a monochrome sensor of 64 whopping MPs, it should in theory blow a 100MP FF coloured sensor out of existence. If we pull this to 40x via a 400mm TL, it drops to 15MP which is still really good.
The drop follows a nice cubic function.
An objective like that is only good for transillumination/coaxial epi due to the WD, which I'm guessing is around 0.5mm... oh it's typically coverslip corrected too. Higher NA brings stacking errors or something, I've never studied this. Only recently I've been encountering such problems. A sunset moth stack turned the lines in the scales into little zig zags.
Due to the extremely short WD, nothing but coaxial epi works for reflected light.
All my experience points to one thing, always follow the manufacturer's specifications. Outside of that, it's unexplored territory. I used to use an ITL200 on my Olympus microscope. While it's true that the 20mm difference is minute on paper, I get weird distortions. I thought this was just the objectives being silly, mounting the U-TLU (180mm) showed otherwise. A 200mm TL, or 180mm, 165mm in the case of Zeiss for the matter is special, it's recommended by the manufacture, the engineers design the lenses to fit these FLs.
The same is true with the 2.5x HR and 5x NA0.28 (also HR) QV-objectives, which to manufacturer's spec, works on a 100mm TL. With a 200mm TL, the results are astounding. The former beats any 5x I have due to the exceptionally high NA, the latter is "almost" as good as a native 10x Mplan Apo. How about the corners though? Well, they aren't the best, the 2.5x HR outperforms my 5x objectives but comes with a lot of distortion; the 5x pushed to 10x is nearly indistinguishable to my native 10x. Of course, if one doesn't care about distortion, it's totally fine.
Longer telephotos have other problems beyond this. For example, Nikon's 200-400mm VR2 version is simply horrendous at infinity. At closer focus, it's very sharp. The newer 180-400mm fixed it to some extent. I can't speak for other telephoto lenses. The 300mms I've used are the F4.5 ai-s, F4 AF-s D version and F2.8 ai-s. Good close-focus results, but not so good at infinity. I guess those non-macro large format lenses will do a lot better as manufactures probably account for infinity focus. But then, the Schneider apo-digitar 120mm M version (M for macro) does better at infinity than the non-M version! See where I'm getting to... unexplored...
If I was tasked with designing a 300mm or 400mm telephoto for FF cameras, I'll have close-mid focus in mind; because when I think of those FLs, I think of wildlife, sports, and portraits for those defocus crazed photogs, landscapes ie infinity comes last. If it's a 100-400 kind of lens, then I'll seek a balance between close-mid-infinity, as such a lens is designed to be versatile, not a specialised optic.
I'm still waiting to see someone test their 100-400mm Canon RF mount or Sony E mount. These lenses are said to be very good at the entire range. Now Sigma offers a slightly slower, but cheaper 100-400mm as well. All this time though, I don't see the keen explorer. Nikon's S-line will be out by next year, but spending nearly $4000 AUD to buy a lens to try out objectives is simply out of the question. Loaning is also not an option, one might as well buy a new copy and then sell it. Typical loaner prices here are $80 a day with seven days minimum!
It's totally fine to explore these things. I'm in the middle of finding a good 100-120mm TL, tested several so far with sad results. The Gretag just flew in and I'm getting an adaptor made soon. I will however not recommend outside of the manufacture's spec sheet, especially when it comes to beginners, until there's sufficient information. For instance the raynox, it wasn't "designed" to be used as a TL, but there's plentiful information and tests to show how well it does, and mounting is trivial at this point; I would not have recommended the raynox 4 years ago.