4x lenses and objectives shootout

[ChrisR]

Unfortunately the graphs on that site don't appear to have comparable, or even enumerated axes, in many cases, so comparisons are guesswork.
Subjectively, I'd say there are detectable differences between the various enlarger lenses, of the order of "I think that's a a bit better". Moving to a 4x NA 0.2 objective or similar though, it's more of "OMG".
Don't expect too much!

[rjlittlefield]

Unfortunately the graphs on that site don't appear to have comparable, or even enumerated axes, in many cases, so comparisons are guesswork.

Chris, could you clarify your concerns about the graphs?

When I look at http://coinimaging.com/s50comp.html, all the axes appear to be well labeled except for y-axis "MTF" on the first graph. I view that first graph as being just a relative indicator to answer "what's the sharpest aperture?", justifying the use of f/4.7 in subsequent tests on that page. Similarly the page on EL Nikkor 50mm f/2.8 at http://coinimaging.com/nikon_el50-28n.html shows f/5.6 as sharpest, so that's what gets used later.

In the graphs farther down the pages at coinimaging.com, it looks to me like there are enough labels to figure out the actual data values, though I notice that the variable scaling of different graphs makes it difficult to do a direct visual comparison.

To check, I pulled into Photoshop the graphs for the EL Nikkor 50mm f/2.8N and the Componon-S 50 f/2.8. Then I overlaid them with transparency and checked to see if I could resize them to permit direct visual comparison of the curves. Here are the results:

Resolution and sharpness, slight advantage to the Componon. Compare the red and blue curves, noting that the MTF numbers now line up on the left.



Loss of corner sharpness, significant edge to the Componon. Compare the green curves, noting that the loss-of-sharpness percentages now line up on the right:



What other issues are you seeing?

--Rik

[Blame]

The graphs at coinimage.com could be better laid out. However there is a lot of good data there and I can only applaud the creator for sharing it.

I do have to agree with ChrisR. The results from the Nikon 4/0.20 microscope lens make it clear that reversed enlarger lenses are giving far from perfect results. The Mitutoyo lenses could be expected to deliver much the same. There is still potential for a better solution between the optimal effective f/11 1x of the sigma 70/2.8 lens and the 0.14NA=f/3.5xMagnification of the 5x Mitutoyo.

Surely we can find a more economical solution than what Mitutoyo is asking for a f/10xmagnification 2x or Canon for the MP-65 and a f/2.8x(magnification+1)? I'ts not as if these are particularly challenging apertures.

[ChrisR]

clarify your concerns

Thanks Rik that does help - probably not just for me.
The site has improved, since earlier looks which prompted contact with the author - unanswered.
I agree entirely, there is a lot of information presented on that site.
[Edit - having looked at the older material: a number of the issues which caused my earlier consternation remain, so if you're new to it, I would urge to look around all of it to see how he goes about things in later additions. There are still some difficulties.]

Having asked for numbers this may sound contradictory, but I'm uneasy about reading too much into the differences reported between very similar lenses, whether numerical or lines on a graph. I remember when the author said he had six enlarger lenses to compare. Only six altogether?
The chance of getting a relative dud is quite high from what I've in seen in my clutch, then different real subjects look different in different light, and can show detail more or less well even within the same image.
Sample variation is hard to imagine if you don't have several of anything. I have duplicates of a number of lenses, but only tend to remember a favourite. There was a time when some 50mm El Nikkors were under $20. Of about 7 of those, two are below par.
Microscope objectives we know can easily be substandard even without suspected intrusions.
But I have two 4x NA0.2, one with a chip, which I find makes no difference whatsoever to its images.
The Canon MP-E is sometimes reported as best at full aperture, sometimes not. I doubt I'll be getting more than the one of those, because mine is best at f2.8. But maybe I need to try...?

[Blame]

ChrisR

Steady on. It is not as if any of us get research funds. Making decisions on rumor, gossip and possibly unreliable data neatly laid out on graphs... well these thing are part of life. Hell, some days we even put faith in the manufacturers data.

If somebody gets a dud and honestly reports it as a poor lens then so be it. That is why one should not trust a single source. Rjlittlefield and that unfortunate mitutoyo 10/0.28 comes to mind. It was regrettable but it didn't stop the rush because so many others were posting excellent results.

He was not responsible for the products reliability and nether is any of us so by all means report your MPE as good wide open and leave it at that. Anybody who finds theirs different is free to say so.

[ChrisR]

Steady on.

All I'm saying is, the reader should be a bit careful.

That is why one should not trust a single source.

Absolutely.
There is the temptation, however, to put more credence on information which is presented more fully.
It becomes very easy to believe that lens X is better than lens Y because a fellow with some fancy technique says so - even though that may be from one sample.

If somebody gets a dud and honestly reports it as a poor lens then so be it.

Sure, but if you don't know it's substandard, and produce a detailed analysis..? Readers have to "be a bit careful".
We don't know anything about Coinimaging's sources, unless I missed it. We can guess they aren't all new out of the box.

It was Rik's Mitty 10x which was one history contributing to what I wrote above. (Even!) Rik took a while to realise that something was amiss, I believe. A new user may never have realised, if he'd bought the lens after whatever trauma befell it.

by all means report your MPE as good wide open and leave it at that.

That's all I've done - all any of us has done.

Anybody who finds theirs different is free to say so.

Of course, but when it's in the context of a website full of tests of similar lenses, the situation is different.

There is only one site I remember which routinely shows results taken from a large number of samples. A surprising spread can appear, on the test bench. How much of that shows in real subjects, is the other question.

[Blame]

ChrisR

Uh. This site that " routinely shows results taken from a large number of samples".

That wouldn't be lensrentals.com would it?

Well, they have lots of samples because they rent them. They also have good business reasons to test them all. It's sort of the exception that proves the rule.

Anyway, we are arguing at cross purposes. You are right in that the data is far from reliable and I am in saying we should be grateful for what we can get seeing as it comes for free. We are both right.

Right now I am trying to extrapolate enlarger lens performance at infinity, or even a diopter beyond and it is hard going. I am taking the data where I can and for 4 element enlargers I am reduced mostly to hearsay and deductive reasoning. There is an awful lot of potentially good glass going cheap out there for lack of data of any kind. I will take what I can get.

[Tim M]

Gents,

Maybe this belongs in a new thread (vice this "4x shootout"), but as you've already touched on a couple of aspects of my real question, which might be restated as:

"for 1:1 to 2.5x max (not 3x, not 4x - already have those nicely covered with a CF N 4x objective), AND you only had say $200 or less (not enough for most of glass mentioned most often, and also likely not enough for more than one sample), what lens would be most recommended?"

I keep reading here that at these mags "there's lots of good choices". But then there's the possibility of duds, and other remarks that seem to point right back to spending 4x that $200 budget. Any advice?

Thanks,
Tim

[Peter De Smidt]

Maybe a Rodagon D 75 f4.5? You'd have to get fairly lucky on the price.

A Rodenstock (or Linos) Magnagon is a good performer. I paid less than $100 for a new one. It's optimized for 1:5 to 2:1, if I remember rightly. These lenses were supposedly used in some of the Imacon scanners.

For my film scanner I tested a Magnagon on a D600 versus a Rodagon D 75 F4 at 1:1. The Rodagon had a slight edge, but both were very good.

[ray_parkhurst]

Deleted OT post

[ray_parkhurst]

Deleted OT post

[ChrisR]

Tests are in short supply, but cheap and "seem to be pretty good" would be stacked lenses, (50 + 50, 105 + 85... trial and error), or the better enlarger lenses (reversed) which are around 80mm f/4 (Not the Nikkor 75, it's a 4 element) from Rodagon, Schneider; the Olympus 80mm f/4.
Micro Nikkor 55mm f/2.8 is cheap, you can reverse it too but it's a bit bulky for the WD.

[Blame]

I will vote for stacked lenses on the grounds that no matter what the intentions shorter focal enlarger lenses are going to be seriously sharper in the limited field of view we actually need at infinity than in the 1:1 to 4:1 range.

In that capacity I would very much include the el-nikkor 75/4 if you have one lying around.

This mind is mostly theory but what info there is suggests that if you keep the field of view reasonably narrow and the aperture high then a 4 element lens is sharper.

Now the problem with stacking lenses at low magnifications is that the front lens is probably going to be fairly wide angle. Used as an infinite lens the rays from a point from the corner of the field of view will exit parallel to one another but angled towards the far corner. The lower the focal length the greater the angle. Worse the central point from where that angling starts is going to be somewhere inside the lens. Not at the back.

So from this we can conclude 3 general but not absolute rules:

1) keep the focal lengths as long as possible bearing in mind that long focal length lenses are often not as sharp as short.

2) Go for physically short lenses.

3) Get them as tight to each other as possible.

From 1 and 2 its tempting to go for 4 element enlarger lenses on the grounds that they are likely to be sharp at longer focal lengths for the fairly modest field of view of a digital camera (as opposed to the medium to large format photography they were designed for).

I certainly intend to try jamming a 75mm and 90mm cheap enlarger lenses together sometime with a black paper stop in between to give 1.2x at f/11 and about 0.8x at roughly f/9. It might work or it might not but the cost and effort wont be high.

[ChrisR]

Why do you favour 4 element lenses?
I've tried (only one each of) Nikon 50mm f/4 and 75mm f/4. Definitely not as good as their 6 element brothers when I used them, and I know of no reason to believe the situation would reverse in another application - but it might.
I'd advise against actually parting with money for them - other than for academic interest.

[Blame]

Why do you favour 4 element lenses?
I've tried (only one each of) Nikon 50mm f/4 and 75mm f/4. Definitely not as good as their 6 element brothers when I used them, and I know of no reason to believe the situation would reverse in another application - but it might.
I'd advise against actually parting with money for them - other than for academic interest.

Well yes. It would be throwing good money on a longshot. However there are a lot of them lying around including a fair number in my "not worth selling on Ebay" box. It was more advice to try them if you have them.

However there IS a realistic chance they will do well and I did give reasons if only based on hearsay. I would point out that at f/8 and as they approaching infinity, if we are to trust the data at http://coinimaging.com , the 75/4 is very low in CA and the corners hold up. It is also close to diffraction limited. F/8 may seem high but a pair of them stacked face to face will be delivering that f/8 as an effective aperture.

They are small enough and the focal length is long enough that one could hope that the light path is not too much off the designed for, especially with a stop placed between them. Finally they gain all the advantages of a symmetrical lens stacked so many of the distortions automatically disappear. To a lesser extent also true for stacking with a slightly different focal length like 90mm.

If you still have those two lenses try them together. You never know.