PDA

View Full Version : Looking for a lens with over-corrected spherical abberation



Patrik Roseen
13-Mar-2008, 07:34
Hello,
I think I might have understood what makes some lenses swirl (like the petzval lenses).
It's probably a combination of over-corrected spherical abberation that should show up as ringed highlights behind the plane of focus and optical vignetting that is quite easily accomplished with a limiting hole in the back of the lens.

most of the lenses I have tested so far seem to have the ringed highlights in front of the plane of focus and that's not what I am after.

Are there any suggestions for lens designs which creates the ringed highlights behind the plane of focus?

lenser
13-Mar-2008, 07:55
I'm not into the science of optics, but I remember from my days as a portrait photographer a trick we used on a filter to break up background highlights. Taking a UV or other plain filter, we cut a lightning shaped piece of electricians tape and laid it over the center area of the filter from side to side. The results were that all highlights in the background were split with a soft but defined jagged shape in the very out of focus center.

I'm wondering if you might not achieve something like this with tiny slices of tape arranged in a concentric pattern around the edge of a filter and maybe in a bout a fourth of the way. Make them random and spread well apart and about the size of a small fingernail clipping and, in the right out of focus situation....maybe, just maybe this will work.

Good luck.

Tim

Ken Lee
13-Mar-2008, 09:25
If you turned the lens around, would that do it ?

Patrik Roseen
13-Mar-2008, 10:00
If you turned the lens around, would that do it ?

Ken, I have tried this but the physics are against me. I have tried a couple of lenses, each with two lens elements, and turned them in every way I could think of and it still does not give me what I am looking for (except for many other strange effects)

Tim, I will try your suggestion but I am not looking at taking light out of the higlights, on the contrary I wish to make them glow and preserve their light.

googling BAD BOKEH seems to indicate a 'Double Gauss' as a lens design to look for.

lenser
13-Mar-2008, 10:46
Patrik.

My idea won't remove any highlights, but will modify them to be less promenent and may create the swirly pattern that you are after. It's just a total guess, but it just might work into something that will be near to your goal.

Tim

wfwhitaker
13-Mar-2008, 10:50
A lens I had which showed indication of over-corrected spherical aberration was the 14" Kern Dagor, version 3. But I don't think that's what you're looking for...

Jim Galli
13-Mar-2008, 12:03
There may only be a dozen people who know if their lens they want to sell is "over corrected for spherical aberation" (I'm not one of the 12) so you're sort of limiting your chances of finding what you want.

A good lens to play with is the triplet Bausch and Lomb projector lens from the 1920's. They're not worth much, and they're simplicity itself. All 3 elements fall out together. You can flip things around backwards and change spacing to your hearts content with very little effort. I listed a 12" a while back with no takers at 40 bucks.

Patrik Roseen
13-Mar-2008, 14:31
There may only be a dozen people who know if their lens they want to sell is "over corrected for spherical aberation" (I'm not one of the 12) so you're sort of limiting your chances of finding what you want.

A good lens to play with is the triplet Bausch and Lomb projector lens from the 1920's. They're not worth much, and they're simplicity itself. All 3 elements fall out together. You can flip things around backwards and change spacing to your hearts content with very little effort. I listed a 12" a while back with no takers at 40 bucks.

Jim, I was mainly asking for advice in which lens (vendor/lens design) to look for. The double gauss design I mentioned is apparently very complex.Seems most cheap/normal lenses include positive elements whereas the over-corrected lenses include negative elements that will push the spherical abberation to the other side of the plane of focus. That is probably the reason for me not getting anywhere by turning the glass around., they are still positive.

Jim, Does the B&L lens you mention include any negative elements? If you still have the 12" for sale you may send me a PM.

As for knowing if the lens is overcorrected or not I agree that most people do not know (I was not even aware there were such things until Ole Tjugen informed me less than 24 hours ago.

Let me repeat myself in the appreciation for Domenico Foschi's photograph of the boy on the bridge which clearly shows the ringed highlights in the swirling background. That is what I am looking for.
http://www.dfoschisite.com/html/~stock_I.htm view the second photo in the 'Stock I' portfolio (BEAUTIFUL!)

Jim Galli
13-Mar-2008, 15:04
It is 3 elements 3 groups. A classic Cooke (http://dioptrique.info/objectifs5/00200/00200.HTM) triplet. The middle element is a double negative. Front and back are positive.

Actually the double gauss design in one of it's forms is borrowed for the dialyt lens (http://dioptrique.info/objectifs4/00157/00157.HTM) and they are common as dirt. Contact me at jimgalli at lnett dot com. I still have the projector lens and also an interesting old brass dialyt. There are many dialyt's. Goerz Celor, Syntor, Dogmar, Artar, to name just a few. Cooke Aviar is one of the finer dialyt's. Many of the cheap lenses of the 1920's were f6.8 dialyt types. The better ones for your purpose would be the more expensive f4.5 versions. More aperture, less corrections. The inherent problem with this design was flare. 8 count em 8 air - glass surfaces that each lose about 4% of the image forming light. They were sharper than anything else but the flare killed them until after WWII when coated surfaces made dialyt's cake and eat it too.

Jim Galli
13-Mar-2008, 15:07
I went and looked at Domenico's picture. Classic Petzval. That's all you need to know. ;)

Ole Tjugen
13-Mar-2008, 15:25
Overcorrection is more common in some "modern" lenses which are optimised for apparent sharpness rather than real resolution...

There is also a trend in lenses for 35mm cameras that telephoto lenses are constructed to give soft backgrounds whereas wideangles usually give soft foregrounds. It's logical when you think about it...

The peoblem is that the correction isn't inherent in any basic lens design. Rather it's a choise that the makers do for each and every lens model! Even with the Heliar, which is intentionally undercorrected, there is nothing preventing a lens made on the same formula from being overcorrected (and some repro versions may be overcorrected).

A fast triplet might be a good place to start. These are very sensitive to small changes in the placement of the central element, which could make them easier to modify.

Mark Sawyer
13-Mar-2008, 15:30
The Verito and Imagon are both capable of the swirlies too. I'm not sure I attribute it to the spherical aberration though...

Peter K
13-Mar-2008, 15:46
As the Airy-disk or the circle of confusion depends strongly of the form of the diaphragm, special soft-focus lenses are equiped with special rounded diaphragms with many blades to get a realy round circle of confusion. So you will get this highlights as Domenico's image of the boy on the bridge.

Special soft-focus lenses are aviable but not for LF. E.g. the AF DC-NIKKOR 105 mm 1:2D. With this lens one can choose the defocussing of the back- or foreground. But it will be hard to find such a lens for LF photography.

Paul Fitzgerald
13-Mar-2008, 18:21
Patrik,

You could try a Universal Heliar, the moving center element is a double negative, but it will not go as far as an undersized petzval will.

Just a thought

BradS
13-Mar-2008, 18:54
"over-corrected spherical abberation"

huh?

cowanw
13-Mar-2008, 19:09
Let me repeat myself in the appreciation for Domenico Foschi's photograph of the boy on the bridge which clearly shows the ringed highlights in the swirling background. That is what I am looking for.
When I saw this photo I was reminded of my mirror telephoto with it's donut highlights, when used with grass or leaves inthe background. It looks a lot like this when there are no specular highlights to give the real doughnut shape.
Regards
Bill

Struan Gray
14-Mar-2008, 01:42
As far as I can see, there is only an indirect connection between spherical aberration and the swirlies. If you have a ring-shaped out-of-focus spot because of spherical aberration it will enhance the swirlies and make them more prominent, but other factors have to come into play before it will work well.

The point is that spherical aberration is symmetrical. If no other effects are present you blur the ideally-focussed point spot symmetrically. You can make various patterns like rings, fuzzy spots and dots with your out-of-focus highlights, but you can't make them swirl around the centre of the field.

To do that, you need an *asymmetric* blur. That can either be an asymmetric aberration like coma or astigmatism, or it can be caused by the way the aperture seems to narrow into a "cat's eye" shape as you go off axis.

Traditionally, asymmetric blur has been regarded as ugly. That's why all the famous early lenses trumpeted their lack of astigmatism. So if you want that blur from a lens designed for photography you need to operate outside its comfort zone. Usually that means using a larger aperture than the manufacturer recommends or using the lens on a larger format than it was intended for.

The latter trick is what makes Petzvals so super-swirly cool. It won't work with all lenses though, because some designs lose all definition as you move outside their specified image circle, and because many lenses - especially modern ones - deliberately vignette their circle of illumination so that it is restricted to the boundaries of the circle of good definition. This is the main reason you need to use older lenses: they tend not to have such aggressive 'field stops' and so they will illuminate a much larger piece of ground glass or film than they are intedended to cover photographically.

Modern Dialytes like the many Apo-process lenses (Artars, Ronars, Lustrars etc) do have mild bright line bokeh behind the plane of focus, but they don't work well to produce swirlies because the definition dies fast outside the image circle and they tend to be deliberatly vignetted to reduce stray light. Older Dialytes like the Goerz Celor and related lenses like the Ross Homocentric are a much better bet, but you have to find one and test it for yourself (and cope with the bad flare from eight uncoated air-glass surfaces).

Triplets are traditional soft-focus lenses, but again they have a restricted field of view. The famous portrait objectives were designed to produce an overall *pleasing* softening of the out-of-focus parts, and deliberately avoided asymmetric swirls. Triplet projector lenses from slide projectors and AV equipment are very cheap right now, so it might be worth playing with, say, a 150 mm lens from a 6x6 projector. You will probably have to cut the lens barrel down to prevent it from vignetting the swirly part of the field. My experiements with free lenses culled from 35 mm projectors suggests that it won't work too well, but it cost nothing to try.

Double Gauss lenses have a very different layout from Dialytes. They do show swirl at the outer edges of the field, but that swirl is very dependent on the aperture. It comes from a mixture of aperture narrowing and oblique spherical aberration (spherical aberration which varies as you move off the axis), and the oblique spherical aberration is quenched even more rapidly as you stop down than first-order aberrations like coma and astigmatism. Since most double Gauss lenses in modern times have been relatively slow wide angles, the swirl is quite well controlled from the outset. You could find a 89 mm Wray wide angle (cheap) or one of the Kodak Wide Field Ektars (more pricey) and experiment on 4x5, but anectodal evidence suggests they will, just like the Dialytes, vignette before they get interesting.

(Patrik: I may be mis-remembering, but don't you have one or more Cooke wide angles? They too are double Gauss designs. Try them wide open and let us know how they perform at the limits of their image circles :-)

Which is why everyone comes back to Petzvals. They have a restricted field, but they illuminate a much larger image circle and the image quality degrades gracefully into that outer zone. If you want to reproduce the look of Domenico's photo the simplest way it to use a similar lens, and not to spend too much time worrying about the physics of how it achieves its funky look. If on the other hand you want to experiment with a wide variety of different sorts of funk, start bidding for the cheapie brass lenses on eBay.de and use them on the wrong format.

Patrik Roseen
14-Mar-2008, 15:52
Thank you all for your response to this thread, I truly appreciate it!


...
When I saw this photo I was reminded of my mirror telephoto with it's donut highlights, when used with grass or leaves inthe background. It looks a lot like this when there are no specular highlights to give the real doughnut shape.
Regards
Bill
Bill, as you say I have also found information about mirror lenses creating the 'donut highlights'. Seems it is quite common in telescopes and microscopes. It would be interesting to understand that type of optics too.

Struan, when reading your excellent explanations I realize how much I still have to learn.

The question is ofcourse why do I at all try to accomplish this swirling effect when there are petzvals to be hunted for out there? Well, it's probably for the fun of it...going from total ignorance...to understanding ... and hopefully doing. Maybe I have this idea that the Petzval isn't this magic lens we think it is! ;)

Jim Galli, I intend to contact you when I come back from the Stockholm archipelago after the weekend. And I am already looking for brass lenses and projection lenses on ebay.de

Peter K
14-Mar-2008, 16:21
Reflector telescopes are quite common but in the field of microscopy a mirror isn't often used. Isaac Newton himself has suggested a mirror lens because a mirror has no chromatical error.

Mirror lenses for photography are used to get long focal lenghts. The ray path is folded, so the length of a 500mm lens is quite short. But this kind of lenses have a doughnut like entrance pupil, so the circle of confusion of this lenses are also ring shaped. Mirror lenses used for photographical purpose are a combiantion of lenses and mirrors, so the chromatical error is still an issue with this lenses.

Peter K

Dan Fromm
14-Mar-2008, 16:46
Um, Struan, there are fast relatively narrow angle double Gauss types that almost work for LF. Examples include the very expensive (when longer than 100 mm or so) Xenotar, Dallmeyer Super Six; f/1.4, f/1.9, and f/2.8 Saphirs; yes, the Planar; and more, I'm sure. How about them?

Peter, I know they're not for LF and are scarce enough to be more myth than real, but do the Makowski Katoptarons have refracting lenses inside or are do they do it all with mirrors?

Peter K
15-Mar-2008, 01:40
Peter, I know they're not for LF and are scarce enough to be more myth than real, but do the Makowski Katoptarons have refracting lenses inside or are do they do it all with mirrors?
Sorry Dan, I've no information about the Katoptatrons. Possible you can throw some light into this field.

Dan Fromm
15-Mar-2008, 05:17
Peter, somewhere I have some Makowski propaganda that Klaus Schmidt sent me, but of course I'm not sure where it is. Klaus has actually bought several Katoptarons, I've touched one at a camera show years ago and that's the limit of my experience with them.

Peter K
15-Mar-2008, 06:18
The Katoptaron can be used with UV up to IR, so only mirrors can be used. Also the ray path is folded without using an element two times, so the entrance pupil is a full circle and no doughnuts formed circle of confusion are made.

Dr Klaus Schmitt
15-Mar-2008, 10:05
Peter, somewhere I have some Makowski propaganda that Klaus Schmitt sent me, but of course I'm not sure where it is. Klaus has actually bought several Katoptarons, I've touched one at a camera show years ago and that's the limit of my experience with them.

Hi Dan + Peter,

the docu may be found on my site here (you know that Dan...): Makowsky kataptaron TS-E500 (http://www.macrolenses.de/ml_detail_sl.php?ObjektiveNr=292)

and this is the much more advanced scientific version LDM-1 which he had developed from that: LDM-1 long distance microscope (http://www.macrolenses.de/ml_detail_sl.php?ObjektiveNr=315)

No doughnot shaped bokeh, due to the Z-folded light path, no glass, just two front surface quartz coated mirrors; very simple, yet clever design ("Kutter Schiefspiegler").

Funny, just made some tests today for another forum:

28mm overview:
http://www.macrolenses.de/bilder/28mm_DSC3849ack.jpg

LDM-1 (f=800mm, f11):
http://www.macrolenses.de/bilder/LDM-1_DSC4120k.jpg

100% crop:
http://www.macrolenses.de/bilder/LDM-1_100_DSC4120ck.jpg

Castle is about 1000 meters / 3000ft away.

Bokeh/DOF test, ca 30m/100ft away, fully open aperture:
http://www.macrolenses.de/bilder/LDM-1_DSC4198k.jpg

overcast, shot at 1/180s @ ISO400, light wind, far from optimum for a 800mm lens!

Sorry, I did not want to hijack that thread, but since there were questions about that exotic lens...

Cheers,

Dr Klaus Schmitt
15-Mar-2008, 10:14
Struan, gents,

I have a high speed Steinheil Jos-Pe-Cassar 1:3 f=18cm (180mm) here. Since the Cassarit is a classic Cooke triplet design, would that produce "the swirl"??

Cheers,

Struan Gray
15-Mar-2008, 14:21
Um, Struan, there are fast relatively narrow angle double Gauss types that almost work for LF. Examples include the very expensive (when longer than 100 mm or so) Xenotar, Dallmeyer Super Six; f/1.4, f/1.9, and f/2.8 Saphirs; yes, the Planar; and more, I'm sure. How about them?

I'm too mean to have indulged my curiosity :-) The only data point I have is my Hasselblad 80 mm F lens, which vignettes before it gets at all interesting.

Klaus, since you actually have the lens I'm reluctant to guess. The only Cooke triplet I own is too long for me to have been able to explore its outer limits.

I distinguish between lenses like the classic portrait lenses which are designed to produce a 'pleasing' bokeh over the whole field of view, and those which give more dramatic effects which vary strongly across the field. I'm not aware of a lens that does both well. I've not done much swirling myself: I'm more interested in how bokeh creates an interesting but even background texture.

Re: relective microscope objectives. Edmund have these:
http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productID=2687. My collegues who know what they're doing with microscopes say they're only worth using if you really need the wide spectral range. They're good, but if you only want, say, the UV performance, a dedicated UV objective is better.

Ole Tjugen
15-Mar-2008, 14:33
The Xenotar is a Planar-ish lens; not a double-Gauss. I don't know about the others, and I'm not going to look them up right now. ;)

Dan Fromm
15-Mar-2008, 14:50
Ole, what is the Planar if not a 6/4 (originally) and later a 6/5 (and other variants too) double Gauss?

Ole Tjugen
15-Mar-2008, 14:54
Ole, what is the Planar if not a 6/4 (originally) and later a 6/5 (and other variants too) double Gauss?

The Planar was so thoroughly redesigned from the double-Gauss that the comparison is about as useful as saying that a Tessar is a reverse Petzval. ;)

Dan Fromm
15-Mar-2008, 16:52
Thanks. Interesting too, but not everyone follows your convention.