Schneider vs Rodenstock - Is it relevant ?greenspun.com : LUSENET : Large format photography : One Thread
I know this may sound like a silly question but I really would appreciate some opinions based on my arguments below.
I am just starting out in large format photography and am looking to get three lenses. The sharpness and overall image quality is of paramount importance to me, especially since I will shoot 6x9 almost exclusively, using a roll film back.
In particular, I am comparing Rodenstock and Schneider lenses and it seems that from the data of the lenses, the Rodenstock lenses are superior then their Schneider counterparts.
For example the MTF curves of the Rodenstock Apo-Grandagon 45/4.5 are vastly superior to the Schneider Super Angulon XL 47/5.6. In fact the MTF curve for this particular lens is so impressive that it pretty much matches the legendary Zeiss Biogon 38/4.5 for the Hasselblad SWC! And beats most other Zeiss wide angle lenses! Truly remarkable, if true!
Similarly the Grandagon-N 75/4.5 beats the Schneider Super Angulon XL 72/5.6. Even standard lenses seem to go Rodenstock’s way as far as the MTF curves are concerned. Each Apo-Sironar-S beats its Schneider Apo Symmar counterpart and the cheaper Sironar-N about matches it.
I know a lot of people will say that they are all the same – performance wise that is, even the G-Clarons! But then most people stop down to f22 or more. At such small apertures diffraction takes over and yes there indeed is very little or no difference at that aperture! In fact at f45 even a hundred year old lens will be as sharp as a modern one!
That really is not a fair way to judge a lens. One must stop down to the optimum aperture of each lens and not f22.
At f22 diffraction limits the resolution to around 60 lines per millimeter. I think f16 is fair since the diffraction limit is then 90 lines per millimeter. But again one must stop to the optimum aperture for that particular lens.
So what is the bottom line here? Are the Rodenstock lenses really that superior to the Schneider equivalents as the data suggests? Or are the MTF curves by Rodenstock rigged?
So which is better? Schneider or Rodenstock?
Thanks for reading and I look forward to some interesting posts!
-- Mike Foster (email@example.com), November 26, 2000
I'm curious. Where are you finding the MTF curves for Rodenstock? Is this from the Web, or do you have product literature? If on the Web, at what address can they be found?
-- 31415926 (firstname.lastname@example.org), November 26, 2000.
I'm curious. Where are you finding the MTF curves for Rodenstock? Is this from the web, or do you have product literature? If on the web, at what address can they be found?
-- neil poulsen (email@example.com), November 26, 2000.
Mike at the end of the day I feel there is very little to choose with regards to the sharpness between makes. I've long abandoned MTF curves etc. for I believe the majority of todays computer designed lenses are going to be first class for use in real world photography. If the final image is first rate I'm happy, remember the chain is only as strong as its weakest link, this of course could be anything from the film holder to the enlarger/printer.
I use both Rodenstock and Schneider lenses for both 6x9cm and 4x5. I choose a particular type/make which fits the bill for my photography, after choosing the focal length I look at its image circle (camera movements), weight and size (backpacking).
The only lens I was ever disapointed with was Schnider's 65 Super Angulon, it had poor edge sharpness (and I was only using it on 6x9 at the time). I exchanged the first one I purchased only to find the second one was the same. I sold it eventually. But to balance things I use their 47XL and 110XL and find them to be first class with all the sharpness you could ask for. Regard,
-- Trevor Crone (firstname.lastname@example.org), November 26, 2000.
Mike... some comments. First, remember that both companies MTF curves are theoretical not measured so the don't account for actual execution of the lens design. Manufacturing variations in polishing, centering and R.I. of glass will effect real MTF.
That said, the Apo-Grandagons really are spectacular. As impressive as the MTFs are, the proof is in shooting. I think the 55 is the sharpest LF lens I have used on 6x9. Also, on the Hasselblad web site, under Arc-Body lenses, you can see Apo-Grandagon and Grandagon- N 75 f/4.5 curves are 40 lp/mm which is important for roll film work. These lenses are quite spectacular at that spatial frequency as well.
Note that when looking at the MTF curves, the Apo-Symmars are tested at 1:infinity, or 1:10 while the Apo-Sironar curves are at 1:10 or 1:20 if I recall. Generally, the Rodenstock's seem to have slightly better on-axis performance while the Schneider curves are a bit flatter at ~3/4 of the image circle. My guess is that these differences are not significant in real lenses. The most difficult aspect in comparing for use on roll films is that Schneider doesn't publish curves for f/11 or f/16 and they don't publish chromatic aberation curves.
I am looking forward to comparing the new Super Symmar 80mm XL to the Grandagon-N 75mm f/4.5. On paper, it looks like the Grandagon-N is a better lens for roll-film, but given the on-film performance of the 110XL, I am curious to see if the paper results translate into reality.
-- Glenn Kroeger (email@example.com), November 26, 2000.
Mike: This has nothing to do with MTF curves, but I have see a lot of lens seperation in Rodenstock lenses that are a few years old, including a 135mm camera lens and two enlarging lenses I own personally. I called a well respected LF technician to see about having the 135 fixed and was told it wasn't worth it. He also said that Rodenstock had a lot of problems with seperation. I hope the problem has been fixed on new lenses, but it is something to look into.
-- Doug Paramore (firstname.lastname@example.org), November 26, 2000.
You are going to drive yourself crazy sitting around looking at data sheets. On them one lens may look superior to another while in the field the whole question may be moot. Measure your results by looking at (and measuring, testing & whatever if you want) images taken with the lenses. Theoretical performance and how the images look after photographing into the sun through suspended ice crystals at sunrise may not be the same. Both Rodenstock and Schneider make good lenses. If you go to a good store and actually test shoot with the lens you think of buying you can do a direct comparison and buy the one that puts the image on the chrome the way you like, getting the best lens possible. So much can happen to even the best theoretical glass. Fail to clean it properly & you might as well be shooting with an e-bay special mistakenly sold as a cookie cutter. Look in the camera bags of many top LF photographers and you will most likely find a variety of lenses represented. Buy one and its theoretical performance is just that... theory, and you replace it with what works, whether another of the same brand or another make. We use what works and the MTF curves and whatever are only helps to making the decision. Size, weight, filter thread size, maximum apertures for focusing, a 'user friendly' feel, or nostalgia are all valid reasons to purchase a lens. Some are better than others and it will always be that way. Try some and buy what you think will work and be willing to sell or trade those that don't pan out for you, with your particular needs and style of shooting. No manufacturer is perfect, nor are we photographers. Theoretical performance and real world use are often far apart, for many reasons. Choose what you find to be the best, which often ends up being a compromise for many varied reasons.
Just so the photos look as good as you want them to.
-- Dan Smith (email@example.com), November 26, 2000.
As others have mentioned, both lens brands seem excellent in actual use. With that being said, it is also only fair to compare lenses with similar design parameters. As I remember, the Rodenstock 45mm lens doesn't cover 4x5 while the 47XL does (131mm vs. 166mm image circles). In a similar fashion, the Rodenstock 75mm has a 195mm IC while the Schneider 72XL has 226mm IC. If you need the coverage, no amount of MTF theoretical analysis can get you there.
-- Larry Huppert (Larry.Huppert@mail.com), November 26, 2000.
Just wanted to add my few pennies (cents!!) worth !! As has been already stated there is little or no difference between LF lenses these days, quality control and computer aided design have all but eradicated dodgy lenses. Everyone has their favourite lens and the majority will often purchase a lens after readin/hearing such recommendations. I once heard that the 100mm Apo Symmar (Schneider) was one of the sharpest lenses available for roll film. I got one and it was!! But only when stopped down to f22 was it okay for landscape (depth of field problems otherwise). You mention that MTF curves suggest that Rodenstock outperform Schneider, they may translate thus, but I no longer pay attention to these graphs. For LF image circle is of paramount performance and most LF lenses NEED to be stopped down to at least f22 (for landscape work) which makes the data included in the graphs pretty useless. Modern LF lenses outperform the film emulsions available anyhow and I doubt whether any one is able to tell which make of lens made a particular picture. Saying that I do own the 110XL and do believe the hype about its quality as I've seen it with my own eyes. In practical terms there are many more factors in LF photography that can spoil a good a good image...film not seated in holder, dust in holder, dodgy tripod, forgeting to remove darkslide, etc, etc,.... puts lens manufacturer into perspective I think!! Regards Paul
-- paul owen (firstname.lastname@example.org), November 26, 2000.
"I am looking forward to comparing the new Super Symmar 80mm XL to the Grandagon-N 75mm f/4.5. On paper, it looks like the Grandagon-N is a better lens for roll-film, but given the on-film performance of the 110XL, I am curious to see if the paper results translate into reality. "
Actually the best ones from Rodenstock for roll film are the Apo Sironar Digital series which are available from 35 up and most cover 6x12 so that would take care of all 45 roll formats except for 45 roll.
They are the most highly corrected lenses for small format in the Rodenstock line.
We can send spec sheets to anyone in the US who wants details.
-- Bob Salomon (email@example.com), November 26, 2000.
My judgement is that the MTF curves for the Rodenstock Grandagon-N 75 mm f4.5 and the Schneider Super-Angulon XL 72 mm f5.6 are essentially tied. One important thing to realize is that the x-axis scales on the MTF curves published by the two makers are different: the x-axis of the Rodenstock curve (in their paper brochures) has units "image radius [mm]", while the Schneider graph has units of percent maximum image height. The Rodenstock graph goes from 0 to 101.2 mm, while the Schneider graph goes from 0 to 100%, with 100% being 113 mm. When comparing curves you have to be sure to compare the MTF values at the same radii in mm. For instance, 50 mm on the Rodenstock graph is 50 mm / 113 mm = 44% on the Schneider graph.
I read off the values for 10 lp/mm and f22 and the curves are very similar. For most values the Rodenstock MTF is slightly above the Schneider MTF, but there are radii where the Schneider MTF is higher. My guess is that the difference is not significant. The curves were calculated by different people, probably using different programs and perhaps slightly different approaches. It is likely that honest differences in the calculation could vary the results comparable to the differences between the curve. Additionally, there is the issue of how closely each manufactured lens matches the calculated MTF curves. (The Schneider graphs are described as calculated; I don't see any statement on this issue in the Rodenstock brochure.)
In short, Schneider and Rodenstock both make superb lenses. We are lucky to have them. Many of us tend to get obessive about lens quality.
-- Michael Briggs (firstname.lastname@example.org), November 26, 2000.
There are also other important curves to look at.
Rodenstock always also publishes distortion, fall off and color curves as well as MTF.
As Rodenstock is ISO 9001 certified the curves published are what Rodenstock expects the customer to receive as the minimum quality level.
-- Bob Salomon (email@example.com), November 26, 2000.
Thank you to everyone who has replied. The posts have been very interesting.
The MTF curves I am referring to are the ones printed in the Rodenstock literature. The Schneider ones are from their web site. Sorry I forgot to mention this.
Trevor says that all modern, computer designed lenses are first class. I agree, but still a few are even better. They are the masterpieces so to speak. Take the new Zeiss lenses for the Contax 645. Zeiss designed 7 lenses for it. All pretty much at the same time. All are first class. But one of them, the 120 Macro is a masterpiece. Why can’t the others be as good as that? Especially since they are all “computer designed”. I don’t know.
So therefore I am trying to pick the best of the crop. And I know that not all the modern computer designed lenses are equivalent. Some ARE for sure better than others. And I want to find those! Am I being greedy? Sure.
Now coming back to the lenses, sadly Schneider only publishes MTF curves for the wide open aperture and f22 with the exception of the Super Symmar XL series. This is a pity. Why don’t they publish the MTF curves for f11 or f16? One possibility is that they are hiding them deliberately, because they don’t perform that well after all. But that is just a possibility. I know that people rave about them so they must be something to them. Moreover, I found a web page that says the Super Symmar XL 110/5.6 can resolve 80 lines per mm at f11. Now that is impressive.
For the Super Symmar XL series Schneider does publish MTF curves for an aperture other than f5.6 or f22. That aperture is f8. And sadly the curves are not impressive at all at f8 for any of them. I compared the new 80/4.5 against the Grandagon-N 75/4.5. The curves were at f8 for the Schneider and at f11 for the Rodenstock. And again the Rodenstock beat the hell out of the Schneider. But the Schneider uses aspherical elements and so should perform very well at f8. Why doesn’t it? This is a mystery.
The debate continues…
-- Mike Foster (firstname.lastname@example.org), November 27, 2000.
There is no mystery why some lMTF curves are at 22 while others are at 8, 16 or 11.
Manufacturers are showing the performaance curves at the diffraction limited stop. That is where you will get the best performance.
-- Bob Salomon (email@example.com), November 27, 2000.
When I became wrapped up in all the technical stuff, my work suffered. I had to decide between being a photographer or a technican. It's all about knowing your equipment, it's limitations and how to use it to the fullest. I've seen beautiful work with what many here would consider to be so so or or not so good equipment. As for me, I'm glad I've passed that hurdle.
-- Jim (firstname.lastname@example.org), November 28, 2000.
Like Jim, I constantly try to be a photographer rather than technician. But I can empathize with Mike because I have agonized over some of the same issues. The problem today, unlike a decade ago, is that with the best lenses, film and roll-backs, combined with digital scanning and printing, MF can essentially equal LF for prints up to 20x24 (in color! B&W still really benefits from larger film for reasons I don't fully understand). This of course assumes that the MF has full movements. The problem with MF is that you are always on the edge. Everything has to be right. Film flatness, not too small an aperture, fine grained film, not too much cropping and so on. With LF, there is more room for slop. Lens sharpness is not as important given enlargment factors of 5x rather than 9x, and more tolerance in film flatness and position. The tradeoff for MF "equipment angst" is less equipment bulk, more emulsion choice, lower cost and sometimes quicker shooting.
-- Glenn Kroeger (email@example.com), November 28, 2000.
I too agree with everything Jim has said, BUT I am eternally grateful to the technicians among us who push for sharper lenses, better films, more functional cameras, etc. I salute you! Who among us wants to go back to the "good" old days of single element landscape lenses and glass plated film? Not me!
-- Pat Raymore (firstname.lastname@example.org), November 28, 2000.
One thing to be careful about is that Rodenstock often uses f16 for the published MTF's while Schneider uses f22, and it's known that an f5.6 lens is going to give the maximum resolution on film around f11- 16. But how often does one shoot at f11-16? For me, it's more like f22-32. The bottom line is that if you need the DOF, you'll have to stop down the lens. It could be that Rodenstock's MTF's suck at f22- 32...who knows? For an LF lens, f22 is just more realistic, IMHO.
-- James Chow (email@example.com), November 29, 2000.
"One must stop down to the optimum aperture of each lens and not f22."
Unless you are shooting planar subjects, or are not trying to render the whole image sharp, what you loose by not having enough DOF renders the issue of lens quality moot. There is a reason why apertures of f22 or smaller are often used. As a result, I believe the difference you're reading about is not significant in practice.
-- Q.-Tuan Luong (firstname.lastname@example.org), November 29, 2000.
James Chow, Q-Tuan Luong: Remeber, Mike wants to shoot primarily roll film. That gives him between 1.2 and 1.5 stops more depth of field than 4x5. I find on roll film that f/11-f/16 are very usable stops for depth of field when the camera has movements, unlike 4x5 where f/22-f/32 are more typically needed. In fact, that is one of the strong points of roll film, since that extra stop+ can be used for higher shutter speed, or slower film.
-- Glenn Kroeger (email@example.com), November 29, 2000.
"One thing to be careful about is that Rodenstock often uses f16 for the published MTF's while Schneider uses f22,"
The Apo Grandagons are diffraction limited at 11 or 16 depending on the focal length. All others are diffraction limited at 22 and that is what the MTF curves are.
Note: the digital lenses are made to work at large apertures rather then small apertures and also have MTF at apertures larger then 22. But then these lenses are for small formats or digital and don't cover 45.
-- Bob Salomon (firstname.lastname@example.org), November 29, 2000.
Good God! I wish I could get Bruce Barnbaum to look at this thread. He'd be laughing his head off..... I think he'd put it in the same file where he puts his information on densitometry. ;-)
-- matthew cordery (email@example.com), November 30, 2000.
Dear James and Q, as Glenn has said I do want to shoot mainly roll film. So I won’t want to stop down that much.
But your argument begs one very important question. Why on Earth do you shoot 4x5?
Since you need to stop down to f22 or more almost always, this SEVERELY limits the resolution you will get. So why not shoot 6x9 on roll film instead? You will get the same movements and other advantages of view cameras. But, you won’t have to worry about those film holders! You can shoot more film and much easily. No more loading film in darkness, etc. The list is endless.
More importantly as Glenn mentioned, you would be able to get your much needed depth of field a good stop or more lower. And according to Q’s link since 50% of the shots are taken at f32, you could now use f22 instead. This implies that the diffraction limited resolution will jump from 46 lp/mm (at f32) to 68 lp/mm (at f22), which compensates exactly for the smaller negative and so therefore you will get the SAME resolution!
Of course you might argue about grain and tonality. But with modern films that is a moot point. And unless you blow up the images to 30x40 or more you won’t notice it.
So why shoot 4x5?
It also amazes me why people shoot 8x10 at amazingly small apertures like f128. Naturally they feel they need to stop down that much to get their desired depth of field. Fine. But then why 8x10? What’s the point of that extra chunk of film?
The 3D world has its limits when trying to photograph and getting the most resolution in the third dimension. This is a big limitation. In the small world we use electron microscopes to get around this. Maybe we need to make an electron view camera :)
In the meantime this is what I would do. For shots that absolutely beg for great depth of field I would take out a roll film holder and shoot in medium format. If the object is pretty 2 dimensional, say like a building then heck you can shoot 11x14 and get resolution beyond belief. Same for shots of infinity.
Sorry but shooting at f32 and smaller apertures just doesn’t make sense to me...
-- Mike Foster (firstname.lastname@example.org), November 30, 2000.
diffraction limits are just theoretical limits. there are many other factors which degrade resolution. in practice, it's much easier to get 30 lpm on film and enlarge two times less than to get 60 lpm or higher.
grain and tonality are not a moot point. in fact, many feel that's where larger film makes a difference. resolution grows linearly, grain and tonality as a square function of film linear dimensions. in my experience, with the best modern films, grain shows up at a 10x enlargment.
individual sheets of film are actually nice to have in order to optimize processing. it's also nicer to look at a 5x7 transparency or contact print than at a 2x3.
your assessment that if the subject is very 3D, then it's not possible to get an extremely sharp image is correct. I do resort to 35mm myself in these cases (for instance for macro). but then, if you have a LF camera, you have at least the option of using the larger or the smaller formats depending on the subjects.
-- Q.-Tuan Luong (email@example.com), November 30, 2000.
Mike: Q makes some good points... as much as I am a fan of roll-film, and appreciate its advantages, there are still some advantages with 4x5. In color, with digital reproduction (Tango scan/LightJet print) there is a subtle difference starting at 20x24 inch prints. It is mainly a grain/tonality difference, and is truly minor. If I never wanted larger prints, or never needed to crop, then I could toss 4x5.
But... as I said before... with roll-film, everything needs to be perfect to get that quality. Sheet film gives me lots of flexibility. I can use faster film, or push an extra stop and not worry about grain. I can crop away 50% of a chrome and still have more left than roll-film.
Diffraction limits aren't the real problem. If you evaluate the situation from an MTF standpoint, the real problem is film. In color, Velvia and Provia loose contrast rapidly past 40 lp/mm. You can recover some of that with digital sharpening, but never restore the contrast you get on film at 20 lp/mm. If we could get films with the grain structure of Provia F, with a decent MTF out to 100 lp/mm, then we could worry more about the lens, but I doubt that we will get color film like that. Much more likely to get digital sensors that perform well at those spatial frequencies. That is why Rodenstock and Schneider are developing their "digital" lenses. Some of those have image circles large enough to consider using on roll-film.
Bottom line for me is that roll-film with the right lenses, film and printing is more than adequate for most of what I want, and I can appreciate a smaller, lighter, faster set of gear. But 4x5 still has significant advantages as well.
-- Glenn Kroeger (firstname.lastname@example.org), November 30, 2000.
I shoot roll film backs and 4x5 on the same cameras (Arca Swiss F-line, Canham DLC and previously, on a Sinar C). so I'll try to answer some of your questions from your second post.
"But your argument begs one very important question. Why on Earth do you shoot 4x5?"
When I shoot 4x5 instead of roll film it is for several reasons. The primary one is the visual information gets spread out over more real estate. The next one is that I don't have to shoot an entire roll to switch emulsion types. The third reason is that I can shoot six sheets and process first one and then if necessary another to determine how to best process the rest.
"Since you need to stop down to f22 or more almost always, this SEVERELY limits the resolution you will get."
I very rarely shoot below f/22. I learned on a Sinar P studio camera and learned how to use the movements of the camera so I wouldn't have to stop down to stops smaller than that unless it was absolutely necessary (and sometimes it is absolutely necessary, even with roll film.
" But, you won?t have to worry about those film holders! You can shoot more film and much easily. No more loading film in darkness, etc."
And neither do I. I use Fuji Quickloads almost exclusively, and I am just as fast as when I use the roll film backs (Sinar 6x7, Horseman 6x7 and Horseman 6x12)
"More importantly as Glenn mentioned, you would be able to get your much needed depth of field a good stop or more lower. And according to Q?s link since 50% of the shots are taken at f32, you could now use f22 instead."
See my earlier paragraph addressing this issue.
" This implies that the diffraction limited resolution will jump from 46 lp/mm (at f32) to 68 lp/mm (at f22), which compensates exactly for the smaller negative and so therefore you will get the SAME resolution!"
...but resolution is not the whole story!
"Of course you might argue about grain and tonality. But with modern films that is a moot point. " I respectfully beg to differ as you are forgetting that when you enlarge to any size you also have to deal with the qualities of the enlarging lens and enlarging systems.
"It also amazes me why people shoot 8x10 at amazingly small apertures like f128. Naturally they feel they need to stop down that much to get their desired depth of field. Fine. But then why 8x10? What?s the point of that extra chunk of film? "
Do you actuually know many people who shoot 8x10?
-- Ellis Vener (email@example.com), November 30, 2000.