Leica SE 18mm Service

[markowich]

Advertisement (gone after registration)

maybe some of you remember the distorsion problems i had with the SE 18mm, if you don't please take a look at the attached jpg. i then sent the lens to leica, where it sat for more than 2 months. today i received the following email, which-for the fun of it, i quote in full (german) detail:

 

Sehr geehrter Herr Prof. Markowich,

vielen Dank für Ihre Mail und das Bild.

Entschuldigen Sie bitte die Verzögerung in der Beantwortung Ihrer Anfrage.

Der techniker hat Ihre Frage wir folgt beantwortet:

 

bei Weitwinkeln ist eine Verzeichnung und Vignetierung nicht zu verhindern. Es liegt somit an der Art des Objektives.

 

Ich hoffe, Ihre Frage beantwortet zu haben. Gegebenenfalls leite ich weitere Fragen an unseren Technischen Infodienst weiter.

 

translation:

 

dear prof. markowich,

many thanks for your mail and the image. please excuse the delay in responding to your request. the technician has answered your query as follows:

 

Distorsion and vignetting cannot be avoided in wide angle lenses. Therefore the issue is due to the type of lens under consideration.

 

i hope to have answered your query. Otherwise i shall refer further questions to our technical information service.

 

i am speechless. i have meanwhile emailed stefan daniel to ask him to brand the SE 18mm as 'Not suited for architectural photography'.

speechless about the depth of response.

peter

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[UliWer]

The answer they gave you was not wrong. The data sheet for the lens shows how much distortion you have to exspect when you buy this lens.

 

That it took Solms two months for the simple reply is a completely different matter.

[markowich]

The answer they gave you was not wrong. The data sheet for the lens shows how much distortion you have to exspect when you buy this lens.

 

That it took Solms two months for the simple reply is a completely different matter.

 

sorry no. the datasheet does not in the very least reflect what we see on my jpg.

peter

[01af]

Sorry, no. The datasheet does not in the very least reflect what we see on my JPEG.

It does. What I see in your image is exactly what I'd expect after having taken a look at the data sheet.

[markowich]

It does. What I see in your image is exactly what I'd expect after having taken a look at the data sheet.

 

well, i have measured the distorsion values and the are definitely above the specs. a horrible lens, completely useless except maybe for landscape if you do not mind a curvy horizon. the zeiss counterpart is phenomenal in comparism. and half the price.

peter

[UliWer]

I just wanted to propose a comparison to the Zeiss 18mm. If it is completely different I should be glad seeing the same motive taken with it to learn that the Leica data sheet is wrong and the Zeiss - which tells us it has 0.2% less distortion - is right. The advertizing text for the Zeiss tells us, that it is "almost" distortion free.

[becker]

Advertisement (gone after registration)

I like mine very much.

Its wonderful.

Both 18 mm se Elm. Handhold

 

the 15 mm cv shall be good, saw some example in the Germ Forum by an arch Prof

he was satisfied with it.

 

Make a comparation to an other 18 mm I f Yours is a Monday Car.

The distortian is like a smooth wave.

 

Hope that helped You.

 

Kr Matt

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[markowich]

I just wanted to propose a comparison to the Zeiss 18mm. If it is completely different I should be glad seeing the same motive taken with it to learn that the Leica data sheet is wrong and the Zeiss - which tells us it has 0.2% less distortion - is right. The advertizing text for the Zeiss tells us, that it is "almost" distortion free.

 

i'd loveto retake the picture with the zeiss, but it would take a travel to jeddah. i avoid this in summer---)))

peter

[markowich]

I like mine very much.

Its wonderful.

Both 18 mm se Elm. Handhold

 

the 15 mm cv shall be good, saw some example in the Germ Forum by an arch Prof

he was satisfied with it.

 

Make a comparation to an other 18 mm I f Yours is a Monday Car.

The distortian is like a smooth wave.

 

Hope that helped You.

 

Kr Matt

 

looks much better. thanks, p

[becker]

I just wanted to propose a comparison to the Zeiss 18mm. If it is completely different I should be glad seeing the same motive taken with it to learn that the Leica data sheet is wrong and the Zeiss - which tells us it has 0.2% less distortion - is right. The advertizing text for the Zeiss tells us, that it is "almost" distortion free.

 

What ever that means and under which conditons, such sww must be well ballanced,

only a very little inaccuracy, disturbs the hole pic. Only my few reflections.

 

 

Kr

 

Matthias

[UliWer]

i'd loveto retake the picture with the zeiss, but it would take a travel to jeddah. i avoid this in summer---)))

peter

 

Don't take too much pains! As you say that the Leica 3.8/18 is unusable whilst the Zeiss 4/18 is ever so brilliant it won't be difficult to show other examples comparing the two lenses where the difference will be evident.

 

This gives me an occasion for greeting Sean Reid, who has done so much to enable users to look at reliable comparisons. I wish him well, hoping that he soon can go on with his work.

[hoppyman]

I agree that the delay to receive a response would be very disappointing.

In my opinion the example shown by the OP represents the worst possible case to show the distortion which is part of the design compromise with such a wide angle. You can clearly see that the distortion is most evident at the edges and corners and not apparent through the centre. That is exactly what you would expect when you look at the published performance figures from the lens.

 

For many other subjects the distortion is much less apparent and not at all objectionable. In general terms an 18mm is of course, not ideal for architectural shots if you require the most accurate rendering. An 18mm is more suitable for when you take advantage of the more dramatic effect. You can choose to exaggerate the effect by your composition as well.

Here are examples from an M8. The ZM 18 is an excellent lens too but it is certainly not exempt from distortion.

 

Uncropped ZM 18 (coded as Wide Angle Tri Elmar @18)

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Uncropped Super Elmar 18

 

Better use of Super Elmar to take advantage of the more dramatic effect.

[AlanG]

I have to say that is a lot of distortion and also a very demanding subject.

 

It was not shot squared to the subject and that adds to the problem.

 

I "squared up" the image the best I could and then drew some boxes over it so that we can better see what is going on. (An image that was shot perfectly perpendicular to the wall and level would have been ideal.) This is the kind of distortion that will need a special correction profile. (I don't know what aperture was used but perhaps the distortion improves a bit when stopping down.) The distortion is wavy and progresses towards the edges.

 

Then for comparison is a very trying image that was made with a 17mm TSE. This is a very large lens but amazingly has virtually no distortion. It is tough to get perfectly aligned on a shot like this but I got pretty close. At some point you have to wonder if everything in the room is actually perfectly square and straight.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[giordano]

Then for comparison is a very trying image that was made with a 17mm TSE. This is a very large lens but amazingly has virtually no distortion.

 

That's really impressive. But a T-S lens is certain to be used for architectural and technical work and the Canon designers would have given low distortion a very high weighting in the design process.

 

Reasoning backwards from the performance of the 18mm Elmar, we can conclude that the Leica designers didn't intend the lens for critical architectural work. And why should they, given the general superiority of reflex and view cameras for that particular task?

[01af]

The Super-Elmar-M 18 mm Asph primarily is an M lens—that is to say, SHARP and SMALL. You simply can't have a lens that covers a 24 × 36 mm frame with an angle of view of 100° that is tack-sharp across the whole field and small and distortion-free. The Canon TS-E 17 mm is sharp and virtually distortion-free—but definitely not small.

 

Generally, Leica accepts fairly high amounts of distortion in their wide-angle lenses (and often the unpleasant and hard-to-remove wave-form kind) in order to optimize other aspects. Zeiss wide-angle lenses usually have less distortion but they are bigger, and/or have more curvature-of-field, and/or vignette more, and/or are not so sharp at the frame's corners. The art of lens design is to find a good compromise. And each lens desiger has different ideas, preferences, and priorities.

[AlanG]

Smaller with distortion vs. larger without. Lens designs are compromises. I must say that I am astounded by how good the 17 TSE is. I think it is an indication of what is possible with current lens technology. Considering that it is a tilt shift lens, a lens with a similar design that has a smaller covering circle could be a bit more compact.

 

Even if you are not an architectural shooter, wide angle lenses on all cameras are often used to shoot buildings and other subjects where distortion would be objectionable.

 

I think Leica (and all other companies) are thinking about digital correction for distortion in the camera firmware or in post processing. This would allow for smaller "distortion free" lenses. It would be helpful if the camera or software automatically knew which lens was on the camera, the focusing distance, and the aperture. So I think this will require manual settings for the Leica either in the camera or in post. I'm guessing that the lack of EXIF lens data may be one reason (besides a small user base) why DXO does not support the M digital cameras.

[hoppyman]

A tilt/shift lens design could never work on an M for a number of reasons of course and image circle is not only affected by the form factor naturally.

Leica Camera is probably an exception regarding correction though, preferring to make the lens as good as possible rather than need those now common corrections (for CA etc) in firmware. The necessary exceptions have been for vignetting/cyan drift of course.

 

I'm not suggesting that moving the corrections in the image chain does not work but it is not the Leica approach. .

I think that philosophy is part of why customers are attracted to Leica lenses in the first place.

[AlanG]

A tilt/shift lens design could never work on an M for a number of reasons of course and image circle is not only affected by the form factor naturally.

Leica Camera is probably an exception regarding correction though, preferring to make the lens as good as possible rather than need those now common corrections (for CA etc) in firmware. The necessary exceptions have been for vignetting/cyan drift of course.

 

I'm not suggesting that moving the corrections in the image chain does not work but it is not the Leica approach. .

I think that philosophy is part of why customers are attracted to Leica lenses in the first place.

 

Regardless of what their approach normally would be... In this case, it seems that Leica left in the distortion in order to make it "as good as possible" within a certain size or some other constraint. So the only potential solution will be to correct it in firmware or post processing. Or live with it. I was suggesting that the 17mm TSE lens has a much larger image circle than is needed to simply cover 24x36mm on a non tilt/shift lens.

 

Thus it should be possible to make a nearly distortion free well corrected 17 or 18mm f3.8 lens (not a ts) that is significantly smaller than the Canon 17mm TSE. Whether this could be small enough to appeal to Leica M users, I can't say. But I think one must consider that many camera and software companies are moving in the direction where for all intents and purpose distortion will be a thing of the past for those who wish to avoid or eliminate it. (One can even use software to get distortion free images from fisheye lenses.)

 

So saying that one must live with distortion in any lens, let alone in an expensive prime, seems unreasonable to me. Adobe makes a system for user testing of any lens. It should be possible for users to come up with corrections for this lens without much trouble.

[lars_bergquist]

I just wanted to propose a comparison to the Zeiss 18mm. If it is completely different I should be glad seeing the same motive taken with it to learn that the Leica data sheet is wrong and the Zeiss - which tells us it has 0.2% less distortion - is right. The advertizing text for the Zeiss tells us, that it is "almost" distortion free.

The Leica data sheet for the Super-Elmar has a graph showing 1.9% barrel distortion. The Carl Zeiss sheet for the 18mm Distagon shows 1.0% barrel.

 

So, 1.9 -- 1.0 = 0.2. Interesting.

 

The old man who missed the New Math

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[hoppyman]

Alan I have no experience regarding the Canon lens whatsoever. It may be great on a dSLR, I have no clue nor opinion there.

The size constraint to work well on an M is obvious of course as is that a T/S cannot work on a rangefinder camera. I'm not an optical designer either. A pretty reasonable assumption that the Peter Karbe and his team are expert though and as a customer, I personally support their approach. Are you suggesting that the Canon design could be modified to do a better job? Have you any basis for that?

 

Yes you certainly can use the most recent correction capabilities in Photoshop etc to remove unwanted effects, that's fine. That's different to firmware correction of course, which is what I mentioned that Leica Camera prefers not to rely on. Profiling may well become a very useful tool as well. I'm sure that others will be interested too if users make available correction profiles for either.

I posted examples from the best two 18mm lenses in M mount that I am aware of. I personally have no problems with either and like any lens they will show their best when used intelligently. I never said either that you must live with anything. It's a very specialised focal length (especially for an M Rangefinder camera) in my opinion and I was pleased to replace my 18 with a 24 once I had my M9. Whatever suits each person is valid there.