50's without distortion?

[Juan Valdenebro]

Hi,
I see the times when lens designers worried seriously about making lenses without distortion are kind of gone... After that, the chase for closest focus and highest sharpness wide open changed things... I like old lenses like the Tessar 3.5 and the Elmar 3.5, which IMO produce images with something "natural" in their rendering, as seeing with our own eyes...
Has anybody seen a list with 50's without distortion? I've read most Leicas have some distortion... Apart from RF lenses (main interest), I'd also like to know about mechanical lenses in F (Nikon) mount: I own a 35 for Nikon SLR's which has no distortion, AI-s I believe, a 2.8 produced for a short time, so maybe some members know about any 50 made with that priority too...
Thanks!
(Perhaps this thread should be moved I don't know where...)

 

[ferider]

Hi,
I see the times when lens designers worried seriously about making lenses without distortion are kind of gone... After that, the chase for closest focus and highest sharpness wide open changed things... I like old lenses like the Tessar 3.5 and the Elmar 3.5, which IMO produce images with something "natural" in their rendering, as seeing with our own eyes...
Has anybody seen a list with 50's without distortion? I've read most Leicas have some distortion... Apart from RF lenses (main interest), I'd also like to know about mechanical lenses in F (Nikon) mount: I own a 35 for Nikon SLR's which has no distortion, AI-s I believe, a 2.8 produced for a short time, so maybe some members know about any 50 made with that priority too...
Thanks!
(Perhaps this thread should be moved I don't know where...)

The Nikkor AIs 50/1.8 and Nikkor-H 50/2 that I have are quite clean, Juan.

Here is what I've tried on the RF side and found to be mostly rectilinear:

- all Summicrons (I've tried coll, rigid, v3 and v4)
- all Elmars (LTM 3.5 and 2.8, and M-Elmar)
- CV CS 50/2.5
- CV Nokton 1.1 (a surprise as the Nokton 1.5 barrels)
- M-Hex 50/2
- L-Hex 50/2.4
- classic Sonnars (Nikkor and Canon)
- Canon 50/1.8 LTM
- ZM Sonnar

Hope this helps,

Roland.

 

[Juan Valdenebro]

Thank you, Roland!

 

[Robert Lai]

I would add that the Nikkor 50 1.8 has no distortion at infinity, but as you get towards the minimum focus distance, I can definitely see some barrel distortion.

Macro lenses usually have no distortion.

 

[Daryl J.]

I'm told the short-lived V.1 50 Lux had little distortion. Can anyone verify?

 

[Jack Conrad]

Micro-Nikkor 55/3.5

 

[Doug]

Roland, I suppose there could be sample variation, but my Voigtlander Color Skopar 2.5/50mm has noticeable pincushion distortion... not really bad but noticeable when the image has straight lines parallel to the frame edges.

 

[Erik van Straten]

I agree with Juan that many 50mm lenses are not totally free from distortion.

The worst I know is the Jupiter 8M, but also some lenses that Roland mentions are not completely distortion free.

This distortion is above all disturbing when viewing the pictures on a monitor, usually the distortion is not so obvious on analogue prints.

The old Leitz 50mm lenses, from the thirties, are usually distortion free, such as the Elmar, the Hektor and the Summar. But the Summitar ... o my!

A lens with huge distortion is also the Summilux 50mm f/1.4 50mm v2. The v1 is practically free from it.

The Color-Skopar 50mm f/2.5 has a very little bit of distortion. The Elmar-M 50mm f/2.8 is distortion free.

Also the Nikkor-H Auto 50mm f/2 suffers from distortion, but the later f/1.8 Nikkor 50mm seems to be free from it.

On fast lenses the designer can choose between distortion or coma. When the coma is corrected, there is distortion. When the distortion is corrected, there is coma. Only on aspherical lenses both faults can be corrected completely.

Example: barrel vault distortion on Summilux 50mm f/1.4 v2.

Leica M3, Summilux 50mm f/1.4 v2, TriX.

Erik.

 

[Juan Valdenebro]

Thank you!

 

[ferider]

Juan,

I'm learning from the replies that some levels of distortion are OK for me, but not for others. Indeed, if you compare - for example - Puts' data of different Leica lenses (http://www.overgaard.dk/pdf/Leica-M-Lenses-Their-Soul-and-Secrets_en.pdf, pages 46-51), you'll see that

distortion(Summilux v2) > distortion(Elmar) > distortion(Elmar-M) > distortion(Summicron),

but in practice, for me, among those 4, only the Summilux is not acceptable (mainly because other, non-Leica lenses are better); the other 3 lenses "feel" similar. So it's subjective, obviously.

Roland.

PS:

I'm told the short-lived V.1 50 Lux had little distortion. Can anyone verify?

Check the data of the first Summilux mentioned in Puts' paper linked to above (page 50).

 

[Huss]

Example: barrel vault distortion on Summilux 50mm f/1.4 v2.

Leica M3, Summilux 50mm f/1.4 v2, TriX.

Erik.

Great shot Erik!

Elmar M 50 2.8 is distortion free.

 

[Emile de Leon]

The Elmar-M 50mm f/2.8 is distortion free.

I was wondering why I liked this lens so much..thx for the info!

 

[willie_901]

I doubt any lens is completely free from rectilinear distortion.

Radial distortions are errors in rectilinear image projection. If we ignore all other optical lens errors we only have to consider barrel and pin cushion and mustache distortion (where both barrel and pin cushion are present).

The issue really is: what level of distortion is acceptable? The answer is subject dependent. Architectural photographs often contain numerous objects with straight lines and rectilinear distortions are more obvious than portraits, sports, wild life, etc., photographs. Different photographers have different opinions on what's good enough.

Some rectilinear distortions such as volume anamorphosis distortion are not cause by lens errors. They are a function lens-to-subject distance and how far the object is from the optical center.

Post-production rectilinear error corrections don't bother me. What I care about is how well the correction model and the model parameters work. I have used a number of F-mount, ultra-wide angle zoom lenses that could not be fully corrected in post-production because higher-order barrel distortions were not included in the correction model. Corrected images from my FUJIFILM 10-24/4 XF lens contain extremely low levels of high-order rectilinear distortions.

 

[Dwig]

I doubt any lens is completely free from rectilinear distortion.

Radial distortions are errors in rectilinear image projection. If we ignore all other optical lens errors we only have to consider barrel and pin cushion and mustache distortion (where both barrel and pin cushion are present).

The issue really is: what level of distortion is acceptable? The answer is subject dependent. Architectural photographs often contain numerous objects with straight lines and rectilinear distortions are more obvious than portraits, sports, wild life, etc., photographs. Different photographers have different opinions on what's good enough.

Some rectilinear distortions such as volume anamorphosis distortion are not cause by lens errors. They are a function lens-to-subject distance and how far the object is from the optical center. ...

Excellent, well said.

If we limit ourselves to rectilinear distortion (barrel, pincushion, and the various "mustache" mixes of the two - oh, iPhone, thou art evil...) the following are reasonable generalizations:

• The faster the lens the more likely there will be visible distortion. Other considerations are likely to have outweighted distortion corrections when the designers decided what was important.
• The more retrofocus or telephoto, as distinct from "long focus", the lens the more likely that there will be distortion.

 

[ferider]

The issue really is: what level of distortion is acceptable?

Very much agree.

If we limit ourselves to rectilinear distortion (barrel, pincushion, and the various "mustache" mixes of the two - oh, iPhone, thou art evil...) the following are reasonable generalizations:

• The faster the lens the more likely there will be visible distortion. Other considerations are likely to have outweighted distortion corrections when the designers decided what was important.
• The more retrofocus or telephoto, as distinct from "long focus", the lens the more likely that there will be distortion.

Not generally true, the Nokton 1.1 is a good counter example.

These are the 5 Seidel Aberrations:

- Spherical Aberration
- Coma
- Astigmatism
- Curvature of Field
- Distortion

For faster lenses, good distortion correction usually means one or more of the Seidel aberrations suffer. For example, my VM Ultron 35/1.7 is quite rectilinear at the cost of curvature of field. The v2 Summilux is documented to be optimized for Coma, at the cost of distortion. The ZM Sonnar is quite rectilinear, but has noticable Spherical Aberration ("focus shift"). Etc.

And what looks like identical designs can be optimized differently. For example, I have several 50/1.5 classic LTM Sonnars that have flat field, but a very late Contax mount Zeiss Sonnar that has noticeable field curvature.

Roland.

 

[f16sunshine]

Roland, I suppose there could be sample variation, but my Voigtlander Color Skopar 2.5/50mm has noticeable pincushion distortion... not really bad but noticeable when the image has straight lines parallel to the frame edges.

Hi Doug
I Have noticed that as well at near focus. Getting closer to infinity it's difficult to see or maybe even gone (doubt it).
Still my overall favorite RF 50mm for b&w film.

 

[ferider]

Elmar M 50 2.8 is distortion free.

It's all relative.

 

[Juan Valdenebro]

Interesting...
As it's not covered in Roland's link (Puts' PDF), where can I get some information on the Collapsible Summicron? Not just distortion...
It would be fine to read a good while about its design and how its formula "preferred" which aberrations, if it can be said that way...
Anyone having a link or any article/book talking about it with some depth? Was it just Mandler?
This subject of someone -or a team- deciding what to correct and what to allow optically, is totally fascinating to me... (Maybe because I know so little about it! )
Thanks!

 

[Erik van Straten]

I'm told the short-lived V.1 50 Lux had little distortion. Can anyone verify?

Almost no distortion indeed.

Leica M5, Summilux 50mm f/1.4 v1, 400-2TMY.

Erik.

Leica M5, Summilux 35mm f/1.4v1, 400-2TMY.

 

[Erik van Straten]

For example, my VM Ultron 35/1.7 is quite rectilinear at the cost of curvature of field. The v2 Summilux is documented to be optimized for Coma, at the cost of distortion.

In general I can accept a little bit of linear distortion on a 50mm lens, but not on a 35mm lens.

A set of the Summilux 35mm f/1.4 v1 and the Summilux 50mm f/1.4 v1, both almost 60 years old, is close to ideal in my opinion.

Erik.