Real world 21mm lens comparisons

[michaelwj]

Advertisement (gone after registration)

Wow I'm surprised on how well this lens controls distortion close up. Looks even better than the WATE or SEM. I'm assuming these are uncropped?

The lens has zero distortion, people don't get egg heads in the corners. This is the benefit of a symmetrical wide angle - you end up with a very natural looking wide shot. This is why those who use it love it.

To get this of course, we have dark corners and coloured edges on digital.

You win some and you lose some! :-)

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[the warrior]

True, in the corners there is less light, you must know the material you use to get good pictures.

I note that many people do not have patience to work with the material that some people do not like, many people carried away by the comments of other people without their own criteria, it seems to me well ask the opinion of the people but I want see for myself the reality, perhaps others do not know how to use well the lens, who knows ....

There are other photographers who themselves know how to use their material and new lenses found in other things, I see the results of SA f3,4 and I like, but I'll try the SAf4 to decide which one will buy, now I have a Elmarit ASPH 21mm and not offer me what I want.

Each photographer uses the material differently, do not all look the same.

A bad lens can become a great tool in the hands of an expert.

Maybe I'm wrong.

 

I have not a viewfinder, and my estimation of light (M3 and MP) and distance is to the eye.

 

Monochrom - SA f3,4

 

All photographs that show you are made very quickly.

 

[Hayek]

 

 

....people don't get egg heads in the corners. This is the benefit of a symmetrical wide angle - you end up with a very natural looking wide shot....

Is this statement correct?

[the warrior]

No.

[Peter Branch]

The lens has zero distortion, people don't get egg heads in the corners. This is the benefit of a symmetrical wide angle -

 

This is another of those statements that appear from time to time on the forum.

 

The lens is not free from optical distortion.

 

Optical distortion may, or may not, be minimised by a, so called, symmetrical design, which is actually a semi-symmetric design for reasons to do with the very large difference between the object distance and the image distance from the lens. 

 

"Retro focus type" i.e. highly non-symmetrical, lenses such as the 18mm f/3.8 SEM, have very low optical distortion. 

 

Perspective distortion is a function of the view point which with these lenses is extremely critical.  Move a few millimetres and it can change radically. 

[the warrior]

True friend Peter Branch.

I use habitualmenbte one 21mm and know the angular distortion that all are capable of performing.

Only the knowledge of our team, our lens makes framing know how to avoid much of this distortion, as you well know.

I insist: we must know the team to know how to use it better.

All teams are very good and all equipment can be very bad if the person behind not know how to use.

But maybe I'm wrong !!.

Now I look in lenses for many years some things that new ASPH lenses do not offer me.

Only one new lens, ASPH, which I think is the best of Leica, The Noctikux f0,95, some of my best works are done with this spectacular lens, but you have to know to know what is inside.

 

A lens is like a bride, you must spend time and effort, if not the bride will go elsewhere ......

[the warrior]

Advertisement (gone after registration)

Near of  the night, M3 and SA f3,4 to 40 centimeters ofthe faces.

 

[michaelwj]

 

Is this statement correct?

 

 

Lets just say I failed to clarify my over generalisation 

 

There are a few types of distortion;

1. Perspective - nothing to do with the lens.

2. Geometric - are straight lines wavy or curved? In general, more symmetrical wide angles are better than less symmetrical ones. The new 3.8/18 and 3.4/21 are examples of very corrected asymmetrical lenses.

3. Not sure what to call it, but is the image stretched into the frame? The more symmetrical the lens, the less it stretches the corners into the frame. The SA is very good at leaving round things in the corer round and hence my earlier statement about it looking natural. See for example http://photo.net/leica-rangefinders-forum/003aZa or check out the Hasselblad SWC - almost the exact same lens design as the SA/3.4, and another where the image is not stretched in the corners.

Most asymmetrical wide angles have a not very wide center and then stretch the image at the edges to fit more in. It just looks wrong.

 

So I stand by my statement that a symmetrical wide angle distorts less than an asymmetrical wide angle. Those who disagree are free to post examples where an asymmetrical wide angle leaves a circle as a circle in the corner.

 

Michael

[honcho]

I've read that amongst the VM lenses, QC issues and sample variance have been particular problems with the CV 4/21 (not as if those things don't frequently affect Leica or Zeiss).

 

Anyone experience this?

 

I once owned the CV 21mm f4 in ltm.   I didn't keep it long, it was one of the worst lenses I have ever owned.  Never really sharp at any setting, flare was a constant problem and the construction was poor compared to the CV 28mm f3.5.

 

A shame really, I was hoping it would be comparable quality to the CV 28mm f3.5 ltm, which I still have and use, but it wasn't.

[lct]

[...] 3. Not sure what to call it, but is the image stretched into the frame? The more symmetrical the lens, the less it stretches the corners into the frame. The SA is very good at leaving round things in the corer round and hence my earlier statement about it looking natural. [...]

 

Interesting indeed. First time i hear of such effect. Would you have some pictures to show it? Just curious.

[michaelwj]

Interesting indeed. First time i hear of such effect. Would you have some pictures to show it? Just curious.

 

Hi lct.

 

If you click through the link you'll see some direct comparisons.

In particular this one for the egg head (front bicycle wheel) http://photo.net/photodb/photo?photo_id=678288&size=lg and here http://photo.net/photodb/photo?photo_id=932271&size=lg the box has been stretched

Seeing the compression in the middle relative to the edges is harder to compare, this one shows somewhat what I'm referring to http://photo.net/photodb/photo?photo_id=678294&size=lg Note the way that the leading lines seem to "accelerate" towards the center in the Elmarit shot, where the SA is a more even progression. 

 

As most modern wide angles are not symmetrical, they all suffer the same stretching in the corners. Here is the fantastic Zeiss 21/2.8 ZF for example, note the head looks weird http://ogiroux.smugmug.com/Friends-and-Family/Maude-2009/DSC1566/586784417_ad8Cv-X3.jpg Compared to a beautiful photo by the warrior in this thread http://www.l-camera-forum.com/topic/259440-real-world-21mm-lens-comparisons/?p=3040434 with a similar placement of a head that looks perfectly natural. It doesn't suffer from the wide angle stretch that makes a typical wide angle shot look unnatural.

 

Michael

[lct]

Thank you Michael but i fail to see the effect you're referring to in those pictures. Our colleague Adan (Andy Piper) would certainly correct me if needed but SA and Elmarit have not the exact same focal length i suspect and the fields of views were not 100% identical seemingly. Also SA has a bit less optical distortion certainly but the difference is probably inferior to 0.5% so its effect on stretching should be minimal in this respect. I have no experience with SA 21/4 nor 21/3.4 lenses though so i remain interested in this subject.

[T*Sonnar]

Isn't that just a matter of geometric distortion? Why is it called symmetrical?

This term is new to me.

 

If you would 'correct' the geometric distortion in the image posted above you'll get the same effect:

 

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!

 

Oh and Hi, yes i'm new. I've been silently observing this topic.

[michaelwj]

Thank you Michael but i fail to see the effect you're referring to in those pictures. Our colleague Adan (Andy Piper) would certainly correct me if needed but SA and Elmarit have not the exact same focal length i suspect and the fields of views were not 100% identical seemingly. Also SA has a bit less optical distortion certainly but the difference is probably inferior to 0.5% so its effect on stretching should be minimal in this respect. I have no experience with SA 21/4 nor 21/3.4 lenses though so i remain interested in this subject.

The most obvious to me is the front bicycle wheel, which remains round in the SA shot (top images) and is an oval in the Elmarit shot (bottom images). I also understand that the difference is small but it does make wide angle people pictures look more natural.

Different people place there priorities in different places, some want straight lines, some want even illumination or even and high sharpness. I like a wide angle that doesn't stretch the image in odd ways.

I have no other 21mm lens to compare to, so I'm sorry but I'm unable to help more other than showing some more SA shots...

Michael

[michaelwj]

Isn't that just a matter of geometric distortion? Why is it called symmetrical?

This term is new to me.

 

If you would 'correct' the geometric distortion in the image posted above you'll get the same effect:

 

s2893_zpswqw9nwvr_edit.jpg

 

Oh and Hi, yes i'm new. I've been silently observing this topic.

Hi and welcome,

The lens itself is symmetrically constructed;

Large format super angulon https://en.m.wikipedia.org/wiki/Schneider_Kreuznach#/media/File%3ASuperAngulon.png

Rather than asymmetrical like a biogon;

https://de.m.wikipedia.org/wiki/Biogon#/media/Datei%3ABiogon21.png

 

Geometric distortions can of course be corrected, and is often done automatically in software or in camera. There is always some penalty for doing so though, loss of resolution or image area.

 

Michael

[T*Sonnar]

Oh i see that is symmetrical indeed. And symmetrical lenses are known to have more geometric distortion then i assume.

So why are we talking about it like they are two different things, when one is just a result of the other?

 

And don't worry it may be my first post here, but i'm no beginner.

Not concerned with corrections, just showing the difference.

[michaelwj]

Oh i see that is symmetrical indeed. And symmetrical lenses are known to have more geometric distortion then i assume.

So why are we talking about it like they are two different things, when one is just a result of the other?

Symmetrical lenses have LESS distortion as the light doesn't bend as much as it passes through the elements.

They also have lower illumination (and resolution) as you move away from the centre and don't play well with colour digital sensors.

Michael

[T*Sonnar]

I think you're mixing up the definitions of distortion.

 

A lens with less distortion (known as 'corrected'), will have faces and objects that are more distorted.

A lens with more distortion (like symmetrical ones apparently), will have faces and objects that are less distorted.

[lct]

I'm learning new things every day on this great forum.

[Advertisement]

Hallo,

Look here: Leica M

[michaelwj]

I think you're mixing up the definition of distortion.

A lens with less distortion (known as 'corrected'), will have faces and objects that are more distorted.

? How does less distortion lead to more distorted faces? Less distortion would surely have less distorted faces. A lens without distortion doesn't need correcting.

Michael