MTF charts for first four S lenses

[dfarkas]

In case anyone missed this, Leica posted the MTF charts for the first four S lenses on the S-System micro site. Each chart has performance wide-open (left) and stopped down slightly (f/5.6). The 35, 70, and 120 are all f/2.5, meaning that the they are only being stopped down about 2.5 stops. The 180 is an f/3.5 lens, meaning that it is only being stopped down about 1.3 stops.

These are very impressive results and backs up the claim that no software correction is needed to achieve maximum performance.

David

[marknorton]

It seems clear that lens quality will be the deciding factor in the S2 image quality. In the current LFI issue, there's an interesting comparison showing the contribution to the close focus quality made by the separate moveable lens group. Without it, the quality falls apart.

In the same issue, the question is asked about the cost of the S lenses. The 70mm is said to be of comparable complexity to the 28mm Summicron-M but has autofocus, electronic aperture, water-proofing in addition. The Summicron is €3000 so that suggests the "cooking" 70mm lens will be, what, €5000?

I unexpectedly find myself the owner of a D3x and it seems clear that just as the M8 is stretched to the limit by its rangefinder, the D3x is limited by the quality of even the best Nikon glass.

Truly, that S glass is going to be decisive - irrespective of price - and Peter Karbe is clearly a genius! As for Nikon, they need some no-holes-barred primes...

[marknorton]

BTW, it's clear the S2 is still a work in progress, not just firmware tweaking. The shooting session just published was with no AF and the LFI internview also mentioned the central leaf shutter is being developed and worked on by Leica. Seems clear the central shutter will add significantly to the lens cost.

[jaapv]

I think, Mark, if the lenses really deliver what these curves show, they will be very close to sensational.

[zapp]

Nice looking graphs. Assuming that we get to see the typical 5, 10, 20, 40 LP graph someone should tell Leica that other manufacterers include the 60 LP graph as well. And you need this high resolution just to resolve according to the pixel pitch. How about distortion? No graphs yet. The 35 mm lens mock up was longer than the 70 mm lens. Resolution may be fine, but I wonder about distortion.
The details provided in the LFI artcile are not worth much. Floating elements are not new and if a lens is designed to have floating elements it will fall apart when you would fix them.
Still. The graphs look very promising so far.

[stephengilbert]

Maybe they should give the camera away and charge 10,000 euros per lens. Didn't Polaroid sell their cameras cheaply to create a market for their film?

Originally Posted by marknorton

the contribution to the close focus quality made by the separate moveable lens group.

It's amazing that Leica starts to promote that feature now. Nikon introduced their close range correction with "floating elements" back in 1968, with the release of the 24/2.8

[dfarkas]

Originally Posted by volkerm

It's amazing that Leica starts to promote that feature now. Nikon introduced their close range correction with "floating elements" back in 1968, with the release of the 24/2.8

Leica has used floating lens elements for a while now. The 50 Lux ASPH and 75 APO both use floating elements, not to mention all four of the new M lenses shown at Photokina and even older R lenses like the 28 Elmarit.

David

[rosuna]

Impressive S lenses... Thank you for posting...

LEICA Camera AG - Das Leica S-System

[rosuna]

Originally Posted by zapp

Assuming that we get to see the typical 5, 10, 20, 40 LP graph someone should tell Leica that other manufacterers include the 60 LP graph as well.

That's nonsense.

Manufacturers usually publish 10 and 30lp/mm curves. Only Olympus publishes 60 lp/mm, but that is due to the 4/3 format being half the 35mm format. Olympus's 60lp/mm curves are equivalent to 35mm's 30lp/mm curves...

[coronita]

I wonder why Leica posts so confusing graphs. Normal procedure is to use different colors for contrast (10 L/mm) and resolution (30, 40 or 60 L/mm depending on sensor format). To use same color for four (¡!) different values and not stating what those values are is misleading, and looks to me like a bad "designer" job.
By the way: one should not forget that those MTF values are usually "computer proyections", not actual performance of "store samples". Anyway, they look really good.

[rosuna]

The graphs are simple to read for me... The higher pair is 5 lp/mm, then 10 lp/mm, then 20 lp/mm and 40 lp/mm, from top to bottom...

[scott kirkpatrick]

I agree that the curves must be the standard 5, 10, 20, 40 (two coarse, two fine values). The dark and light lines are radial and tangential resolution, I suppose, but which is which? And what is the horizontal axis? It's distance from the center of the image as a fraction of something, but is that the image diagonal, the horizontal extent, or what? The faint writing didn't answer that for me.

They do look very crisp across the field at f/5.6 (a good artificial light working setting) but a touch soft at the edges wide open.

scott

[stunsworth]

Originally Posted by scott kirkpatrick

And what is the horizontal axis? It's distance from the center of the image as a fraction of something, but is that the image diagonal, the horizontal extent, or what?

It's over 27mm - see the text underneath the graphs.

[R2-D2]

Originally Posted by dfarkas

Leica has used floating lens elements for a while now. The 50 Lux ASPH and 75 APO both use floating elements, not to mention all four of the new M lenses shown at Photokina and even older R lenses like the 28 Elmarit.

The new Elmar-M 24mm ASPH. has no floating element.

[scott kirkpatrick]

Originally Posted by stunsworth

It's over 27mm - see the text underneath the graphs.

You are right, I didn't look hard enough. That is the distance out to the corner of the frame from its midpoint.

scott

[coronita]

Originally Posted by rosuna

The graphs are simple to read for me... The higher pair is 5 lp/mm, then 10 lp/mm, then 20 lp/mm and 40 lp/mm, from top to bottom...

But... different colors, as is standard procedure anywhere, would make interpretation much more easier... don't you agree?
I insist: this is a bad graphic designer decision... in my humble opinion and based on my long experience.

[rosuna]

Well, different colors may help...

[leicar7]

I suggest reading E. Puts's articles on MTFs and other image quality issues on his web site.

There is no standard for using different colors for the line pairs per millimeter graphs.

The solid lines are the radial and the broken lines the tangential contrast lines.

Zeiss does not show 60 lppm on their mid-format lenses: only 10, 20, and 40 lppm. If they don't, then 60 lppm is not standard.

As for edge performance, I challenge anyone to show anyone's lenses where the performance at the edges is such a high percentage of the central performance for wide, normal, and short tele lenses. Very few lenses, even from Leica, and none from other makers, render contrast so uniformly and at such high levels across the image frame at these wide, normal, and short tele focal lengths. The S lenses, at least as far as the MTFs indicate, look like a new level of contrast rendering.

Given the economy, I begin to wonder about the chances that the R system may end up being shelved and Leica carry on with the S system. It will certainly redefine reflex photography and image quality from portable cameras. But I certainly hope that Mr. K has the deep pockets to ride this out while giving us a new R. If new wide and normal R lenses are to have this level of contrast rendering at the edges, they will set a new standard for reflex 35mm photography.

[AlanG]

Originally Posted by leicar7

As for edge performance, I challenge anyone to show anyone's lenses where the performance at the edges is such a high percentage of the central performance for wide, normal, and short tele lenses. Very few lenses, even from Leica, and none from other makers, render contrast so uniformly and at such high levels across the image frame at these wide, normal, and short tele focal lengths. The S lenses, at least as far as the MTFs indicate, look like a new level of contrast rendering.

I think it is great if Leica makes wonderful lenses. But from my perspective, there may be diminishing returns for some photographers. I have a long background shooting MF - Hasselblads and Rollei 6000 series for 6x6 and Linhof with Zeiss, Rodenstock, and Schneider glass for 6x7, 6x9, 6x12. And of course 4x5 and 8x10 too.

You know what? I think I shot almost all of my MF photos at f8 or smaller. (The view camera lenses were almost always at f11-22) I once compared my Zeiss 80mm Planar/Rollei 6006 with a $60 Ukranian Arsat 80 on a Kiev 66. At f8 the images looked excellent from both lenses. (I can't recall what I thought of them at f2.8) You don't always need such great lenses for good results on 6x6. (That's why lots of cheap TLR's gave such good results.)

So let's say the Leica glass is great at the corners and center wide open. When do you shoot MF wide open? Maybe portraits when you desire minimal depth of field. But you don't usually need sharp corners for this. Most static objects are shot stopped down - generally using a tripod if one is after maximum sharpness. Perhaps the longer lenses will often be used wide open for fashion and the like. But I don't think I used my 350 TeleTessar wide open too many times. And I don't see that many sport shooters going with any MF system.

So while the MTF charts may be incredible, all that will really matter is if a typical user can really see a difference in the type of photos he/she shoots with the S2 versus another system. In the case of the S2, perhaps more users will be shooting at larger apertures or wide open if the camera truly can displace some of the work that is typically done on 35mm.

While the S2 has a number of f2.5 lenses vs.what is commonly 2.8-4 on 6x6 systems, the Rollei system has a 110mm f2 Planar and a 180 f2.8 for those who want to shoot MF with less depth of field. The H3D has a 100 f2.2. I don't know how good those lenses are wide open but they are also working on a larger format.