CV 21mm F4 Color Skopar on Leica M9

[bocaburger]

Advertisement (gone after registration)

Those that are compact show the most red-edge: 15 SW Heliar, 21/25 Skopars, 21 Super-Angulons, Zeiss 21 f/4.5. The waters are muddied a bit because most of them are not Leica lenses, and therefore have no dedicated code to correct them precisely.

 

Digital sensors just behave better with light delivered through a long tunnel, at least for the time being.

 

I think the size of the lens is a red herring, and you touched on the real reason: no dedicated code. The reason I believe this is that the pre-ASPH 21 Elmarit which is less compact than its ASPH successor, suffers worse red-edge on the M9. I believe it comes down to less adequate firmware. It is the oldest ultra-wide for which Leica formulated a code, and being a discontinued lens they clearly have less to gain from spending $$ to refine its code than for lenses they're still producing and selling. Still I hope they will tweak it in a future update. It doesn't have that far to go.

 

Another thing, my 28mm f/1.9 Ultron, physically longer than the 28 Summicron let alone the 28 Elmarit-ASPH, also exhibits severe red-edge which I have not found any substitute Leica code to correct.

[Advertisement]

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

[thighslapper]

Shootist.....

 

I'm no expert but from what I recall the Kodak sensor consists of a Bayer filter mosaic in a fixed pattern of 50% green, 25% red and 25% blue. The arrangement must be lopsided with the red or the blue filters either being on the Right or Left of each cell. The cells are all stacked uniformly across the sensor, with the RGBG filters always in the same pattern. From the Leica blurb this is overlain in the periphery of the sensor by microprisms to compensate for the very oblique rays from wide angle lenses and attempts to attenuate this spreading of the rays and loss of sharpness. Presumably as this is fixed and only works up to a point there must still be some uncontrolled aberrations in the periphery. Is it possible that this effect selectively stimulates more of the left (or right) side of each cell the further out into the periphery you go giving a specific colour cast depending on the arrangement of mosaic pattern ? My knowledge of optics is insufficient to know whether this is possible, but the bayer mosaic pattern arrangement is the only 'handed' part of the optical path that can explain a Right/Left arrangement of the effect.

 

I assume Leica dial out both vignetting and the worst of this by doing exactly what cornerfix does - taking a uniform 18% grey image from each lens and subtracting out the abberrent parts that don't match the centre.....

 

I'm surprised some enterprising nerd hasn't hacked into the firmware and found out exactly how Leica code this and supply profiles us simple folk can load up ..... the firmware must consist in large part of a library of 'cornerfixes' for all the coded lenses.

 

If Leica were helpful they could use a selection of spare 6 bit codes to produce a matrix of corrections that users could trial for non Leica lenses which could correct the worst of these. The list of non Leica lenses that are commonly used is pretty short .... and only a few have problems that are bad enough to be worth correcting...... it would be nice to think that one of their software team (or is it Jenoptiks?) with a few hours to spare would be kind enough to knock out the codes for us.........

Edited November 17, 2010 by thighslapper

[adan]

"I think the size of the lens is a red herring,...."

 

There are always exceptions, but statistically, a shorter compact wideangle design is more likely to run into red-edge or other problems on digital. What works "best" for digital is long, tunnel-like optics.

 

(Take "long, tunnel-like optics" to be shorthand for a page-long dissertation on telecentricity, angle of incidence, exit pupils, etc. - which I'm going to skip).

 

Neither of the Zeiss ZM 21 lenses have dedicated "Leica" codes - but the f/2.8. with its much longer, tubular optical path (see image), is more correctible than the stubby f/4.5.

 

per digilloyd's reviews:

 

diglloyd: Zeiss ZM 21mm f/4.5 C-Biogon for Leica M9/M8/M and Zeiss Ikon

diglloyd: Zeiss ZM 21mm f/2.8 Biogon for Leica M9/M8/M and Zeiss Ikon

 

Short, stubby wideangles - for the most part - fall outside the range of corrections Leica offers, whereas longer, more digitally-friendly optical designs of the same focal length DO fall within the correctible range.

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!

[Keith (M)]

"I think the size of the lens is a red herring,...."

 

Neither of the Zeiss ZM 21 lenses have dedicated "Leica" codes - but the f/2.8. with its much longer, tubular optical path (see image), is more correctible than the stubby f/4.5.

 

I have just ordered the Zeiss ZM 21 f2.8 for my M9, plus the Coder Kit 2.0 (I also have the ZM 35 & 50mm). Once I have had a play, I'll let you know how the 21mm performs!

[adan]

Thighslapper. there is no "handedness" to Bayer patterns. Any green pixel has either red, or blue, pixels on both sides of it. In the image below, ALL of the represented "cells" can be found on any part of a Bayer-patterned sensor - depending on where one arbitrarily draws a box around any four pixels.

 

It should be thought of not as a tiling for 4 pixels in a "cell", but as a tiling of single pixels (each with their own microlens overhead). Just as a checkerboard is not a black surface with red squares on it - or a red surface with black squares on it - but simply a combination of red and black squares.

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!

[adan]

Keith - note that digilloyd recommends coding (or manually selecting) the corrections for the "21mm f/2.8 pre-ASPH (11134)" as working well with the 21 f/2.8 Zeiss ZM.

 

The coding for the 21 pre-ASPH Elmarit seems to be the "go-to" code, with the strongest corrections, for a number of non-Leica lenses. Ironically, as bocaburger points out, it doesn't work perfectly (yet) for the intended lens.

[Keith (M)]

Advertisement (gone after registration)

Keith - note that digilloyd recommends coding (or manually selecting) the corrections for the "21mm f/2.8 pre-ASPH (11134)" as working well with the 21 f/2.8 Zeiss ZM.

 

Thanks, Andy. As part of my research into lenses before taking the plunge into Leica-World (!) I took out a subscription to digilloyd and it has been very useful indeed. Your reminder re the optimum coding for the ZM 21mm is very timely.

[thighslapper]

Ok Andy...... just a thought.....

 

I cannot see any other way that colour casts can appear only on one side of a sensor other than some peculiarity in the way the sensors cells resolve the colours of the incident light rays at the periphery.

 

There must be an explanation that involves something to do with the Bayer array and its arrangement in the sensor......

[thighslapper]

I've just checked all my Leica 18mm/3.8 pics and cannot find any easily noticeable vignetting or any colour aberrations in the periphery (apart from the expected magenta fringing of acute light/dark boundaries) so Leica must have very successfully dialed this out in-camera (M9).

 

Shame they have not produced a matrix/set of spare 6 bit codes to try for non-leica lenses as I suggested above......

[bocaburger]

Leica doesn't put any codes in the menu for lenses that have only been produced with 6-bit coding, but that doesn't mean one couldn't hand-code C/V lenses like we did with the M8. Maybe the code for the 18mm, or 21 Summilux, or one of the new 24s might be appropriate for some of the C/V lenses.

[lars_bergquist]

Ok Andy...... just a thought.....

 

I cannot see any other way that colour casts can appear only on one side of a sensor other than some peculiarity in the way the sensors cells resolve the colours of the incident light rays at the periphery.

 

There must be an explanation that involves something to do with the Bayer array and its arrangement in the sensor......

 

Or possibly the microlenses. Or quite possibly something else entirely.

 

The fact is that I have not seen one single convincing, scientifically credible explanation of the entire phenomenon. Which, as we should remember, is a general digital image sensor problem, not just a Leica M9 problem. Also, it is not just a matter of "rededge" but also of "greenedge" to the right. But a cyan shift blends in better with most backgrounds than a red one. This goes to show the general lack of scientific or technical methodology here -- wild guesses are a dime a dozen, and are sold as Revealed Truth.

 

Leica Camera, or possibly Kodak, should tell us the Reason Why. Or maybe they think we have all signed up to Do or Die.

 

The old man from the Age of the Red Left

[lars_bergquist]

Now back to the theme of this thread. The C/V 21mm has not only a rededge problem. It has a fuzzy corners problem. I have used that lens on film, but was finally fed up with it, because no amount of stopping down could produce acceptable corners. I can well believe that the lens performed acceptably on the M8, because both the red/green edge and the corner fuzz was cropped away. But no more.

 

The old man, as above

[bocaburger]

Lars, my perception is there may be a fair amount of sample variation among these. My 21 Skopar performs crisply in the corners even at f/4, improving a bit to f/8. It also doesn't have more than about 1 1/2 stops of vignetting wide open, also improving by a couple stops down. A friend of mine had the same lens and sold it because it vignetted severely (about 3 stops). (My current copy is actually my second lens, the first one got stuck at infinity the first time I used it, and I had to return it for a replacement.)

[artspraken]

Thanks for all the valuable comments.

 

I decided to get the Zeiss 21mm Biogon F2.8 for my Leica M9.

 

Works perfect. No vignetting or color shift on edges. For manual lens setting, I selected Elmarit 21mm F2.8

 

For those of you using M9 looking for a 21mm (and cannot afford Leica), the Zeiss Biogon is certainly viable choice.

[Keith (M)]

Thanks for all the valuable comments.

 

I decided to get the Zeiss 21mm Biogon F2.8 for my Leica M9.

 

Works perfect. No vignetting or color shift on edges. For manual lens setting, I selected Elmarit 21mm F2.8

 

For those of you using M9 looking for a 21mm (and cannot afford Leica), the Zeiss Biogon is certainly viable choice.

 

Agreed! My ZM 21mm arrived a week ago and in terms of colour rendition and crispness etc results have been superb. Most of my wide-angle photography up until now was with the EF 16-35mm f2.8 on my 5D MkII and whilst it may not be completely fair to compare zoom against prime, the ZM 21 results have been a revelation. Some examples from my first outing are on my Flickr page (see below for link).

[Xmas]

Hi

 

The sensor micro lens compensation is (mainly) dependent on the position of the rear nodal point and the aperture of the taking lens (and I think - the focal length - as the angle of the arriving cone of light is the actual dependency).

 

I wont try and explain what a nodal point is but...

 

http://en.wikipedia.org/wiki/Cardinal_point_(optics)

 

If you are lucky CV might bring out compact 'digital' wides...

 

Noel

[Washington]

A sensor and film are two different things and, being so, act differently.

Apples and oranges.

[Sonnyside]

CV 21 MM test | Flickr - Photo Sharing!

M8

don't have M9 yet but i am happy .and yes resolution as shot with Leica brand lens a lot better

Edited December 20, 2012 by Sonnyside

[Advertisement]

Hallo,

Look here: Leica M

[Notman]

I use a V C 21 mm on my M 9. To set 21 mm lens detection I went to Menu, Lens detection,Manual, then 21 f2.8 11134. Try this I found that it helped quite a lot.

[jaques]

I think the VC wide angles are fantastic lenses despite any red edge issues. You can always black and white- or use cornerfix- or crop:

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!