Jump to content

Wideangles magenta shift


satureyes

Recommended Posts

  • Replies 147
  • Created
  • Last Reply
However, retro-focus or not, what is generally being talked about in this thread is the degree to which a lens exhibits image-space tele-centric-type behaviour ...

No, this is not what's being talked about.

 

 

Image-space tele-centric-type behaviour, (i. e. exit-pupil distance tending towards infinity) ...

Image-space telecentricity indeed means an infinite exit-pupil distance. However in Leica M lenses, the exit pupils are nowhere near infinity, so they most definitely are no "image-space telecentric-type" lenses.

 

I think it's Olympus' marketing division who originally brought up this "near tele-centric" nonsense—which is even more stupid than the nonsense term "retrofocus-like". So please note: Regular lenses are not, repeat: NOT telecentric or "near-telecentric" or "telecentric-like" on the image side.

 

What actually is meant by all this inflated buzzword bragging just is a (slightly) increased exit-pupil distance. But when this distance in a modern wide-angle lens is, say, 25 mm whereas in an older model it used to be, like, 18 mm or so then this has nothing, repeat NOTHING to do with telecentricity. Not even "nearly". Not even remotely. Not even when the lens was a true retrofocus.

Link to post
Share on other sites

I am sorry you are having this trouble but glad you posted about it, another reason I think I will keep the M9 and not "upgrade."

 

I used to use Cornerfix on 21mm Zeiss images but then had the lens 6 bit coded. Can't help with how to code out your lens though. What I did find was that with a little bit of experimentation I was able to create a Lightroom preset that did almost as good a job as Cornerfix.

Link to post
Share on other sites

........Regular lenses are not, repeat: NOT telecentric or "near-telecentric" or "telecentric-like" on the image side........

 

Your advocacy of the “True definition” of retro-focus is admirable – but irrelevant.

 

If you want to stop the adverse effects of high angles of chief ray incidence on digital sensors then the issue under consideration is the distance of the exit pupil from the sensor – plus having the correct rear element diameter.

 

My post does not suggest that Leica M lenses are tele-centric – hence there is a problem – but this has little or nothing to do with whether the lens is “retro-focus like”.

 

The words are not “inflated buzzword bragging” they are an attempt, clearly failing in certain quarters, to use terms which have precise meaning.

Link to post
Share on other sites

I have used the 18se with my m9 and no problem.odd that this should occur with the M240

 

Well, I think the experience you describe is part of the problem. Before the problem was effectively reduced for the M9 by a new firmware in 2011, I saw the red edges regularly when I used the 18mm Super-Elmar.

 

But there were certain situations where your photos were completely free of the slightest traces of this problem. This depended on the "condition" of light. Someone from the German forum put it right when he observed this, saying: "What can we do about it - shall we shift the light?"

 

With the M 240 we may experience it the other way round: the problem is normally effectively reduced - but it still exists under certain light conditions. So the firmware needs some more very subtle tweaking: to reduce it under any conditions which might lead to some compromises as we saw for the late M9.

 

When the new sensor for the M240 was introduced I misunderstood the descriptions and thought the problem was completely solved. Now I learn this was wrong. Since Leica is no longer dependant on exterior support for the camera electronics one could expect a better solution than last time.

Link to post
Share on other sites

ok so.. first trip out with the new M..

Loving it all - but I'm totally shocked at the purple band along the edge of the frame in my 21mm Super Elmar shots..

 

Earlier I saw it happening massively on the Voigtlander 12mm but I put this down to the non-Leica lens and how wide it is.. but looking at almost all my new M images with the 21SE there's a magenta band along the edge of shot.. as well as some pretty annoying vignetting.

 

Shame - these were not there on the M9 - and I know firmware etc etc.. and corner fix etc etc but surely for a New(isn) released Leica lens this shouldn't be an issue on the new M?

 

See what I mean.. here..

 

500px / Photo "Oh I do Leica to be beside the sea-side" by Rick Bronks

 

I am seeing the Italian Flag Syndrome (reddish/magenta cast) with my M. Oddly, it comes and goes--I haven't seen a pattern as to when it is likely to appear--so far at least. It is most noticeable along the right hand side of the image. I see it occasionally with my 24mm Elmar ASPH.

 

What seems strange to me is--I see it more often and more strongly with my almost brand new 35mm Summicron ASPH. Both coded correctly of course.

 

If you had the M9 early on, I would think you would have seen the fringing with your wides up to the point where they corrected for it in a FW fix in the spring of 2011. Leica did a great job of correcting the problem with their wides at that point. I trust the same can be done with the M, and I hope Leica will test for this and address sooner rather than later. I don't think Leica owes it to us to do a fix for non-Leica lenses, though it would certainly be a welcome gesture.

 

Cornerfix is a wonderful solution, and we are all in Sandy's debt for providing it.

 

Bottom line: when I buy a Leica M-mount lens and a Leica M body, I shouldn't experience anomalies like this assuming they could have been fixed via FW--period. It is not my responsibility as a customer shelling out many thousands of hard-earned dollars to wait for solutions that should have been created prior to releasing the M body. I would think that given their experience with this issue with the M9, someone would have watched for it with the M and addressed it prior to the first M delivery.

 

That's my opinion for what it's worth.

Link to post
Share on other sites

Advertisement (gone after registration)

Here a comparison using the 21/2.8 ASPH at f/8 on the M9 and M (typ 240). Extreme upper right hand corner of the image at 100%. Same settings in LR5. (tiff filesize 25 Mb). The 21/2.8 ASPH profiles were not used on either camera.

 

compareM9M240with21uncoded.tif

 

Form your own opinion based on these images or do your own experiments.

 

Interesting, I see just a faint cyan cast on the M9 and a very faint magenta cast on the M240.

Seems quite acceptable to me.

I am still on Lr4... are you using Lr5 lens correction profiles ? What if you open the DNGs with a plain preview app ?

Link to post
Share on other sites

Just a different point to add - I shoot with the 14mmL on my Canon 1DX and this also randomly shows vignetting, usually in one corner. No colour cast though..

 

If it is always the same corner, some elements might be decentered.

Link to post
Share on other sites

Hi,

 

I believe Royal_Corona and UliWer to be correct. From the posted photos we are talking about "the Italian flag syndrome" (IFS), that is colors on the sides due to sensor/wides combination.

 

For future reference please post photos taken against a neutral grey or white background.

 

I do a far amount of winter snow photographing, and have to say that all wides have this problem to some extent on the M9 (the 28/2,8 ASPH being one of the worst modern wides I have used). The degree of color misconfiguration depends on the nature of the light, (direct sun, softened by clouds, no sun, etc), the lens AND the chosen aperture. So a solution to solve this will have to take into account all these parameters.

 

The FW on the M9 handle most of the Leica lenses ok for summer photo's and winterscapes given that you are not very critical. So thats the easy way around this. For some of the CV / Zeiss and other wides the built-in profiles are not optimal as to the the IFS.

 

The best solution is Cornerfix where you can make profiles handling ITS and vignetting. The profiles can be tailor made for your lens/sensor for each aperture, as well as for different light if you care. The batch job of correcting photos are quite easy to integrate in your workflow. It requires a little work to make profiles since you have to make photos against a neutral background per profile.

 

So you can perhaps expect Leica to improve this by FW for their lenses, but don't expect any easier solution than Cornerfix for non-Leica lenses, nor for really critical work. Its really the same for other mirrorless.

Link to post
Share on other sites

Perhaps the next Leica M model should be a type C, where the C stands for curved sensor?

 

I think that I have read somewhere, that there exist a patented curved sensor, designed to counteract just these phenomena, which are discussed here in this thread.

Link to post
Share on other sites

Back on track again... I thought I'd give Cornerfix a go.. so in 20mins I had it all sorted.

 

I am running out right now - BUT - corner fix and the CV12mm are perfect. I shot a bunch of shots with the dreaded magenta cast, plus a vignette thrown in for good measure.. lens as uncoded.

 

Ran them through Cornerfix.. PERFECT results.. did it a few times. It's like Voodoo.

 

Yeha - it's a bit of a PITA to have to run this on the shots in post, but hey.. hopefully they'll be worth it.

 

Ok.. I'm a bit happier now and even more so that I'm off to shoot 3 evenings of a mad festival so will have ample time to run the M through the mill with various lenses.

Link to post
Share on other sites

If you use Lr do try the flat field convertor as I have now found it does a perfect job on the colour problem and all within the Lr workflow

 

Alas I'm an aperture user - so switching to LR isn't really an option- I am pretty certain the next release of Aperture will contain much of what LR has.

 

To be honest - the 12mm is pretty specialised as a wide and ill only be using it when I need the look it gives.

 

I'm more concerned about when Alec's will fix the issues that affect their current lenses.

Link to post
Share on other sites

Alas I'm an aperture user - so switching to LR isn't really an option- I am pretty certain the next release of Aperture will contain much of what LR has.

 

To be honest - the 12mm is pretty specialised as a wide and ill only be using it when I need the look it gives.

 

I'm more concerned about when Alec's will fix the issues that affect their current lenses.

 

I hear Alec is software engineer extraordinaire at Leica's Solms facility. He'll have all sorted in due time. ;)

Link to post
Share on other sites

...

For future reference please post photos taken against a neutral grey or white background.

 

...

 

Photos of any background might lead to accidential differences - as you cannot influence the light which hits your background. With very wide angles you also might have problems to find something which fits.

 

My recommendation would be to place a diffusor right in front of the lens - or better flat on the hood. Some semi-transparent devices for measuring the white balance are not good for this purpose since they do influence the light rays too much, which might lead to unrealistic extreme results.

 

I use a plastic sheet protector which is not transparent but dim white and put a sheet of dim white semi-transparent paper in it. The whole "test chart" must let the light through, but you should not be able to recognize any object behind it. Put the Camera in direction of natural light which doesn't show much differences and especially no green or red colours (e.g.: avoid having green trees or red flowers in front of your lens). Overexpose by +1 EV.

 

This gives you the most reliable and reproducible idea if there is a red shift pattern and/or italian flag issue and in which parts of the frame it shows.

Link to post
Share on other sites

God damn IPad autocorrect. Clearly has no class.

 

Obviously that was meant to say Leica!!

 

I've been a victim of autocorrect plenty of times.

 

I hope the humor come off as friendly rather than antagonizing.

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...