M9 and Red Edge Redux--Part II, with 28mm Elmarit?

[mjh]

Advertisement (gone after registration)

His assumption that the red edges are a result from the in-camera correction of vignetting and cyan drift (which is applied especially for wide angle lenses) is is the only theory which I have found convincing ever since the issue was discussed.

That’s what I used to think, too, but clearly it isn’t the whole story and I’m not even sure anymore that it is a part. It’s not too difficult to find an explanation for red or cyan drift – the effectiveness of the IR-absorbing filter on the sensor (and also the transmission characteristics of the dichroic UV/IR filter on the lens in case of the M8) would depend on the incident angle, resulting in red vignetting and thus cyan edges; overcorrection of this effect would then produce red edges. But this effect would be approximately symmetrical around the optical axis, just like vignetting. I am not aware of any satisfying explanation of the markedly asymmetrical effect we are actually observing. Neither am I aware of any explanation that could account for noticeable variance between different cameras and/or lenses – while there might be differences between different copies of a camera or lens, these differences should be quite small and the corresponding differences in the effect produced should be equally small.

 

Also the Italian Flag phenomenon isn’t new; MF photographers knew about this issue before we did, implying that there is nothing special about the M9 or M8, no secret ingredient that wasn’t found in other cameras before. Some people did suggest that the asymmetry of the Bayer filter pattern might account for it, but frankly I don’t see how it could. I am not aware of any explanation how exactly the Bayer pattern would create the Italian Flag effect. When I tried to simulate the effect of varying angles of incidence, I found the resulting colour shift to be largely symmetrical, even when the Bayer pattern itself was not.

[Advertisement]

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

[UliWer]

Sorry: the example above does not show the results from the M9, but from the M8 with the 2/28 asph.

 

(Got mixed up in my collection...)

[UliWer]

....But this effect would be approximately symmetrical around the optical axis, just like vignetting. I am not aware of any satisfying explanation of the markedly asymmetrical effect we are actually observing. ...

 

I have no valid theory about this most striking effect (the left wing tendency) at all.

 

What I noticed is that you may get results which are very regular under special light conditions: regular artificial light from the top shows no left tendency but - if the edges are underexposed - rather strong and regular edges.

 

I thought at first: it's the sun, stupid! You often have sun from the right, underexposure on the left, so red edges in the underexposed parts. It did not proof to be true, when I tried to make photos with sun from the left: the effect was different, but still there on the left.

 

Though I still believe that the lighting situation of the individual photo is crucial. I think everybody who has seen the red edges has also seen some photos with the same camera and the same lenses where you can't notice any of this phenomenon.

 

I am not sure and don't have the necessary knowledge of optics. Though I think there is some corresponding between strongly polarized light and weekly polarized light and the red edges. If you use a polarizer filter you notice that sometimes it's effect is very strong, when blue sky turns very dark: there is a lot of polarized light. In different situations the filter has almost no effect: low polarized light, so the filter hasn't much to do.

 

Those photos taken with a wide angle lens which typically produce the red shift but show none of the effect in a certain situation are made in lighting conditions of strong polarized light. The effect is strongest when there is not much polarized light. I may be wrong again, and may see some proofs of the contrary again, but until then I assume that there should be some common factor between the two effects.

 

It's not the camera and not the lenses - it's the light.

[ho_co]

... It's not the camera and not the lenses - it's the light.

Hmm. Maybe we need to dust off Goethe's Farbenlehre?

[parsod]

Hoppyman and UliWer thank you for your comments/explanations. I guess you are right about the effect being most obvious in photos that are slightly underexposed which often is the case in snowy conditions (even if I mainly use manual exposure and try to compensate). Maybe this is not such a big issue if I pay more attention to the exposure in these situations. As I mentioned I am very pleased with the results in the majority of shots taken with this equipment so far.

 

/Paer

[thrice]

If the filter/Microlenses/CFA are symmetrical then the problems have to be from software corrections or some kind of internal reflections in the camera. I performed a totally unscientific test and tilted the camera while shining a light into it and having the sensor exposed, it seems like I can see the silver 'edges' of the sensor on the top and right edges at less of an angle than the bottom/left edges (of course remember the projected images is flipped horizontally and vertically - so maybe reflections off this silver edging or the edge of the IR extinction filter). Of course this wasn't very scientific, just a hunch.

[thrice]

Advertisement (gone after registration)

Here is a quick diagram, if the IR filter or edging around it is even slightly assymetrical this could account for the red edge.

 

Thanks to the red wavelength being the least refracted, a really high angle incident ray could possibly have it's longer wavelengths internally reflected, while the shorter wavelengths are refracted normally; either to land on the sensor or go right past it, but the problem is the redder wavelengths would cross over and go further into the image.

 

This is of course a very wild theory cooked up at 1:45am

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!

[parsod]

Since I have only seen this effect in strong sunshine so far I decided to make a test at home in artificial light. When I looked through all my pictures from Portugal I could actually see a red left side in one picture which was also taken in strong sunshine at the coast and that might be similar conditions to my snow landscapes from this weekend.

 

I took two photos of a white wall with my Elmar 24 at f/5.6. The fist one is "over exposed" 2EV to place the white wall in zone VII which in fact should be a fairly correct exposure. I did nothing with this one in LR and I think the red edge effect is quite mild in this example. Yes there is some vignetting and some might come from the fact that the wall was not so evenly illuminated. Maybe there is more tendency of red on the left side but I also see it in the right top corner.

 

Then I took a second photo exposed after the meter which means the wall will be in zone V. I have the lifted the exposure 2EV in LR and by that placing the wall in zone VII also in this picture. I don't think there is a very large difference between these two examples even though I brighten the second one quite substantially in PP. Compared to the two following photos taken outside in a snow landscape I think the red edge effect is less obvious under artificial light even if under exposed and lifted in PP.

 

I don't know if using a filter under these light conditions could reduce this red edge 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!

[charles-k]

I have a M9 with similar issues. Any FL 35 and less, the red edge does show up when you have a light, semi over exposure and then PP with some saturation. In speaking to an authorised Leica repairer, he suggested that there is a quite a wide variation in CCD sensors. Yes the in camera software should correct, but it is very difficult. The camera could be sent back to Solms, and recalibrated, but yes the edge dissappears, but the rendering may not be the same either.

 

Best to incorporate in the workflow, either with corner fix or Capture One, take an exposure with an expodisc and have the correction in the PP workflow. Only takes 2 seconds. Another method is to use ACR graduated filter, hue set to 203, and adjust the level to remove the red edge. Again very quick.

 

My feeling the red edge problem will not disappear in the short term with CCD sensors, so it is relatively easy fix to incoporate in the PP work flow.

[viramati]

Corner fix is a useful tool but for me the problem is that you can't use it from within LR3. It would be great (probably will never happen) if leica could create a plug in in conjunction with Adobe so that we can work on the image from within LR3 to correct the issue, After all LR3 is provided now with leica cameras so they must have a working partnership and if cornerfix has been developed I don't see that it should be so hard for adobe to do the same

[mjh]

In speaking to an authorised Leica repairer, he suggested that there is a quite a wide variation in CCD sensors.

A wide variation of what, I wonder.

[charles-k]

A wide variation of what, I wonder.

 

CCD sensors are quite difficult to manufacture, and the raw output of the sensor do vary a lot, but yield the sought after rendering. Leica could have done like a lot of manufacturers and heavily filtered all the RAW data, but they have taken a minimilist approach. However I do believe there should be more available software to allow PP options.

[mjh]

CCD sensors are quite difficult to manufacture, and the raw output of the sensor do vary a lot, but yield the sought after rendering.

The output of CMOS sensors varies even more, for that matter. But the real issue is: How could this variation explain the red edges phenomenon? Frankly I don’t see how it could. Nobody has offered an explanation so far, rather than just pointing fingers at something or other. “It’s the CCD!”, “It’s the Bayer pattern!”, “It’s the microlenses!” … these are not what I would call an explanation.

[tobey bilek]

I hope they notify me when the problem is gone. I may reconsider a purchase.

 

I have purchased 4 Nikons in the past four years, D40 to D3, and the only problem I had was there was a fleck of black paint floating around on top of the focus screen of the D40.

I repaired it myself.

 

I am still sitting on numerious film bodies and several dozen lenses I wish I could use on a digi body. And not a $25,000 camera with one lens and a truncated system that I would not spend $100,000 + on that is even to big to be portable. I wouold take M or R, don`t care. The toy Leicas like the Delux XX whatever are not considered.

 

Someone else suggested it first, but I believe the fix for now is a preset in PS or LR with a grad filter.

[jaapv]

Seems a bit like overreacting to me. In actual use the problem is minimal to nonexistent, except for one or two Leica lenses which may need some postprocessing in certain light. And of course some third-party lenses, but Leica cannot be held responsible for that.

[bocaburger]

Seems a bit like overreacting. In actual use the problem is minimal to nonexistent, except for one or two Leica lenses which may need some postprocessing in certain light.

 

+1. My pre-ASPH 21 Elmarit could benefit from some finessing of the firmware correction, but the rest are very well corrected. Third-party lenses are another story altogether. I haven't found a single instance where a substitute coding works for all ISO values, so I shoot them uncoded and use Cornerfix as necessary. I can't in good conscience be upset with Leica for not spending money to R&D firmware for someone else's lenses.

[UliWer]

Well, I should not call the problem nonexistant nor minimal when I look at the example of snow scenes in #28.

 

I have seen similar photos with the problem in this forum with the Wate and the 2.8/21 Asph. I have many examples with the problem taken with the 3.8/18 Super-Elmar and a few with the 2/28Asph; I have seen it on one photo taken myself in a snow scene with the 1.4/35 Asph (first version) with the M8.-

 

The Leica representatives at the Forum meeting during the Photokina said, the problem exists; the notes for the last firmware update dealt with the problem, even if it was not solved.

 

Leica's policy is to get the best photos with their camera and lenses. So it must be their policy to solve this problem. I can only wish them success in going further towards this aim. But I don't think it will help Leica or their customers to say the problem is nonexistant or minimal or limited to third party lenses.

[sandymc]

Corner fix is a useful tool but for me the problem is that you can't use it from within LR3. It would be great (probably will never happen) if leica could create a plug in in conjunction with Adobe so that we can work on the image from within LR3 to correct the issue, After all LR3 is provided now with leica cameras so they must have a working partnership and if cornerfix has been developed I don't see that it should be so hard for adobe to do the same

 

Right now, Lightroom just doesn't have the "hooks" to allow raw data to be manipulated, and the lens profiles/DNG camera profiles just don't quite have the capabilities built in to correct red edges. I have suggested some (minor) changes to what is done now in LR, but Adobe haven't come to the party....

 

Sandy

[Advertisement]

Hallo,

Look here: Leica M

[thighslapper]

Seems a bit like overreacting to me. In actual use the problem is minimal to nonexistent, except for one or two Leica lenses which may need some postprocessing in certain light. And of course some third-party lenses, but Leica cannot be held responsible for that.

 

I have to agree

 

I use the 3.8/18 and 2/28 an awful lot and looking through the last 5000 exposures with both my current and previous m9 I cannot find a single example of red edging

Precious little vignetting of any noticeable consequence either.

 

I suppose if I were a professional photographer and lived in antarctica it might be a issue.....

 

I have to admit I'm fascinated by the possible explanations but in reality I'm not sure its the 'deal breaking' issue its made out to be.

 

Everybody seems happy enough to crop, sharpen, alter contrast, white balance and highlights/shadows to make up for the deficiencies in ANY form of photography, so adding another minor infrequent 'aberration' to the list isn't a big deal.

[atufte]

Seems a bit like overreacting to me. In actual use the problem is minimal to nonexistent, except for one or two Leica lenses which may need some postprocessing in certain light. And of course some third-party lenses, but Leica cannot be held responsible for that.

 

Try taking pictures of snow with anything wider than 50 with the light source (the sun) from one of the sides, even the newest Leica lenses have this problem, actually my 35 Biogon has less red corners than my 35 cron, and i also thought this was a camera problem, but now i have tried several M9's, and several different lenses and all of them show similar red corners, which i did not see before the snow landed... So this is not overreacting, this is BAD, really BAD...it almost makes the M9 a B&W only camera in environments like this, which is really far out...

 

This is with a new 35 cron, today...

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!