Leica M11 Review by Jono Slack

[adan]

Advertisement (gone after registration)

53 minutes ago, elmars said:

As good as the 35mm Summilux FLE is, it is prone to purple fringing.

Yep - this forum has had lots of "what's this purple edge with my 35 Summilux" over the years - including the pre-FLE - and long before the M11.

Combination of some inherent (but often invisible) CA - high lens contrast - and blowing out a sky or specular highlight!

Why more on the M11? Because (as tailwagger says) a 0.1mm fringe will spread over 20% more M11 pixels than with an M10R, and 50% more pixels than with a 24 Mpixel M.

A better sensor will improve any lens - but it will also reveal things that are too small to notice as easily at a lower resolution.

Now, the microlenses on any sensor can also introduce their own chromatic aberration - they are, after all, still lenses bending light. We'll probably need to see some more "controlled" experiments comparing the M11 with "camera x" - to sort that out.

Edited January 29, 2022 by adan

[tom0511]

I have also seen more purplefringing with the M11 compared to M10r, for example with the 28 Summilux.

 

[Advertisement]

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

[Chaemono]

This has been known ever since the SL2. 75 Noctilux on SL2 wide open quite a bit of green and purple fringing, same lens on M10-P, none. I may show the 75 Noctilux on SL2-S and on the Z7 (don’t have the SL2 anymore). That’s why fast M lenses are best used wide open on the M10/P, on the SL2-S, perhaps on the S5 (there’s some color shift in corners with that, though), or the upcoming (in 2023, I wish) M11-S. 

[FrankX]

vor einer Stunde schrieb tom0511:

I have also seen more purplefringing with the M11 compared to M10r, for example with the 28 Summilux.

 

Do you see it also more pronounced at 36 MP? Would be interesting to answer if it’s a sensor or resolution effect…

[Tailwagger]

4 hours ago, FrankX said:

Do you see it also more pronounced at 36 MP? Would be interesting to answer if it’s a sensor or resolution effect…

I cant say, as I haven't shot a 36Mpx, however, I did do a day at 18Mpx (accidentally) and I would say in reviewing those shots that, as you'd expect, fringing is less visible. I think it would require a bit of side by side testing with the 10-R to say anything definitive as to whether or not the sensor, other then MPx, is having any effect. But I think the more interesting question you raise is the effect of pixel binning itself. One could imagine baking in some intelligence into the algorithm to detect and de-fringe as a part of the compression process, if not today, perhaps in a future update. 

[mmradman]

8 hours ago, ELAN said:

I've also been seeing an increase in purple fringing wide open.  Here's a 100% crop from a 35 lux (pre-FLE) wide open, SOOC, rising sun at my back.  This lens never fringes this much on the M10-P (the 28 lux does get close).

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!

Non issue, who shoots into the daylight sky fast lens at maximum aperture other than to pick nits?

I have seen same purple fringing on Summilux 21, 28 and 35 FLE used with M240 which is, reminder, only 24Mp. Perhaps, as mentioned by others, effect enhanced on 60Mp M11 but it is principally lens thing.

[ELAN]

Advertisement (gone after registration)

1 hour ago, mmradman said:

Non issue, who shoots into the daylight sky fast lens at maximum aperture other than to pick nits?

I do.  And it's an issue for me.  As stated, the sun is just rising behind me, hardly daylight sky.  Click twice on that cropped image for full res and observe the purple fringing. I've never seen so much CA from my Summiluxes before, including the 28 lux.  Here's the unedited full image, not really a shot into daylight sky.

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!

Edited January 29, 2022 by ELAN

[tom0511]

vor 6 Stunden schrieb FrankX:

Do you see it also more pronounced at 36 MP? Would be interesting to answer if it’s a sensor or resolution effect…

I haven't checked at 36MP. By the way - in LR it was easy to get rid of the purple fringing. So I don't expect it to be a problem. But the M10r seems to do better, or maybe its better corrected in software.

Edited January 29, 2022 by tom0511

[mmradman]

1 hour ago, ELAN said:

I do.  And it's an issue for me.  As stated, the sun is just rising behind me, hardly daylight sky.  Click twice on that cropped image for full res and observe the purple fringing. I've never seen so much CA from my Summiluxes before, including the 28 lux.  Here's the unedited full image, not really a shot into daylight sky.

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!

Premium Helga 😂.

It seems we have “flawed” yet premium priced optics in our hands.  Remedy, close down or use sliders in ACR. Perhaps Leica could include software correction, or maybe trying to incentivise us to use L system with APO prime lenses.  It would be interesting what Leica Chief Optical Designer would have to say, or Beta Tester in Chief.

Edited January 29, 2022 by mmradman

[adan]

5 hours ago, Tailwagger said:

One could imagine baking in some intelligence into the algorithm to detect and de-fringe as a part of the compression process,

But how does the CPU "know" if the purple is a fringe - or just an out-of-focus bathing-suit in the background?

I've watched Camera Raw's de-fringe controls at work - and they have a nasty habit of eating into the edges of things that are actually purple - or red - or green - unless carefully supervised.

At this point I'm even careful about which lenses I buy - do they produce red fringes (90mm Summicron v.3, 90 Summarits) or purple fringes (90 Elmarit-M). Because there are a lot more "primary red" objects in the world than "primary purple/magenta" - so it is easier to remove purple fringes with  "dumb automation" (if that's not a tautology - ) without nibbling into the actual colors.

[adan]

4 hours ago, ELAN said:

I do.  And it's an issue for me.  As stated, the sun is just rising behind me, hardly daylight sky.  Click twice on that cropped image for full res and observe the purple fringing. I've never seen so much CA from my Summiluxes before, including the 28 lux.  Here's the unedited full image, not really a shot into daylight sky.

A blown-out sky is a blown-out sky.

Doesn't matter if it is shot at dawn, or in noon daylight, or midnight. If it is virtually white in the picture (but was not virtually white in real life) - it has been grossly overexposed, and "blown."

And that will exaggerate whatever fringing is present.

I took measurements from your image - your sky is pure white (250+ brightness level) in the places that the worst purple is present. Here's the threshold mask - everything pure white here was pure white ("blown out") in your picture.

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!

The M11 may have very good dynamic range, but it is not infinite. If the sky is in your image - meter with the sky in mind, or expect to have problems.

Basic rule of digital imaging - "Expose for the highlights; post-process for the shadows."

[mmradman]

Before M8 and M9 Leica were adamant that digital RF would not be feasible mostly due to light ray angle hitting the sensor surface, in the end they delivered miracle, with few bumps along the way, like IR contamination, corroding sensor, etc.

Leica also developed dedicated Monochrom RF digital cameras which are very good with pesky colour fringing.

[jonoslack]

5 hours ago, ELAN said:

I do.  And it's an issue for me.  As stated, the sun is just rising behind me, hardly daylight sky.  Click twice on that cropped image for full res and observe the purple fringing. I've never seen so much CA from my Summiluxes before, including the 28 lux.  Here's the unedited full image, not really a shot into daylight sky.

 

 

32 minutes ago, adan said:

A blown-out sky is a blown-out sky.

Doesn't matter if it is shot at dawn, or in noon daylight, or midnight. If it is virtually white in the picture (but was not virtually white in real life) - it has been grossly overexposed, and "blown."

And that will exaggerate whatever fringing is present.

I took measurements from your image - your sky is pure white (250+ brightness level) in the places that the worst purple is present. Here's the threshold mask - everything pure white here was pure white ("blown out") in your picture.

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!

The M11 may have very good dynamic range, but it is not infinite. If the sky is in your image - meter with the sky in mind, or expect to have problems.

Basic rule of digital imaging - "Expose for the highlights; post-process for the shadows."

Hi There Elan

I'm absolutely with Andy here - overexposing the sky makes a perfect storm of purple . . . . but either way, CA is a product of the lens, not of the sensor, although of course if you have more pixels you'll have more purple pixels! 

It certainly isn't something I've had any extra problem with using the M11 (even with some very old lenses). 

As Andy says (and puts it so nicely)
 

"Expose for the highlights; post-process for the shadows."

 

all the best

[Tailwagger]

44 minutes ago, adan said:

But how does the CPU "know" if the purple is a fringe - or just an out-of-focus bathing-suit in the background?

I see my short treatise on ML was lost on you. No worries, I didn't really get it either. 😃  That said, I'd bet that detection and removal of fringing is a good candidate for ML (but likely not in camera).

Conventional or otherwise, it would have to be noticing edges with thin lines of purple pixels abutting them and rejecting areas that didn't fit that pattern. But I suspect it would probably require more HP than it's worth.  As an aside related to purple bathings suits, one of the annoyances I have with Adobe and camera raw is that the full suite of defringing tools is not available when using filters. When I mentioned I was finding fringe removal a little easier, it was exactly this problem of eating into other areas you mentioned I was channeling. 

 

[ELAN]

8 hours ago, adan said:

A blown-out sky is a blown-out sky.

But the sky isn't blown out.  Here's the image processed with exposure at -1, including the histogram, so you can see than not a single pixel is clipped!

I understand about purple fringing and have been removing it in LR for ages.  Especially with the 28 lux that tends to fringe more than others.  And I understand too that 60 MP would amplify all flaws in the lens compared to 24 MP.  I am simply confirming others' observations that purple fringing appears more pronounced with the new sensor, even without pixel peeping.

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!

 

 

Edited January 30, 2022 by ELAN

[M Journey]

23 hours ago, ELAN said:

I've also been seeing an increase in purple fringing wide open.  Here's a 100% crop from a 35 lux (pre-FLE) wide open, SOOC, rising sun at my back.  This lens never fringes this much on the M10-P (the 28 lux does get close).

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!

I’ve done test shots like this where I am shifting into the sun peeking behind branches and leaves and have seen zero fringing.  I wonder if this is more prone on older lenses?  The m11 has thinner glass on top of the sensor and I wonder if this optimizes newer lens at the cost of performance on older lense designs?

[SrMi]

44 minutes ago, ELAN said:

But the sky isn't blown out.  Here's the image processed with exposure at -1, including the histogram, so you can see than not a single pixel is clipped!

I understand about purple fringing and have been removing it in LR for ages.  Especially with the 28 lux that tends to fringe more than others.  And I understand too that 60 MP would amplify all flaws in the lens compared to 24 MP.  I am simply confirming others' observations that purple fringing appears more pronounced with the new sensor, even without pixel peeping.

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!

 

 

Lightroom does not tell you reliably if highlights are clipped.

[oka]

...I noticed this problem noticeably on Noct50 (with M11). I haven't done any test so this might be good next comparison test to compare SL2S, M10R and M11 for purple fringe.

[adan]

2 hours ago, ELAN said:

But the sky isn't blown out.  Here's the image processed with exposure at -1,

If you have to drop the exposure in post-processing a full stop - and still get grays instead of blues in your sky gradient... (e.g. between the two tree trunks)

- then the camera exposure was clipped and the sky was blown out in those places. The gray is the image telling you "I don't have enough usable data here!"

I learned that all the way back to the M9 in 2009. Correct exposure for digital highlights is something you do taking the picture, not trying to rescue it later.

..............

Especially with the newer Ms  with improved dynamic range. The extra range is mostly in the shadows.

They have a lot of shadow detail that can be pulled out (using the Shadows slider instead of the Exposure slider). Which is hidden SOOC due to a strong default contrast curve.

M10 images below - SOOC (top), and with all the available shadow information revealed (bottom). Maybe even more than some would want (it looks a little HDR-ish) - but at least it is in there for the taking.

So again, expose in the camera for the highlights, and post-process for the shadows.

Thus protecting the areas (to get back on-topic) where the purple fringes may raise their nasty heads.

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!

Edited January 30, 2022 by adan

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Chaemono]

vor 35 Minuten schrieb adan:

If you have to drop the exposure in post-processing a full stop - and still get grays instead of blues in your sky gradient... (e.g. between the two tree trunks)

- then the camera exposure was clipped and the sky was blown out in those places. The gray is the image telling you "I don't have enough usable data here!"

I learned that all the way back to the M9 in 2009. Correct exposure for digital highlights is something you do taking the picture, not trying to rescue it later.

..............

Especially with the newer Ms  with improved dynamic range. The extra range is mostly in the shadows.

They have a lot of shadow detail that can be pulled out (using the Shadows slider instead of the Exposure slider). Which is hidden SOOC due to a strong default contrast curve.

M10 images below - SOOC (top), and with all the available shadow information revealed (bottom). Maybe even more than some would want (it looks a little HDR-ish) - but at least it is in there for the taking.

So again, expose in the camera for the highlights, and post-process for the shadows.

Thus protecting the areas (to get back on-topic) where the purple fringes may raise their nasty heads.

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!

I remember this post when the M10 first came out. Ever since I expose similarly. There’s is perhaps an understandable temptation with new cameras with better and better FF sensors and higher and higher MP count that according to reviews have greater and greater highlight detail recoverability to push the limits of exposure to the right in order avoid shadow noise. I remember this was an argument with the M10-R as well. Personally, I would refrain from doing it and would continue to expose just like with the M10. 

Among the last five here: https://www.smugmug.com/gallery/n-JkgxSt/, there are four pictures (two are crops of the other two) where the S3 was exposed to protect highlights and pushed to the permissible limit to the right. The crops show that the difference in the noise level is negligible and one must zoom in a lot to discern it.