Leica M11 -purplish tint ???

[Velo-city]

Advertisement (gone after registration)

7 minutes ago, LocalHero1953 said:

I dip into this thread every five pages. Nothing changes, neither those posting nor arguments - just groundhog day.

I dont think there's anything that can change really, not in the short term.

- Some people find the white balance on their M11's off - and particularly magenta-shifted under certain conditions - and that's unacceptable and frustrating to them.

- Others don't experience it, or don't see it, or just dislike anyone complaining about it.  A few will seemingly defend it to death, blame the users (bl**dy amateurs), push workarounds, and generally try and squash dissent.

Chances are Leica won't address it until the next numbered model, just my guess. Despite finding it annoying, frustrating and disappointing, in practice there are other quirks of the M11 that are equally if not more annoying....

Meanwhile, reading some of the justifications is quite amusing it has to be said, so I'll check back just to see what else appears.

 

[Advertisement]

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

[pedaes]

1 minute ago, junix said:

test it myself

Thats the way to go. You don't know what funny habits others have.

 

[Lee S]

57 minutes ago, CDodkin said:

You need to look at the whole image pipeline - trying to guess which element may be contributing to the issue doesn't really get you anywhere. The pipeline includes the following key components which will impact WB:

Sensor Sensitivity: The M11's sensor might have a slightly higher sensitivity to red wavelengths compared to green and blue, leading to a magenta tint in images.
Color Filter Array (CFA) Pattern: The specific CFA pattern used in the M11 could influence the color balance, especially under certain lighting conditions.
Calibration Issues: Potential calibration inaccuracies in the camera's white balance algorithm could contribute to the magenta shift.

White Balance Algorithm: The white balance algorithm used by the M11 might have a tendency to shift towards magenta, especially in daylight scenarios.

It's been interesting to watch Fuji, and now Leica, grapple with the same daylight WB issue, both since switching to Sony sensors in their cameras. The length of time this has been going on (7+ years) would suggest that it's not a trivial problem to solve for the camera manufacturer!

Thanks for taking the time to respond in depth. Didn’t realise it was affecting other manufacturers. 

They have got the WB/Tint right on the rest of their 60MP line which I expect uses the same silicon with different cover glass and micro lens design optimised for steep ray angles.

Regards,

Lee  

[CDodkin]

Automatic White Balancing Algorithms

Automatic White Balancing (AWB) is the technique used in all digital cameras to correct color casts in images caused by different lighting conditions. It aims to make the image appear as if it were captured under standard lighting. The algorithm in the M11 likely uses some/all of the following techniques in order to determine what the WB should be in any given image.

 

1. Gray World:

   • Assumes the average color of a scene is gray.
   • Adjusts the red and blue channels to match the green channel's average intensity.
   • Simple and effective but can be inaccurate in scenes with dominant colors.

2. White Patch:

   • Assumes the brightest pixel in the image is white.
   • Adjusts the red and blue channels to match the maximum intensity of the green channel.
   • Sensitive to noise and outliers.

3. Iterative White Balancing:

   • Identifies white or gray pixels in the image.
   • Iteratively adjusts the red and blue channels based on these pixels.
   • More robust to noise and color casts but can be computationally expensive.

4. Illuminant Voting:

   • Models the image formation process using linear combinations of illuminant and reflectance spectra.
   • Uses a voting scheme to identify the most likely illuminant.
   • More accurate but computationally intensive.

5. Color by Correlation:

   • Precomputes a correlation matrix between image colors and different illuminants.
   • Correlates the image colors with the matrix to determine the most likely illuminant.
   • Provides a measure of uncertainty in the illuminant estimation.

These methods are often combined or used in a hierarchical approach to achieve more accurate and robust white balance. The choice of method depends on factors such as computational complexity, accuracy requirements, and the specific characteristics of the camera and imaging conditions.

When presented with a calibrated transmission target, such as the Expodisc, we help solve for 1, as we present only a calibrated gray image. 2, as we remove any bright spot pixels in the image. 3, as all pixels are calibrated gray in the image. 4, as we have a single transmitted illuminant. And 5, again because we have a single illuminant and all neutral colors in the image.

[T I N O]

vor 34 Minuten schrieb hdmesa:

Everyone seems to agree the M11 has a magenta tint bias. We can ask Leica to address the issue

I asked a salesperson at the Leica Berlin store this summer about the magenta issue of the M11, and you could witness in real time how the guy put on his best poker face, while he told me that he never heard of a "magenta issue". 
He also told me that he had not heard about any freezing issues.
Seriously.

My point being that while Leica apparently are dealing with the freezes by offering new firmware versions (to the point where, almost three years after release of the M11, the camera is now widely considered to be "pretty stable" - congratulations) they did not really ever acknowledge that there's an issue with the color / white balance -
so I wouldn't get my hopes up that they will "adress the issue" now if we were to email them - they probably rather just sit it out and (hopefully) do better with the M12 🤷‍♂️

 

 

 

 

[hdmesa]

21 minutes ago, T I N O said:

I asked a salesperson at the Leica Berlin store this summer about the magenta issue of the M11, and you could witness in real time how the guy put on his best poker face, while he told me that he never heard of a "magenta issue". 
He also told me that he had not heard about any freezing issues.
Seriously.

My point being that while Leica apparently are dealing with the freezes by offering new firmware versions (to the point where, almost three years after release of the M11, the camera is now widely considered to be "pretty stable" - congratulations) they did not really ever acknowledge that there's an issue with the color / white balance -
so I wouldn't get my hopes up that they will "adress the issue" now if we were to email them - they probably rather just sit it out and (hopefully) do better with the M12 🤷‍♂️

A Leica Store salesperson versus Leica corporate – very different folks. I've had extremely good luck communicating via that email address:

• We've not seen that issue reported yet, are there specific steps or settings that trigger the issue?
• Thanks for the detailed information. I'll forward it to the engineers to look into it.
• This is a known issue we're working on, but we can't guarantee we'll have a fix by the next firmware update.
• Thank you for the feature request. I will forward it to the appropriate team.

 

Edited November 24, 2024 by hdmesa

[T I N O]

Advertisement (gone after registration)

Wait - you're the guy that actually

Am 19.11.2024 um 22:49 schrieb hdmesa:

owned at least as many M11s as anyone else on this board and from the beginning. Five M11s, one M11M, one M11D. I’ve lost enough money selling/trading them to have paid for one outright.

..and, after all that hassle, you do consider yourself "extremely lucky" with email feedback from Leica as quoted above?
Wow.

No offense intended, it just baffles the mind 😅

[hdmesa]

12 minutes ago, T I N O said:

Wait - you're the guy that actually

..and, after all that hassle, you do consider yourself "extremely lucky" with email feedback from Leica as quoted above?
Wow.

No offense intended, it just baffles the mind 😅

I've had extremely good luck communicating with Leica. I've had extremely bad luck with M11s.

I'd have something to say about your photoshopped (AI?) profile pic, but I don't want to end up saying "no offense" or having worry about baffling any minds. Oh well, I'll say it anyway: you seem very smart (and type very well) for a cat 😂

Edited November 24, 2024 by hdmesa

[T I N O]

vor 20 Minuten schrieb hdmesa:

I've had extremely good luck communicating with Leica. I've had extremely bad luck with M11s.

..fair enough! 😀

 

And yes, the profile pic is just an AI rendering of the prompt "kitten in front of a burning house holding a Leica M6" 🙂

[hdmesa]

24 minutes ago, T I N O said:

..fair enough! 😀

Honestly I wish it had been the other way around: good luck with M11s and bad luck communicating with Leica 🙃 

But I will definitely let others freeze-test the M12 this time around before buying one.

[Smogg]

16 minutes ago, hdmesa said:

But I will definitely let others freeze-test the M12 this time around before buying one.

I hope they have learned lesson and will not make such a global mistake again.

[hdmesa]

44 minutes ago, Smogg said:

I hope they have learned lesson and will not make such a global mistake again.

I thought I remembered reading someone say that the M11 was a complete firmware rewrite. Hopefully the M12 will be iterative, requiring fewer firmware changes, and be more stable and not less stable.

[otto.f]

Would digging out the UV-IR filters for the M8 help in the meantime that Leica is working on that firmware? 

[jaapv]

Most likely not-but it is worth a try. 

[nineteenfocus]

I’ve read several post where the auto white balance isn’t reliable with the M11. Recently, I have had several noticeable failures where the camera set the WB to a very wrong from the ‘normal’ value that the scene require

See below the picture out of the camera ‘original’, it’s a sunset scene and it has set the WB to 4900K where it should be at least 5500K like a daylight. This results the image to be cold and not representative of the scene

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!

Does anyone experiencing same problem? most of the daylight I get from the M11 are cold as the WB is generally automatically set to 4500K. 

Another example auto WB in a clear day at noon, should at least be 5500K but the file was set to 4650K by M11, how can it be so off?

I guess that this could be easily tuned with a firmware update but there’s not been a mention about it. Is the WB a firmware issue or a hardware problem? I am wondering if this might be an issue of my camera

Thanks

 

[Velo-city]

5 hours ago, nineteenfocus said:

I guess that this could be easily tuned with a firmware update but there’s not been a mention about it. Is the WB a firmware issue or a hardware problem? I am wondering if this might be an issue of my camera

I don’t think there’s a consensus. The logical default is to assume a firmware issue - the coding around the white balance and AWB, but it could be a number of other things - like sensor calibration or different batches of certain components.
From what I’ve read here I don’t believe Leica themselves have acknowledged the issue, and many owners don’t experience it (or don’t see it). 
 

I also initially thought it was only AWB, but later realised the set WBs are off too in certain conditions - particularly sunny blue skies when imported into Lightroom have a very strong magenta tint and have offset this with a colour temp that also seems off. 
It’s not consistent which is why sensor calibration alone feels less likely. 
 

I’ve only used a handful of the cameras that have ever existed, but I’ve certainly never experienced anything this off that I can remember.  Also it doesn’t happen on the Q3 - at least not so obviously. That’s why so many people are surprised/disappointed I think - especially when combined with the price tag of the body and the ‘Leica heritage’.

Edited December 1, 2024 by Velo-city

[jgeenen]

vor 6 Stunden schrieb nineteenfocus:

I’ve read several post where the auto white balance isn’t reliable with the M11.

 

I think, the Leica AWB algorithm is probably not very intelligent, but it is absolutely predictable. In your sample pictures, you have a very dominant blue sky that is obviously being ignored by the camera. The remainder of color tones is on the warm side, so the camera bases the WB based on that. So, the decision of the M11 to take the 4900K value is a best guess without sophisticated image analysis capabilities. I compared with pictures in my archive: Blue skies are largely ignored and the WB is based on the parts of the images that have less dominant colors. Same observation with other dominant colors (e.g. green grass, red cars, etc.). 

The purple tint that can be detected in both pictures is a different animal - but that is widely discussed. 

[nineteenfocus]

It would be useful to share temperature Kelvin on the pictures you (users on this forum) get out of the camera for a very sunny day. Surprisingly most of my values are lower than 5000K. Interestingly, some of them Lightroom Auto adjust to 7500K… I don’t know how this is calculated either. For my eye it should be something around 6000K. Anyone can give some light on what’s the algorithm to calculate the WB and how does it affect to the calculation? 

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[don daniel]

vor 14 Stunden schrieb nineteenfocus:

 

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!

 

Yes, the color temperature might be a bit cool here. But that's not so bad. What's not okay: the magenta cast in the image.

[otto.f]

7 hours ago, nineteenfocus said:

It would be useful to share temperature Kelvin on the pictures you (users on this forum) get out of the camera for a very sunny day. Surprisingly most of my values are lower than 5000K. Interestingly, some of them Lightroom Auto adjust to 7500K… I don’t know how this is calculated either. For my eye it should be something around 6000K. Anyone can give some light on what’s the algorithm to calculate the WB and how does it affect to the calculation? 

I think you’re on the wrong track. In the M9 the WB was also under 5000 too, also with images which were quite neutral and realistic. Besides, these figures are not absolute and not consistent over different app’s. If the right Kelvin is 5200 in LR, it’s not necessarily in C1, etc. The sunset colors you want to have in your example are easy to get, as we see in the ‘last’ image. The magenta problem however can’t be solved in many cases and not via Kelvin or Tint settings.