M10M gate coming?

[jplomley]

Advertisement (gone after registration)

16 minutes ago, tgray said:

First I've seen of this topic.  FWIW, I figured if my camera had this issue, I had one photo in mind that I took on a day with clear blue sky and a 21mm lens with an orange filter.  Sure enough, run the Dehaze filter up a bit and there it was in the sky.  I could find it on a bunch of other photos as well.  Dehaze really seems to bring it out and it mostly shows up on areas of smooth transitions (sky, walls, etc.).  A little bit of the Grain effect completely wipes it out, as does higher ISO shots.  One other thing I noticed is it seems to be mostly contained in the upper third of the image, though that could just be a result of composition and subject matter.  

Any ideas from the forum on what causes it? Are we dealing with two sandwiched 20 MP sensors and we’re observing the seam? Saw something like this on the original S2. 

[adan]

5 minutes ago, jplomley said:

Any ideas from the forum on what causes it? Are we dealing with two sandwiched 20 MP sensors and we’re observing the seam? Saw something like this on the original S2. 

Did you read the very first page of the thread? Did you notice post #2?

[Gobert]

What I noticed is that many of people with problems use orange or red filters. I use mostly yellow from a prime brand. Could that have caused it?

[jplomley]

I’m using a Leica orange filter (and green where applicable). Only one stop loss of light. If the line is caused by use of color contrast filters, then that’s a major concern for B&W photography that includes inclusion of sky. I will have to test this out. 

[jplomley]

2 hours ago, adan said:

Did you read the very first page of the thread? Did you notice post #2?

Thanks Adan. I actually found this thread when searching for "vertical line". Took me to post #48! I've now read through the entirety of the postings. Somewhat shocking the effect is not corrected by Leica, somehow. 

[radialMelt]

I am currently in possession of an M10M and fortunately have not noticed this issue either, even when pushing the files to the max (e.g., cranking up dehaze in LR)

Knock on wood

[Keith (M)]

Advertisement (gone after registration)

In almost fifteen months of M10M ownership, used with lens from 21 to 135mm and frequently with either yellow or orange filters, can't say as I have ever seen any unusual lines, artefacts, whatever.  My only 'problem' with the camera is that it is so seductive that the other bodies (digital and film) are feeling very neglected!

[jplomley]

2 hours ago, Keith (M) said:

In almost fifteen months of M10M ownership, used with lens from 21 to 135mm and frequently with either yellow or orange filters, can't say as I have ever seen any unusual lines, artefacts, whatever.  My only 'problem' with the camera is that it is so seductive that the other bodies (digital and film) are feeling very neglected!

Hmm, in my one month of trying two M10M’s, both have had issues. Agree the files are incredible, but wondering if the QC issues are pandemic related? Whether that be a defocussed labour force involved in final test, or perhaps an issue in parts supply/manufacturing. Hard to really know. 

[Rob L]

Seems odd that the lines in your photos are not vertical as would be indicated by the sensor seam theory. Did you crop and straighten the photos to correct some tilt?.

[tgray]

4 hours ago, Keith (M) said:

In almost fifteen months of M10M ownership, used with lens from 21 to 135mm and frequently with either yellow or orange filters, can't say as I have ever seen any unusual lines, artefacts, whatever.  My only 'problem' with the camera is that it is so seductive that the other bodies (digital and film) are feeling very neglected!

To be fair, I've had my M10M since release and never noticed this issue either until I went looking for it.  I'm not sure if I'm actually bothered by it all other than as a matter of principle.  It only shows up with in particular situations with particular processing.  Fortunately I never process with that processing – the grain effect, which I often use, seems to make the issue not appear even with significant Dehaze, which I also never use.

ANY texture in the image seems to mitigate the effect, i.e. I only really see it in skies or similar backgrounds.  Presumably this is because the Dehaze tool is some variation of local contrast, and any scene variation swamps the sensor 'defect'.  Looking through some photos right now, I see it in one shot in the sky, but not the picture taken literally 30 seconds earlier with the same lens, ISO, etc., with sky in the image.  

I plotted up some horizontal slices through the image.  It's very hard to find the artifact on a given row, but averaging through several rows (20-30) begins to show the issue.  Happens right at pixel 3929/3930 for me.  Oh well!

[RMF]

2 hours ago, tgray said:

To be fair, I've had my M10M since release and never noticed this issue either until I went looking for it.  I'm not sure if I'm actually bothered by it all other than as a matter of principle.  It only shows up with in particular situations with particular processing.  Fortunately I never process with that processing – the grain effect, which I often use, seems to make the issue not appear even with significant Dehaze, which I also never use.

ANY texture in the image seems to mitigate the effect, i.e. I only really see it in skies or similar backgrounds.  Presumably this is because the Dehaze tool is some variation of local contrast, and any scene variation swamps the sensor 'defect'.  Looking through some photos right now, I see it in one shot in the sky, but not the picture taken literally 30 seconds earlier with the same lens, ISO, etc., with sky in the image.  

I plotted up some horizontal slices through the image.  It's very hard to find the artifact on a given row, but averaging through several rows (20-30) begins to show the issue.  Happens right at pixel 3929/3930 for me.  Oh well!

I think you are right @tgray. It certainly seems to only makes itself sometimes present in certain conditions with dehaze in the +50-100 range, with perhaps some correlation with underexposure in the area. It might just be that LR's Dehaze algorithm just doesn't play nicely with M10M files under these conditions. Has anyone seen this issue in Capture One's Dehaze?  

For reference if you shoot a smooth blank wall or a gradated lit wall, it might make itself apparent if you really look- push the LR dehaze to 100 and play with the exposure slider. Monitor quality might make the issue more or less apparent.

 

 

Edited May 1, 2021 by RMF

[jplomley]

So is this a fundamental design flaw? Some people report it, others do not, so appears not to be universal? Once I hear from Leica, I’ll feed back the response on how best to proceed. Not sure insisting on a replacement body will correct the issue. Firmware fix?

Edited May 1, 2021 by jplomley

[adan]

1) Here's a little item regarding your "waffle pattern." Happens to be with the Q2 Monochrom - not sure of the relationship between the M10M and Q2M electronics (same manufacturer?) Anyway, a waffle-like pattern in shadows (actually, overlapping moires) - with 2-stop attempts at shadow recovery. You may need to really zoom in on the picture and look in the subject's hair, and the extreme shadows in the bridge structure.

2) Regarding the dividing line in skies, Leica obviously has some issue, from camera to camera, with balancing the left and right outputs of the one-piece M10M sensor. That appears to be an electronic problem in the multiple chips that take the flow of voltage off the sensor and turn it into a picture. It may be something that can be "tuned out" (just like adjusting the left/right balance of a stereo sound system) - or it may be a faulty downstream chip that needs replacing (like a faulty pre-amp unit in a sound system).

From reading around, Leica will address either of those problems under warranty - adjust the balance settings, or replace some part of the electronics, depending on what problem they find. As well they should.

And, yes, it can very well be a variable problem, present in some cameras more or less so than in others (or not at all). Always keep in mind that digital photography depends on quantum physics. The interactions of sub-atomic particles bouncing around in the silicon - and the home of the Uncertainty Principle. About all one can predict is a probability of photon X producing electron Z, and electron Z eventually getting converted into a bit value of 00000000 ••000001 (M10M produces 14-bit data). Usually the probabilites average out to reasonable consistency.

The real question is, how consistent? Out of, perhaps, 9635 M10-Ms sold so far - how many actually show this problem? Straight from the camera. Or under major tonal manipulation. How does that compare with the other 40Mp, 24x36 monochrome cameras out there?

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

Now for the tough-love:

3) Leica does not test every camera/sensor-electronics set out of the gate with every possible raw-processing software, with their settings all turned up to "11," or with the exposure grossly off-base - just to see what happens. They (or their sensor maker) test random samples, during sensor development, and during production. And their pre-production beta-testers (usually experienced photographers) can push the limits however they choose.

https://en.wikipedia.org/wiki/Category:Raw_image_processing_software

Of course, there may be a conceptual problem with that - to make something "foolproof," one has to give it to fools to test.

Experienced photographers are generally used to getting the picture correct "in the camera" as much as possible, and counting very little on extreme processing to hype up a weak image of a weak situation. So they may never touch the "De-Haze" slider.

4) "But....but... if Leica doesn't test to perfection, why am I paying C$11500??" Well, because a Leica M-A already costs C$6800 - with no electronics at all. That leaves C$4700 (of which Leica probably gets C$3500 after dealer markups) to commission and buy the bespoke electronics and test them to a reasonable extent ("Does the camera record what is in front of the lens with no artifacts, using correct camera technique and minimal post-processing?"

5) I have a fairly simple way of dehazing pictures. If I don't want haze, I don't photograph on hazy days. I, too, did a 750km  documentary "road project" over the past 4 years, and frankly there were places I went back to several times, until the light and weather were "right." Getting good photographs is about spending time and effort - not saving it.

I think you are counting far too much on "sledge-hammer" post-processing to try and turn mediocre light into something different. Your crop in post #72 seems to show outrageous sharpening halos - although that may also be too high a de-hazing setting (see very last link below)

I'll admit to being lucky in living at an elevation of 1609m/5280 feet (and shooting at up to 4300m in our Rockies). My "normal clear day" even in Denver is visibility 100km, with skies as shown (accurately) below (plain vanilla M10 (no filtering), and a B&W conversion).

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!

But sometimes even we get days like this (M10, 135 lens, Rocky Mountain foothills, 70km downwind from Colorado's 2nd largest wildfire on record). When life gives you haze - err, lemons - make lemonade. Even a wildfire 1500 km upwind can produce haze - if you don't want it, wait for another day.

6) Don't blame Leica for Adobe's products.

I did a little research, and found two sources regarding M10M artifacts. This one, and the original digilloyd site.

I found double that number of online comments about issues with the Adobe De-Haze processing itself.

It does a specific thing well (rescue some pictures shot in poor conditions - for a given value of "rescue" and for a given level of photographer), but it is not a magic bullet. Even Adobe's mavens call it "extreme post-processing." It's a sledge-hammer - it can produce dented pictures from any camera.

https://community.adobe.com/t5/lightroom-classic/dehaze-artifacts/m-p/9239541

https://feedback.photoshop.com/conversations/camera-raw-and-dng/camera-rawlightroom-dehaze-altering-the-colorbalancewhitebalance-of-photos/5f5f461d4b561a3d4274337b

https://fstoppers.com/lightroom/good-bad-and-ugly-lightrooms-dehaze-tool-330394

https://www.dennystips.com/photoshop-dehaze-trick/

[tgray]

Thanks for the lecture of how we should post process our photos.  For what it's worth, I can see the line in one photo with zero adjusts if I look hard, and only a relatively minor exposure adjustment makes it even clearer now that I know it's there.  No dehaze or anything.

Anyway, I'm out.

 

 

[jplomley]

Update: Leica is "willing" to replace the sensor. How nice they are "willing" to do this on a new camera, as if its a privilege. I'm not even convinced the sensor is the issue. The most annoying aspect is the waffle pattern, its a royal PITA to have to correct this every time I need to straighten an image. I had zero issues with the two M10's I owned and the M10-R I currently own, so I don't quite understand why the Monochrom is so plagued with issues. As I need the camera for a number of projects, I will send it in during the down-times of Nov/Dec. I'll update the post again once the camera has been serviced.  

[SoarFM]

Just so happened to happen upon this article...

 

"It’s still sometimes possible to make patterns appear in the noise floor of an M10 Monochrom file when you rotate or geometrically distort an image.  Anyone who’s seen this with either of the earlier Monochroms will continue to see it with the new M10 version. The extent to which these patterns are visible has always been dependent on a RAW converter’s interpolation algorithms. ...It’s a rather minor problem with many workarounds, "

https://www.ultrasomething.com/2020/01/paradox-view-the-m10-monochrom/

 
 Whether this purported problem is really a showstopper seems dependent on how you use your M10M.

My take is that Lloyd Chambers is making a mountain out of a molehill. That's his business model. It's up to the enduser to decide. A lot of wonderful photographs have been made with equipment that Mr. Chambers finds fatally and irredeemably flawed.

 

[Malcolm Kays]

Just checked a large selection of M10M images and no sign of the problems you have, even on shots with clear skies.

 

[jplomley]

On 5/27/2021 at 3:27 AM, Malcolm Kays said:

Just checked a large selection of M10M images and no sign of the problems you have, even on shots with clear skies.

 

Congrats Malcolm. Unfortunately, mine has now become about as useful as a door stop. I don't trust it to give me an image without artifact, so i have stopped using it completely and gone back to converting my M10-R images to B&W. Zero problems with that camera, regardless of the image manipulations in either LR or PS. No doubt about it, I got a serious lemon.

[jplomley]

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 latest issue....full image

 

[Advertisement]

Hallo,

Look here: Leica M

[jplomley]

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!

 

Right hand side, lifting the shadows ever so slightly....and voila, a nice dark line down the entirety of the image. Most expensive door stop I've ever purchased :-(

Edited May 28, 2021 by jplomley