41 vs 24 MP

[onasj]

Advertisement (gone after registration)

I’ve noticed from the discussion on this forum that some seem to think that shooting 41 MP will be much harder to focus/capture/process than 24 MP.  I just don’t think that’s the case based on extensive shooting of 24 MP (M10, CL even with M lenses, M10P), 42 MP (a7riii, even with IBIS off and using M lenses), 61 MP (a7riv), 100 MP (IQ3 back with no IBIS and tech cam MF lenses), and 151 MP (IQ4, also with no IBIS and MF lenses).  

Yes, more MP means it’s harder to get everything sharp at the per-pixel level. But more MP does not make it more difficult to retain sharpness at the per-image-height level (which is what really matters unless your clients’ walls and screens suddenly also grow in size), and only as the square root of the resolution.  The 41 MP M10 Monochrom only has 1.3x the number of pixels in any linear dimension than the M10.  

I doubt shooters will really find it significantly harder to focus or capture in practice than shooting the M10.  Even when shooting the 151 MP IQ4 handheld with a MF tech cam lens I can still capture sharp photos (virtually all of my photos here were captured this way: https://www.alpa.ch/en/photographers/liu-david-r- ). I just had to learn to not obsess with pixel peeping... admittedly difficult for a technically oriented guy!

Indeed, I anticipate that the two stops of better ISO performance and thus much shorter shutter times will improve sharpness under challenging conditions much more than 1.3x more pixels in any linear dimension will impede sharpness.

[shirubadanieru]

@onasj I’d just commented the below on the main thread of the M10M the below, but it is something I’m also trying to figure out if it’s a valid concern or not, given that if it is I’d probably refrain from shooting with the M10M...seems you’ve used cameras with much higher megapixel count with no issues though, so maybe those concerns are not to be worried about...
 

I read reidreviews on the M10M in which he states there are no concerns regarding focus accuracy / blur, or lens resolution with the M10M 41MP sensor. I wonder why so many people here are expressing concerns around that when no reviewer who has actually used the camera states this as being an issue or concern. 

I’m pretty used to shooting a 35mm lens at 1/15 or 1/30 and never had any issues with the M10 or any other Leica cameras; if that’d be an issue with the M10M it would be counterintuitive and defeat the purpose of a rangefinder over an SLR (ie ability to shoot at slow shutter speeds without camera shake). Would love to hear more on this from people in this thread that have actually used the camera 

 

[elmars]

In my review about the M10-M I wrote this:

„One last word to focus accuracy: 

There is a long lasting (beginning with the M8) discussion, if the mechanical system of the rangefinder is precise enough to handle the requirements of the digital age. Leica had great problems in the M8/M9 days with calibration, but since the M240 it is much better. With the M10 Leica improved the tolerances and the magnification of the viewfinder once again - I think very sufficient. The first thing I did after getting the M10-M was testing the focus accuracy. It was perfectly calibrated with my Apo-Summicron 2,0/50 mm. I had no problems focussing in practice too. And I never noticed anything negative due to the higher resolution of the sensor: Not more blur, no need for higher shutter speed to avoid blur, no problems with sharpness. I shot nearly all photos at f 2,0.“

[shirubadanieru]

6 minutes ago, elmars said:

In my review about the M10-M I wrote this:

„One last word to focus accuracy: 

There is a long lasting (beginning with the M8) discussion, if the mechanical system of the rangefinder is precise enough to handle the requirements of the digital age. Leica had great problems in the M8/M9 days with calibration, but since the M240 it is much better. With the M10 Leica improved the tolerances and the magnification of the viewfinder once again - I think very sufficient. The first thing I did after getting the M10-M was testing the focus accuracy. It was perfectly calibrated with my Apo-Summicron 2,0/50 mm. I had no problems focussing in practice too. And I never noticed anything negative due to the higher resolution of the sensor: Not more blur, no need for higher shutter speed to avoid blur, no problems with sharpness. I shot nearly all photos at f 2,0.“

Thanks for the shots and the review, really awesome stuff! And thanks for sharing the reality on focus accuracy too, seems the people complaining haven’t had the chance to use the camera. On top of focus accuracy, it also seems that higher shutter speeds are not necessary to avoid blur compared to any other Leica rangefinders...so hopefully that should put to rest people’s concerns  thank you!

[marcg]

This is good to hear – but looking through this forum and elsewhere, some people seem to say that higher resolution equals greater focus problems and greater problems of blur – and others seem to say that it doesn't make any difference.
It seems to me that more people are in the former category – so I'm looking forward to see a well-developed discussion about this when more people have actually got their hands on the camera and have tried it out.

I'm surprised that the reviews that have been posted so far have either not referred to this at all or who have simply referred to it in passing

[lct]

Will be interesting to see how sharp 100% magnification outputs will look with difficult lenses like 90/1.5 on high rez sensors. No problem with EVFs when focusing at working aperture but such lenses reach the limits of RF accuracy already. 

[Herr Barnack]

Advertisement (gone after registration)

39 minutes ago, marcg said:

This is good to hear – but looking through this forum and elsewhere, some people seem to say that higher resolution equals greater focus problems and greater problems of blur – and others seem to say that it doesn't make any difference.
It seems to me that more people are in the former category – so I'm looking forward to see a well-developed discussion about this when more people have actually got their hands on the camera and have tried it out.

I'm surprised that the reviews that have been posted so far have either not referred to this at all or who have simply referred to it in passing

A lot of times, internet based photographic advice is worth every dollar we pay for it.

 

 

Edited January 19, 2020 by Herr Barnack

[marcg]

.... an a lot of times Internet-based photographic advice is extremely helpful.

[RexGig0]

Depth-of-Field,  or, if one prefers, Depth-of-Focus, is unaffected by the resolution of the camera’s sensor. Any DOF scale, for any lens, is going to the be the same, for any camera. Focusing accuracy is linear, with the axis of the lens. The blur problems we see with higher-resolution cameras are more likely the result of shooter movement, across pixels.

If one stands, within a tiny box, one is more likely to touch the sides, than if one is standing within a large box. Well, individual pixels can be seen as tiny boxes. More pixels, per unit of area, means each pixel is smaller. Crashing into multiple boxes makes a mess. Shooting with technique that is is inadequate for the task makes a mess, across those tiny pixels.

My first two DSLRs had/have 10MP sensors. The third, a Canon 7D, at 18MP, was quite a leap, and required an improvement in technique. The next big leap was the 50MP Canon 5Ds R. Macro and close-range shooting was important, during this MP progression, as I was doing evidentiary/forensic/crime scene photography. DOF is very thin, as the range gets closer. I never noticed focus accuracy becoming more difficult, whether the camera was a 10MP 40D, or a 50MP 5Ds R. As a mater of actual practice, the huge files of the 5Ds R were a bit much for the proprietary software in the police department’s computer, so I only used the 5Ds R for one crime scene, a limo bus that had been shot full of holes.

Trivia: One compensates for fore-and-aft movement, at close range, by using Continuous AF, or, in Canon-speak, AI Servo, just as if shooting a ball game, at longer range. Yes, hand-held macro is possible. It is amazing how few shooters know this.

 

[lct]

1 minute ago, RexGig0 said:

Depth-of-Field,  or, if one prefers, Depth-of-Focus, is unaffected by the resolution of the camera’s sensor. [...]

DoF is based on CoC (circle of confusion), among other factors, and CoC values depend on magnification. No problem as long as one views low rez and high rez images of the same size at the same viewing distance (about 25 cm for 0.030mm CoC if memory serves) but things change with magnification, especially when viewing 100% crops on screen. 

[caissa]

41 Mp makes a difference with M lenses. I know from the SL2 with 47 Mp.  Somebody showed pics with the Elmarit-M 24. With 24 Mp everything was ok at f 2.8. With 47 Mp the corners were much worse and it was necessary to use 5.6 to get decent quality.  With the new Monochrom it will be somewhere in the middle, but not much better than with 47 Mp.

 

Edited January 19, 2020 by caissa

[RexGig0]

1 minute ago, lct said:

DoF is based on CoC (circle of confusion), among other factors, and CoC values depend on magnification. No problem as long as one views low rez and high rez images of the same size at the same viewing distance (about 25 cm for 0.030mm CoC if memory serves) but things change with magnification, especially when viewing 100% crops on screen. 

Well, on a practical level, at the distances at which one shoots M cameras, I do not see higher-MP presenting focus accuracy challenges.

[lct]

1 minute ago, RexGig0 said:

Well, on a practical level, at the distances at which one shoots M cameras, I do not see higher-MP presenting focus accuracy challenges.

No problem if you don't magnify too much but otherwise what looks sharp at, say, 50% magnification can look soft at 100%. Focusing must be nailed to avoid or reduce this effect and rangefinders are not the best tool for that.

[RexGig0]

1 minute ago, lct said:

No problem if you don't magnify too much but otherwise what looks sharp at, say, 50% magnification can look soft at 100%. Focusing must be nailed to avoid or reduce this effect and rangefinders are not the best tool for that.

I do appreciate the correction. On a technical level, I was incorrect.

[RexGig0]

11 minutes ago, caissa said:

41 Mp makes a difference with M lenses. I know from the SL2 with 47 Mp.  Somebody showed pics with the Elmarit-M 24. With 24 everything was ok at 2.8. With 47 the corners were much worse and it was necessary to use 5.6 to get decent quality.  With the new Monochrom it will be somewhere in the middle, but not much better than with 47 Mp.

 

This is not a focus accuracy problem, but an Elmarit-M 24mm problem, quite apparent when used on 18MP M9 cameras. I am not saying that this lens cannot be used to create wonderful images, but its field curvature, alone, is a significant problem for a shooter to work-around.

[RexGig0]

37 minutes ago, caissa said:

41 Mp makes a difference with M lenses. I know from the SL2 with 47 Mp.  Somebody showed pics with the Elmarit-M 24. With 24 everything was ok at 2.8. With 47 the corners were much worse and it was necessary to use 5.6 to get decent quality.  With the new Monochrom it will be somewhere in the middle, but not much better than with 47 Mp.

 

To use an example from another system, I quickly learned to keep my EF 16-35/2.8L II away from my Canon 5Ds R. Sharp in the center, so, yes, accurately focused, but, So very, well, uninspiring elsewhere. Of course, Canon had already advised of this, by releasing a list of recommended lenses, for the 50MP 5Ds and 5Ds R. Many pre-2015 L lenses were not on that list.

Actually, that 16-35/2.8L II was not all that wonderful on a 7D Mark II, either, even though the APS-C crop factor removed the worst parts of each image. 

Edited January 19, 2020 by RexGig0

[onasj]

As a proof-of-concept experiment to simulate a worst-case scenario, in which the lens and shooter add absolutely zero additional information when going from 24 MP to 41 MP, simply scale your favorite 24 MP photo to 41 MP and take a look.  I strongly suspect the 1.3x longer height and width will result in no significant loss of image quality at any practical magnification or viewing distance.  After shooting with 12, 15, 24, 41, 42, 61, 100, and 151-MP sensors, I conclude their is little practical loss of sharpness when sensor resolution increases, even if one is tempted to print or view somewhat larger as a result.  But there IS a practical benefit in being able to use smaller crops in most shooting circumstances, as with this 8.2 MP crop (a 5.4% crop!) of a 151-MP original: https://www.dropbox.com/s/o86hrqn8qfilj8c/P0001639-Tight Crop.jpg?dl=0

Edited January 19, 2020 by onasj

[caissa]

There is a difference. Whatever the reason is. It happens similarly with many M lenses that I tried. This was just the easiest example to display the problem, as it was already online. (It is not my lens.)   So I am probably not going to upgrade my M246. I do not want to replace all my M lenses with the latest constructions.

For higher resolution I will mainly use L-mount lenses. They are also less expensive than the latest M. (Of course the Apo 50 has no such problems, it has enough reserves).

This is not saying that all lenses are like that, as some here assume. But it is a counter example, to the statement that there is no difference (with all lenses).

Edited January 19, 2020 by caissa

[Agent M10]

1 hour ago, caissa said:

There is a difference. Whatever the reason is. It happens similarly with many M lenses that I tried. This was just the easiest example to display the problem, as it was already online. (It is not my lens.)   So I am probably not going to upgrade my M246. I do not want to replace all my M lenses with the latest constructions.

For higher resolution I will mainly use L-mount lenses. They are also less expensive than the latest M. (Of course the Apo 50 has no such problems, it has enough reserves).

This is not saying that all lenses are like that, as some here assume. But it is a counter example, to the statement that there is no difference (with all lenses).

One solution is some new M lenses along the lines of the 50SL Summilux. Oh wait, there is the new 90mm.

 

[Advertisement]

Hallo,

Look here: Leica M

[SrMi]

2 hours ago, caissa said:

41 Mp makes a difference with M lenses. I know from the SL2 with 47 Mp.  Somebody showed pics with the Elmarit-M 24. With 24 Mp everything was ok at f 2.8. With 47 Mp the corners were much worse and it was necessary to use 5.6 to get decent quality.  With the new Monochrom it will be somewhere in the middle, but not much better than with 47 Mp.

 

I assume that is only if you do not scale the output to the same size, i.e., when pixel peeping?