M10-R vs M10-P high-ISO, acuity, and noise side-by-side comparison results

[onasj]

Advertisement (gone after registration)

5 hours ago, rramesh said:

I am unsure of the effectiveness of these tests. When focusing with a rangefinder there is no focus confirmation. It’s really the photographer who decides when focus has been achieved. So two shots, two different cameras will always be subjective, wouldn’t it. 

There wasn’t any perceptible difference in focusing between the shots used.  The lens ended up set to the same distance, the tripod remained in the same position, and the f/5.6 aperture means the depth of focus was quite generous in its coverage of the relevant parts of the scene. I used live view at maximum magnification to verify that focusing was not different between the cameras.  You can download the files to confirm all this.

[LBJ2]

https://www.photonstophotos.net/Charts/PDR.htm#Leica M10,Leica M10-R,Leica Q2,Leica SL2

 

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!

[scott kirkpatrick]

Interesting measurement.  The main takeaways I get are:

 Leica has gotten a full stop advantage from their work on a "bespoke" sensor from a smaller supplier, used in the M and S series (go to the source and take a look at the s[007] specs) as opposed to the more commercial standard sensor used in the SL (and in Panasonic's S1R, I suspect).

basing Auto ISO on 100 seems right for the M10-R, as is the current default, and you can see from the curvature of the M10 line why it was safer for the M10 to use ISO 200 as a base.  Straight lines on this plot lead to simpler auto-ISO adjustments.  So the M10-R has two straight line sections.  And to get the absolute most highlight relief on the M10-R it might still help a bit to use M to get ISO 125, or more simply, add -0.3 in exposure compensation.

There seems to be some improvement in the extreme high ISOs, which may also translate into getting more information out of the deepest shadows

[Guest Nowhereman]

For me, the conclusion from these curves is quite different: how close the M10-R and M10 curves are. That makes me wonder why some M10-R reviewers and early buyers have felt that the new camera has a significant improvement in dynamic range, especially in the highlights, over the M10. Indeed, some have said that the improvement in highlight recovery of the M10-R is more significant than the increased resolution. I wonder.
____________________
Frog Leaping photobook

[convexferret]

1 hour ago, Nowhereman said:

For me, the conclusion from these curves is quite different: how close the M10-R and M10 curves are. That makes me wonder why some M10-R reviewers and early buyers have felt that the new camera has a significant improvement in dynamic range, especially in the highlights, over the M10. Indeed, some have said that the improvement in highlight recovery of the M10-R is more significant than the increased resolution. I wonder.
____________________
Frog Leaping photobook

Because highlight dynamic range is not a simple thing, relying as it does on the RAW converter "inventing" the missing data. For that reason the highlight DR improvements don't necessarily show up in a standard chart as above. You can see this by looking at the M10R vs M10M and seeing that they have the same DR according to Photons to Photos.

[Guest Nowhereman]

^

1 hour ago, convexferret said:

Because highlight dynamic range is not a simple thing, relying as it does on the RAW converter "inventing" the missing data. For that reason the highlight DR improvements don't necessarily show up in a standard chart as above. You can see this by looking at the M10R vs M10M and seeing that they have the same DR according to Photons to Photos.

Actually, no: as one would expect, the M10M has a substantially higher dynamic range than the M10-R in the Photon to Photo curves. Extrapolating from the M10 curve, which is very close that of the M10-R, one could expect that a Monochrom based on the M10 sensor, could have a DR close to that of the M10M, don't you think?

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!

____________________
Frog Leaping photobook Edited August 7, 2020 by Nowhereman

[elmars]

Advertisement (gone after registration)

I did a simple test: I overexposed the same motive with both cameras up to 3 f-stops. With the M10-R a one f-stop more highlight recovery was possible.

[SrMi]

6 minutes ago, elmars said:

I did a simple test: I overexposed the same motive with both cameras up to 3 f-stops. With the M10-R a one f-stop more highlight recovery was possible.

Highlight recovery and dynamic range are not the same. The former depends on the exposure. 

[convexferret]

1 hour ago, Nowhereman said:

^

Actually, no: as one would expect, the M10M has a substantially higher dynamic range than the M10-R in the Photon to Photo curves. Extrapolating from the M10 curve, which is very close that of the M10-R, one could expect that a Monochrom based on the M10 sensor, could have a DR close to that of the M10M, don't you think?

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!

____________________
Frog Leaping photobook

Can you explain to me why a monochrome camera would have a better dynamic range than a colour camera with the same base sensor? I'm absolutely serious.

Edited August 7, 2020 by convexferret

[scott kirkpatrick]

4 minutes ago, convexferret said:

Can you explain to me why a monochrome camera would have a better dynamic range than a colour camera with the same base sensor? I'm absolutely serious.

the same noise figures, but a brighter sensor with no Bayer filter on top.  

Something people seem to be missing is that the smaller pixels in the 40+ MPx cameras are in fact giving the same dynamic range as those in the 24 MPc=x generation.  That is not easy to do.

[elmars]

vor 59 Minuten schrieb SrMi:

Highlight recovery and dynamic range are not the same. The former depends on the exposure. 

Same exposure of course. 

[convexferret]

1 minute ago, scott kirkpatrick said:

the same noise figures, but a brighter sensor with no Bayer filter on top.  

Something people seem to be missing is that the smaller pixels in the 40+ MPx cameras are in fact giving the same dynamic range as those in the 24 MPc=x generation.  That is not easy to do.

No, that explains the higher ISO for the same dynamic range, not a greater base dynamic range. Check the M9 vs M9M and M240 vs M246 graphs.

[Guest Nowhereman]

2 minutes ago, scott kirkpatrick said:

...Something people seem to be missing is that the smaller pixels in the 40+ MPx cameras are in fact giving the same dynamic range as those in the 24 MPc=x generation.  That is not easy to do.

Good point, especially for those who want the greater resolution. In my case, I don't want the greater resolution of the M10-R and would be interested in the M10-R if it had decisively better dynamic range than my M10. Actually, I don't know what to make of the Photons to Photos curves showing virtually the same dynamic range for the M10-R and M10 versus the statements of some reviewers and early users that the dynamic range of the M10-R being substantially better than that of the M10.
____________________
Frog Leaping photobook

 

 

[scott kirkpatrick]

Bill Claff's curves don't tell you how the camera actually shoots, they just show its range.  The reviewers' comments seem to be that the lower basic ISO and other parts of the camera setup (straight lines on the DR plot that don't leave you running out of headroom at lower ISOs) make it a pleasure to shoot.  Lower noise means good shadow recovery.  My experience so far is consistent.

Edited August 7, 2020 by scott kirkpatrick

[Jeff S]

48 minutes ago, scott kirkpatrick said:

Something people seem to be missing is that the smaller pixels in the 40+ MPx cameras are in fact giving the same dynamic range as those in the 24 MPc=x generation.  That is not easy to do.

True, and as Stefan Daniel has explained, the M10-R sensor performance was improved despite the expected challenge. Specifically he mentioned four improvements:  greater active sensor area vs inactive (supporting electronics); 10% more photons; dual gain architecture; and reshaped pixels (to correct a small irregularity).  Other tests seem to confirm better dynamic range and improved ISO performance vs the M10. 
 

Jeff

Edited August 7, 2020 by Jeff S

[SrMi]

2 hours ago, elmars said:

Same exposure of course. 

Some cameras need underexposure to reach their potential (e.g., highlight recovery). M10-P is best underexposed a bit, while M10-R work without underexposure.

[SrMi]

2 hours ago, scott kirkpatrick said:

the same noise figures, but a brighter sensor with no Bayer filter on top.  

Something people seem to be missing is that the smaller pixels in the 40+ MPx cameras are in fact giving the same dynamic range as those in the 24 MPc=x generation.  That is not easy to do.

Bill's PDRs are measured on normalized (not to 8MP but apparently similar) images.

If you check Nikon's, Sony's and Panasonic's 24MP vs. +40MP models, you will notice very similar DR.

Measurements are nice, but they are not the final word.

Edited August 7, 2020 by SrMi

[Stevejack]

Anyone know what's going on with the banding on the top of the card in the M10R? I can see this in all of the ISO samples for the M10R, but the M10 doesn't show it. 

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!

[Stuart Richardson]

Looks like just regular moire. It happens whenever the frequency of the detail is aligned with the resolution of the sensor, and it causes a sampling error which leads to this kind of repeating pattern. If you move slightly forward or back, or stop the lens down enough that diffraction is limiting the detail, then you will find it will go away. In fact, it is still present in your second photo, but it happens to be visible in the high frequency detail in the vase to the left, rather than the picture of the man (see the rainbow noise in the detail...that's it). This is why cameras use Anti-Aliasing filters, but you can get sharper results without them, which is why Leica and many companies now leave them out. But notice how they are included in video centric cameras like the S1H...for video it is better to be slightly softer and not have moire than to be sharp with moire.

[Advertisement]

Hallo,

Look here: Leica M

[roverover]

Great work on these comparisons and thank you for your efforts, it's always interesting to see Your samples, you do a great job!

Question:  Since the physical sensor size in the M10 and M10R are the same, yet the M10R has a higher pixel count, what i see is that the smaller pixels in the M10R
are seemingly creating smoother noise patterns, especially noticeable at the higher ISO examples although I see better detail in the M10R at EVERY ISO setting.