Leica M11 Sensor Measurements at PhotonsToPhotos

bclaff wrote:

SrMi wrote:

bclaff wrote:

SrMi wrote:

bclaff wrote:

jjcha wrote:

I'm looking very far down the line - e.g., ISO25,600 and comparing against the old M10, since your chart doesn't go that far down for the M10-P.

I'm assuming the M10-P was slightly worse than the M10, though I'm not sure exactly why that would be the case.

But further down the line, at ISO25,600 is 4.13 for the M11 vs 3.71 for the M10, so 0.42 in difference.

At that end of the scale, my photographs, which are shot at night under poor lighting conditions at F4.0 or F5.6, really are pushing it as to what's "acceptable" with the noise and lack of dynamic range.

A third to half a stop of improvement, for me at least, might, depending on the processing, keep a photo from completely falling apart by being able to recover just a little more depth and being that slightly bit less noisy.

It always felt with the M10-P, for my photos, it was fighting to get the benefits of that last 0.25 to 0.5 stop push of exposure, while not completely losing the fact that it was a photograph (and not falling apart into some comic book image.)

Sure. In extreme situations Fixed Pattern Noise (FPN) will affect how pleasing the results are visually.
You can get a sense of that on the Sensor Heatmaps page.
The M-10 seems to have an advantage here.

For the row and column metrics lower is better.
Also check the link for a visual representation.

Thanks, I was not aware of that metric!

Yeah, there's quite a bit at PhotonsToPhotos that is underutilized !

Word! You need to start doing those monetizable YouTube videos ;-).

The higher the illuminated row and column numbers, the more pleasing the results are visually, right? Do you have any additional information about that metric? The higher the numbers, the less banding?

Leica SL2-S, known for very pleasing high ISO results, has numbers 0.025 and 0.013.

No. As I said initially. Lower is better with zero (never seen) perfect.

Lower is better? I am confused, as you said that M10-P (0.022) seems to have an advantage over M11 (0.012). Doesn't that mean the opposite, that M11 is better than M10-P?

If you look above I clearly said "The M-10 seems to have an advantage here."

Got it. You highlighted M10-P, which made me think that you referred to M10-P, even though you clearly said M10.

Really would like to learn more about that metrics.

Kindof "chicken and egg" the heatmaps and metrics are not used much so it's unclear that it's a good use of my time writing about that at this point.
If you know any statistics it's a sort of covariance and you don't want any correlation so lower is better.

Anyway, thanks a lot for your work and for answering my questions.

Always willing to help people understand things better.