DxOMark tested M11

[nicci78]

Advertisement (gone after registration)

Photons to photos has its perks too.

For example Leica TL2 ranked among the top 10  in High ISO and low light !!! 
Or Pentax APS-C K3 mark III ranked into top 3 for high ISO and low light. 
Clearly P2P is not reliable. Too many weird datas. 

[Advertisement]

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

[SrMi]

19 minutes ago, nicci78 said:

Photons to photos has its perks too.

For example Leica TL2 ranked among the top 10  in High ISO and low light !!! 
Or Pentax APS-C K3 mark III ranked into top 3 for high ISO and low light. 
Clearly P2P is not reliable. Too many weird datas. 

K3 III and high ISO of TL2 are clearly marked to have in-camera noise reduction, hence the good measurements. How does DXOMark handle cameras that have in-camera noise reduction?

The only issue with P2P is that the ISO is not normalized and therefore it is tricky to compare different cameras.

[LBJ2]

On 5/5/2022 at 5:24 PM, FlashGordonPhotography said:

Well the SL2, which you call sub par, shares its silicon with the S1R, which scores 100 on DXO Mark. I have both and they are remarkably similar. And I couldn't get extreme banding in images from my SL2 whereas it was not uncommon when really pushing my A7R2 or A7R3 files. The M8 and M9 are known for their stunning black and white conversions.

I don't subscribe to the *sub par* sensor discussions. They are what they are and like every other part of the camera you learn the limitations and go make photographs. Some of my favourite images were made with the M9. A *better* sensor would not make those better images. It's knd of like saying slide film was sub par to colour neg because it had vastly lower DR (less than 6 stops in the case of Velvia).

Absolutely, a sensor with wider DR or lower high ISO noise can make some things easier. I too, choose cameras with high DR, bigger sensors and low noise when shooting large landscapes. But mostly these are things seen by photoraphers and forums. I've never, ever had an image critiqued with, "too bad it's too noisy" except by some muppet on a forum. If what people are noticing in a photo is a lack of DR or some noise then the image had already failed.

Gordon

I believe that banding you were able to push on the A7rII/III is a result of PDAF. Not something we would see on the CDAF SL2 sensors. 

[nicci78]

Easy DXOMark refuses to test any “weird” sensors : hence no Fujifilm X-trans. Nor Ricoh/Pentax with their noise reduction Co-Processor. 
TL2 did not get tested either. 
 

DXOmark and DXO Labs gave frequently interviews in french. They are french after all. So their work is may be better understood here, than elsewhere. 
 

[SrMi]

7 hours ago, nicci78 said:

Easy DXOMark refuses to test any “weird” sensors : hence no Fujifilm X-trans. Nor Ricoh/Pentax with their noise reduction Co-Processor. 
TL2 did not get tested either. 
 

DXOmark and DXO Labs gave frequently interviews in french. They are french after all. So their work is may be better understood here, than elsewhere. 
 

DXOMark has tested cameras with built-in noise reduction (e.g., Canon R3, R5, Pentax K3 II). AFAIK, and unlike P2P, DXO does not report whether there is a built-in NR.

[nicci78]

Effectively DXOMark only reports out of analog range data. 
However R3, R5 or K-3 II are “normal” cameras ! Hundreds of them applied NR at high ISO for decades

On the other hand : K-1 II, K-P, K-70, GR III, K-3 III has an accelerator unit. Which is a specialised chip baking NR at almost every ISO settings !! It is far from being “typical” behaviour.

Edited May 9, 2022 by nicci78

[SrMi]

Advertisement (gone after registration)

45 minutes ago, nicci78 said:

Effectively DXOMark only reports out of analog range data. 
However R3, R5 or K-3 II are “normal” cameras ! Hundreds of them applied NR at high ISO for decades

On the other hand : K-1 II, K-P, K-70, GR III, K-3 III has an accelerator unit. Which is a specialised chip baking NR at almost every ISO settings !! It is far from being “typical” behaviour.

Canon R3 has NR baked in for the whole ISO range (link). On the other hand, R5 applies NR for low ISO only (link). Pentax K-3 II applies NR starting with ISO 1600 (analog range).

[nicci78]

But neither Canon R3 nor K-3 II used a kind of accelerator unit. Which may imply much more aggressive treatments. 
 

For now I am using my GR III and IIIx with factory settings for noise reduction. Both has the infamous Accelerator unit.

I am not sure how to tweak it yet. One thing is sure, CL has more attractive noise pattern than GR III/IIIx. The Ricohs beats CL only at 25000 and above… but they are not useful at all !! 
With IBIS and AF201FG flash gun they almost never go beyond 6400 (only 1%) On the other hand my CL has over 20% of shots beyond 6400 ISO. 25000 ISO is quite frequent with CL (no IBIS, no flash) 

 

Just to sum, I can see the kind of effect Accelerator unit does : it softened the image by reducing both luminance and colour noise. Then apply sharpening.
 I am not sure to like it at all. But it does not bother me neither for now. 
Only side by side comparison, makes the Ricoh less attractive than CL. 
But GR III never goes beyond 6400

Edited May 9, 2022 by nicci78

[FlashGordonPhotography]

On 5/9/2022 at 6:30 AM, LBJ2 said:

I believe that banding you were able to push on the A7rII/III is a result of PDAF. Not something we would see on the CDAF SL2 sensors. 

Yes, this is correct. It's something the measurement sites don't report on as they use downsized files for their testing which disguises it. In the real world at full resolution the SL2 is not inferior to the A7R3/4 when banding is taken into account. Often it has more *usable* DR because it can be pushed harder.

DXO/P2P figures can be useful for some but they don't tell the whole story. Relying on them as an absolute is not something I will ever do.

Gordon

[SrMi]

5 hours ago, FlashGordonPhotography said:

Yes, this is correct. It's something the measurement sites don't report on as they use downsized files for their testing which disguises it. In the real world at full resolution the SL2 is not inferior to the A7R3/4 when banding is taken into account. Often it has more *usable* DR because it can be pushed harder.

DXO/P2P figures can be useful for some but they don't tell the whole story. Relying on them as an absolute is not something I will ever do.

Gordon

According to Jim Kasson's measurements, a7rIV does not have PDAF banding but PDAF striping that appears only in certain lighting situations.

DPR reported PDAF banding with Z7. However, it was also not detectable in typical operation (link vs. changed lighting link).

SL2 has more banding than SL2-S when lifting shadows, and neither camera has PDAF.

P2P has a page that should inform about the banding potential: Sensor Heatmaps.