SL2 vs. S1R difference in IQ (image thread), or is the S1R a good back-up for the SL2 IQ-wise?

[Chaemono]

Advertisement (gone after registration)

Okay, folks, how good are these SL2 files really?  The short answer is, at base ISO they are awesome!  We are talking beyond α7R III, Z7 awesome.  In order to illustrate, see here how they compare to the α7R III.  Apologies for not framing the pictures the same in this example, but I had three cameras with me in the Leica store and I was in a rush to put the Sony away as quickly as possible. 😁  The important thing is that they are equally exposed with the Sony shot at slightly slower shutter speed since the Planar FE 50/1.4 doesn't let as much light in as the 50 Summilux-SL.  Again, we are exposing in order to protect the highlights at base ISO and to see how much we can push the files.  RAWs can be downloaded.

Less compressed JPEGs here: https://www.smugmug.com/gallery/n-tzdLng/

α7R III + Planar FE 50/1.4, link to download the ARW here: https://cc2032.zenfolio.com/img/g206444809-o750076470.dat?dl=2&tk=8DEukkNX73KQStoOiMlynrm9IaMQXjyVBwli11tZIdw=

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!

ISO 100 f/2 @1/320 sec.

SL2 + 50 Summilux-SL, link to download the DNG here: https://cc2032.zenfolio.com/img/g314517939-o750076470.dat?dl=2&tk=Dn00eoyHJTs2c410BbJ6Z4N-hTzTm0zioECIAybUHYk=ISO 100 f/2 @1/400 sec.

Edited November 9, 2019 by Chaemono

[Chaemono]

Now with the exact same adjustments in LR, AWB by the camera tweaked to try to match, Exposure +2.8, Shadows +100, Highlights -100, Sharpening +40, NR +60.  Maybe Highlights at -100 was a bit too much on the Sony because the α7R III picture gets this over-the-top HDR-look. 😂 The important observation, though, is how easily banding appears in shadows of the Sony file.  This is not PDAF striping, obviously, but more likely PDAF banding or just plain pushed shadows banding.  Who cares, it stinks.  See how clean the SL2 picture looks on the other hand.  The SL2 files are simply amazing at base ISO.

Less compressed JPEGs here: https://www.smugmug.com/gallery/n-tzdLng/

α7R III + Planar FE 50/1.4

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!

SL2 + 50 Summilux-SL

Edited November 9, 2019 by Chaemono

[Chaemono]

And now the crops of the upper right corner. Again, the adjustments are Exposure +2.8, Shadows +100, Highlights -100, Sharpening +40, NR +60.

α7R III + Planar FE 50/1.4 crop from the above 😂

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!

SL2 + 50 Summilux-SL 😍

Enjoy!

[jrp]

Thanks for taking the trouble. 

What this does illustrate is that, at web resolution, the differences are not apparent. But I suppose that we knew that even at 12Mpx.

But this gives rise to a more interesting question: in what circumstances (use cases) is camera a better than camera b? At that point, other usability factors may come to the fore. It’s no good having a better sensor if you didn’t get the shot in the first place (and might have done with camera b). If the alternative is out of focus shots, some shadow striping may be an acceptable trade off. 

[Chaemono]

vor 42 Minuten schrieb jrp:

Thanks for taking the trouble. 

What this does illustrate is that, at web resolution, the differences are not apparent. But I suppose that we knew that even at 12Mpx.

But this gives rise to a more interesting question: in what circumstances (use cases) is camera a better than camera b? At that point, other usability factors may come to the fore. It’s no good having a better sensor if you didn’t get the shot in the first place (and might have done with camera b). If the alternative is out of focus shots, some shadow striping may be an acceptable trade off. 

All good points that my comparisons are not meant to answer.  Instead, this thread is inspired by the DPR review of the original SL four years ago here: https://m.dpreview.com/reviews/leica-sl-typ601-in-depth-camera-review/3 where they concluded that because of the pushed shadows banding issue which the SL exhibited, like the Q, the DR of the camera was limited.  Quote from the DRP article linked above: “The Raw dynamic range of the Leica SL shows some significant limitations. It falls well behind class-leaders like the Nikon D750, D810, or Sony a7R II.  [...]  What this ultimately means is that you'll be limited in your ability to decrease exposure to expose so as to not blow highlights in high contrast scenes, because you'll be limited in your ability to correct (brighten) dark tones in post-processing.”

I wonder what DPR will say now because the SL2 kicks some serious DxOMark score 100 FF butt in this respect. 😂

Edited November 9, 2019 by Chaemono

[dritz]

On 11/9/2019 at 8:49 AM, Chaemono said:

All good points that my comparisons are not meant to answer.  Instead, this thread is inspired by the DPR review of the original SL four years ago here: https://m.dpreview.com/reviews/leica-sl-typ601-in-depth-camera-review/3 where they concluded that because of the pushed shadows banding issue which the SL exhibited, like the Q, the DR of the camera was limited.  Quote from the DRP article linked above: “The Raw dynamic range of the Leica SL shows some significant limitations. It falls well behind class-leaders like the Nikon D750, D810, or Sony a7R II.  [...]  What this ultimately means is that you'll be limited in your ability to decrease exposure to expose so as to not blow highlights in high contrast scenes, because you'll be limited in your ability to correct (brighten) dark tones in post-processing.”

I wonder what DPR will say now because the SL2 kicks some serious DxOMark score 100 FF butt in this respect. 😂

Perhaps this is off topic... but if you plan to shoot at base ISO and then leverage the dynamic range, why not buy a used S?

[Chaemono]

Advertisement (gone after registration)

vor 35 Minuten schrieb dritz:

Perhaps this is off topic... but if you plan to shoot at base ISO and then leverage the dynamic range, why not buy a used S?

Your question implies that the S has an ISO-less sensor.  I actually don’t know that.  I know that the X1D sensor is not ISO-less.

[scott kirkpatrick]

Just to take a few more swipes at this particular dying horse -- I wanted to get some facts to use in setting upper ISOs for Auto ISO on my SL2, and I was wondering if the extra grittiness resulting from the amplification needed to get the same signal from the SL2's 4.5 micron pixels as we are used to from the SL's 6 micron pixels could be compensated for by reducing the SL2 output to the same 24 MPx as the SL's.  So here are some comparisons.  The SL2 have been output at 70% and the SL at 100%.  I then cut out a 1220x900 pixel slice from each shot of the same part of my standard bookshelf, comparing the results at various ISOs.

Here's ISO 12500:  first 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!

U1000759 by scott kirkpatrick, on Flickr

then SL

S1000426 by scott kirkpatrick, on Flickr

and then ISO 6400, first SL2

U1000758 by scott kirkpatrick, on Flickr

and SL:

S1000425 by scott kirkpatrick, on Flickr

and finally ISO 3200, first SL2:

U1000757 by scott kirkpatrick, on Flickr

and SL:

S1000424 by scott kirkpatrick, on Flickr

I'll spare you all 25000 and 50000 ISOs.  They are for emergency use only.  Also, I took all shots in not very contrasty lighting, with exposure offset -1, to bring the white books into midtones. The color shift is from the addition of indirect daylight. 

I shot the same series with the S1R.  There is a difference between the S1R and SL2 when you go above 3200.  The S1R starts to put some noise reduction into the raw file, while the Leica raw images are grittier but give higher resolution. 

edit:  So my conclusion is that the two results are equivalent in resolution of fine detail and texture at 3200 and 6400, but the SL2, when rendered down to 24 MPx, does a little better at 12500.  There is some mottling in the 12500 treatment of broad smooth areas with the SL that the SL2 controls.  I wouldn't use 25000 in either camera, but the SL (not shown) survives a bit better there.

 

Edited January 9, 2020 by scott kirkpatrick

[Chaemono]

vor 4 Minuten schrieb scott kirkpatrick:

Just to take a few more swipes at this particular dying horse -- I wanted to get some facts to use in setting upper ISOs for Auto ISO on my SL2, and I was wondering if the extra grittiness resulting from the amplification needed to get the same signal from the SL2's 4.5 micron pixels as we are used to from the SL's 6 micron pixels could be compensated for by reducing the SL2 output to the same 24 MPx as the SL's.  So here are some comparisons.  The SL2 have been output at 70% and the SL at 100%.  I then cut out a 1220x900 pixel slice from each shot of the same part of my standard bookshelf, comparing the results at various ISOs.

Here's ISO 12500:  first 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!

U1000759 by scott kirkpatrick, on Flickr

then SL

S1000426 by scott kirkpatrick, on Flickr

and then ISO 6400, first SL2

U1000758 by scott kirkpatrick, on Flickr

and SL:

S1000425 by scott kirkpatrick, on Flickr

and finally ISO 3200, first SL2:

U1000757 by scott kirkpatrick, on Flickr

and SL:

S1000424 by scott kirkpatrick, on Flickr

I'll spare you all 25000 and 50000 ISOs.  They are for emergency use only.  Also, I took all shots in not very contrasty lighting, with exposure offset -1, to bring the white books into midtones. The color shift is from the addition of indirect daylight. 

Did you underexpose and then push exposure in post?  WB not only looks off for the SL2 pictures, there’s a green tint to them.  

[scott kirkpatrick]

No post processing at all.  The color shift is because the SL2 shots were taken with LED and flourescent lighting, while the SL had indirect daylight for about half the illumination.  (Or maybe it was the other way around.)  I wasn't interested in the color, so ignored it.  I underexposed because I wanted to see when noise would corrupt the lower midtones, zones 3-5 in the old way of looking at it.  This really shows when you push to 25000 and above -- there is no DR left, despite what the websites say.  Rendering at a lower number of pixels pushes the noise back down and is necessary to use the higher ISOs.

Edited January 9, 2020 by scott kirkpatrick