Jump to content
Jared

Image Quality--CL vs SL vs M10

Recommended Posts

Advertisement (gone after registration)

I did what I thought was an interesting "shoot out" this past Sunday in my back garden.  I compared the following:

 

Leica CL with 18-56 TL zoom set to 22mm, f/4

Leica CL with 21mm Summilux-M, f/4

Leica CL with 21mm Super Elmar M, f/4

Leica CL with 11-23 TL zoom set to 22mm, f/4.4 (wide open at this focal length)

Leica M10 with 35mm Summilux M FLE, f/5.6

Leica SL with 24-90mm Vario Elmar set to 34mm, f/5.6

 

Cameras were tripod mounted.  A shutter delay of 2s was used to reduce the chances of blur from camera movement.  Image stabilization was turned off with the SL.  ISO was set to base value.  Same point of focus was used in all cases.  The objective was to see whether there were any obvious differences in image quality on a real, three dimensional subject when shooting under optimum conditions.  The focal ratios chosen were to provide optimum image quality for the lens and to ensure depth of field differences between formats were minimized.  The lighting varied slightly from minute to minute since those was outdoors, but all images were shot at approximately the same time under overcast conditions.

 

I repeated the experiment from a slightly longer subject distance to get essentially the same framing with a different focal length:

 

Leica CL with 18-56 TL zoom set to 35mm, f/4.9 (wide open at this focal length)

Leica CL with 35mm Summilux-M FLE, f/4

Leica CL with 35mm Summilux TL, f/4

Leica M10 with 50mm APO Summicron M, f/5.6

Leica SL with 24-90mm Vario Elmar set to 50mm, f/5.6

 

End result?  At optimum aperture and when one's technique is good, I really couldn't find a consistent difference among any of these lenses.  Frankly, the biggest challenge was in getting the point of focus the same in all images such that depth of field was distributed evenly in front of and behind the point of focus.  In one case in particular, I simply missed focus a bit (despite it being an AF lens--probably wasn't careful enough in my selection point).  Some lenses had slightly different field curvature characteristics.  There were very slight differences in edge of field sharpness. Very slight differences in color separation on fine details.  That's about it.  Could I tell the difference in noise characteristics between the two full frame cameras and the APS-C camera?  Not at base ISO I couldn't.  Maybe if there had been some sky in the field of view or some other really smooth surface I could have. Maybe not.  What about dynamic range?  Not under overcast conditions where all three cameras were capable of capturing the entire range of tones without any clipping.  Were there differences in how out-of-focus areas were rendered?  Sure, but it would be nearly impossible to pick "winners" and "losers" with the particular garden scene I was using--no bokeh balls or specular highlights for evaluation. What about color cast or saturation?  Well, I've already got my M10 dialed in the way I want it, and the SL and CL were actually closely matched, so the colors came out very nearly identical with the largest factor being the slightly variable lighting conditions. Certainly, I wouldn't be able to pick out a lens/camera combo that had "better" or even "different" colors.

 

Of course, I wasn't pushing any boundaries here.  The comparison was at optimum aperture for each lens or very nearly so.  The lighting and dynamic range were easy to handle.  There was ample light for shooting at base ISO. The tripod minimized or eliminated camera blur. The subject was static. I was taking my time so that focus wouldn't be an issue (and even so, I missed one slightly--just enough to be visible at 100% magnification on my 5K monitor).    

 

What does this tell me?  Nothing shocking, really.  I would choose my lens and camera based on size/weight, field of view needed, control over depth of field needed, available light, need for autofocus, and need for image stabilization. All the combinations above are equally capable if you aren't pushing the edges too much. We all worry more than we probably need to about image quality, lens rendering, sharpness, and similar factors.  Frankly, the worst image I created from a technical perspective was using one of the best lenses--the 35mm Summilux TL.  Even when I was trying to be careful and really nail the "technique" for any factors that could impact image quality, a small difference in focus point was the single largest factor in sharpness! 

 

I'd post the images, but frankly there is no point.  At web sizes and with JPG compression, they are identical.  

Share this post


Link to post
Share on other sites

Now here are four crops from the center of the frame.  They are:

- 35mm Summilux M FLE mounted on the Leica CL, f/4

- 21mm Super Elmar M mounted on the Leica CL, f/4

- 35mm Summilux M FLE mounted on the Leica M10, f/5.6

- 50mm APO Summicron M mounted on the Leica M10, f/5.6

 

Can you tell which is which?  Or even tell me which of the crops you like best and why?

 

Let's call the "top left" image 1:1, the "top right" image 1:2, the "lower left" 1:3, and the "lower right" 1:4.  The next batch to follow...

 

Share this post


Link to post
Share on other sites

Here is the second set of crops...

 

This includes the following lenses and cameras:

- 24-90 Vario Elmar @ 50mm mounted on the Leica SL, f/5.6

- 11-23 Super Vario Elmar @ 22mm mounted on the Leica CL, f/4.4 (wide open for this focal length)

- 18-56 Vario Elmar @ 22mm mounted on the Leica CL, f/4

- 24-90 Vario Elmar @ 34mm mounted on the SL, f/5.6

 

Can you tell which is which?  

 

 

Let's call the "top left" image 2:1, the "top right" image 2:2, the "lower left" 2:3, and the "lower right" 2:4.

 

Obviously, the actual order does not match the list.

 

Share this post


Link to post
Share on other sites

And here is the last batch.  

 

- 35mm Summilux TL mounted on the Leica CL, f/4 (this is the one I slightly mis-focused)

- 18-56mm Vario Elmar TL @ 35mm mounted on the Leica at f/4.9 (wide open for this lens)

- 21mm Summilux M mounted on the Leica CL, f/4 (not it's best aperture, but I wanted to be consistent with everything else)

 

Place your bets!  You've got a better shot on this one since there are only three to choose from.

 

3:1 is upper left, 3:2 is upper right, 3:3 is the lower image.

 

Share this post


Link to post
Share on other sites

I'm not good at comparing different apertures but as far as my M lenses on M240 and CL are concerned, i did not see significant differences in the centre of the frame around f/4 - f/5.6 either somewhat expectedly. Now edges and corners can be another story, especially with M wides at fast apertures. Did you do some comparisons there? Just curious. 

Share this post


Link to post
Share on other sites

Advertisement (gone after registration)

I did. At these apertures, the differences in the corners are similarly small. Basically, as soon as you have a real-world subject with some depth, loss of sharpness was due almost exclusively to being outside the plane of focus. I could post crops from the edge and it would be a very similar result. The extreme corners for this subject didn’t have anything in focus, but some of the leaves on the far left edge work. There is some variation in field curvature, thus some variation in which leaf is best, but there are still more similarities than differences.

Edited by Jared

Share this post


Link to post
Share on other sites

I did. At these apertures, the differences in the corners are similarly small. Basically, as soon as you have a real-world subject with some depth, loss of sharpness was due almost exclusively to being outside the plane of focus. I could post crops from the edge and it would be a very similar result. The extreme corners for this subject didn’t have anything in focus, but some of the leaves on the far left edge work. There is some variation in field curvature, thus some variation in which leaf is best, but there are still more similarities than differences.

 

Interesting indeed but it is not my feeling i must say. Corners look softer on my CL out of some of my lenses like 35/2 asph v1 or 28/2.8 asph v1 at full aperture at least and perhaps one or two stops below. I didn't have the time to do serious comparisons though so your findings would be welcome here if you can share them already.

Share this post


Link to post
Share on other sites

I did a less extensive version of this sort of test a while back with the 11-23 and 23/2.0 CL and the 28/1.4 and 35/1.4 last model M Summiluxes using a thin layer of green things growing on the side wall of our house as a target.  As Jared reports, all four are more than "good enough for government work,"  but there were differences visible on a reasonable-sized screen.  My takeaways were:

 

f/5.6 is better than f/4.0 with all of these lenses for controlling sharpness, especially at the edges, without any loss in contrast

 

The 11-23 at 23mm is sharper than the 23/2.0 and focuses a bit closer in addition, although differences at the center of the frame are very tiny.

 

My 28/1.4 is a better lens than the 35/1.4, as it should be, since the design is about ten years more recent.

Share this post


Link to post
Share on other sites

Interesting indeed but it is not my feeling i must say. Corners look softer on my CL out of some of my lenses like 35/2 asph v1 or 28/2.8 asph v1 at full aperture at least and perhaps one or two stops below. I didn't have the time to do serious comparisons though so your findings would be welcome here if you can share them already.

At full aperture, the differences in the corners are much larger, but I would never shoot a subject where depth of field mattered (like this one) at full aperture. I’ll post the crips, though, so you can see.

Share this post


Link to post
Share on other sites

My point isn’t that there are no differences at all across all apertures. There are. There are some very subtle but consistent differences in color separation and saturation on extremely fine detail. There are certainly differences in corner softness and field curvature. Those differences are quite noticeable wide open. There may even be some differences in how depth of field is distributed in front of and behind the plane of focus. However, in the real world these differences are extremely subtle (except corner sharpness wide open), and with a large percentage of subjects—including the sample image Inposted—these differences won’t matter one bit. None of my corners, for example, have anything that is contributing to the picture in a useful way, and nothing that is remotely in the plane of focus, so there is nothing that is remotely sharp. That is the rule rather than the exception in most images.

 

I would recommend choosing your lens based on angle of view desired, depth of field required (or amount of light available), the need for autofocus, and size/weight for carrying. The rest just isn’t going to matter for the vast majority of subjects. All the current lenses are excellent anywhere near the center of the frame at any aperture from wide open to the diffraction limit of the format.

Share this post


Link to post
Share on other sites

[...] There are certainly differences in corner softness and field curvature. Those differences are quite noticeable wide open. [...]

 

Matches my feelings as well and so not only wide open but to a lesser extent at medium apertures but it depends on lenses apparently. I would be interested to see what happens with M wides in corners and edges as i suspect the CL's sensor stack is thicker than those of my M8.2, M240 and Sony A7s mod. Just a guess though but the CL is less sensitive to moiré and IR contamination so i suspect the difference may come from this thickness w/o any evidence so far. 

Share this post


Link to post
Share on other sites

Jared ...... you have my deepest sympathy....

 

I have carried out this exercise for a number of Leica cameras in the past and it is a thankless task...... achieving absolute repeatability and valid comparisons is very difficult.

 

My findings were that there was little difference between similar resolution sensors ...... certainly not enough to show up in normal usage.

 

The older 28mm and 35mm (2 & 2.8) perform poorly wide open in the peripheries on digital sensors not specifically designed for M lenses (ie. everything outside the M stable). 

 

Everything you have found is what I would have expected from my tinkering with M, SL, TL2 and CL. 

Edited by thighslapper

Share this post


Link to post
Share on other sites

Matches my feelings as well and so not only wide open but to a lesser extent at medium apertures but it depends on lenses apparently. I would be interested to see what happens with M wides in corners and edges as i suspect the CL's sensor stack is thicker than those of my M8.2, M240 and Sony A7s mod. Just a guess though but the CL is less sensitive to moiré and IR contamination so i suspect the difference may come from this thickness w/o any evidence so far. 

I've seen moiré on the CL...

Share this post


Link to post
Share on other sites

I can't give you the extreme corner since there is nowhere in the image where the extreme corner was close to the plane of focus.  However, there is a branch in the upper right near the edge that I can use.  Because of the difference in perspective when using the shorter focal lengths vs. the longer focal lengths, it isn't in the exact same location in every image, but it's reasonably close and will give you an idea of the variability across multiple lenses and cameras at moderate aperture.

 

Since nobody seems to want to guess (not surprised), I'll label them as well.  It's actually the same order as the first batch.

 

I need to make one substitution for this batch.  In the image made with the 21mm SEM mounted on the CL the twig I am thinking of is blurred in the f/4 image due to a breeze, so I will use the f/5.6 image instead.  Otherwise, it looks like all the rest can be the same.  Using f/5.6 on the 21mm gives it significant advantage in DOF, so the details in the wall look much crisper, but you'll just have to ignore that and look at the branch itself.  

 

Here is the first set...

 

Share this post


Link to post
Share on other sites

Here is the second batch...  No caveats on this one.  All lenses mounted on the CL were shot at f/4 or minimum aperture, whichever was higher; all lenses mounted on full frame were shot at f/5.6.

 

Share this post


Link to post
Share on other sites

The problem with lookin at the edge of the field here (almost the upper right corner) should be fairly obvious.  The scenario is not nearly as well controlled.  Specifically, the focus point.  This branch is well behind the point of best focus for all of the camera/lens combos.  So you get a fair amount of softening just due to the limited depth of field.  But, did I set the focus the exact same in all cases?  I certainly tried to, but based on the variability in the detail in the background wall (bokeh), I doubt very much that I succeeded.  I will point out a few things, though...

 

The 21mm SEM mounted on the CL had an advantage in DOF due to needing to show you an f/5.6 image rather than an f/4 image.  Even so, the corners look a little soft on the CL.  Perhaps the earlier supposition is correct--that there is thicker cover glass on the CL which may pose some minor challenges with certain wide angle lenses?  The 21mm 'Lux on the CL didn't seem to do any better.  There could be some truth to that, but, honestly, the differences are pretty small considering these are 100% crops from pretty close to the corner of the field of view and are outside the plane of focus.

 

The 35mm Summilux TL that I had earlier said didn't do well in the center because I slightly missed the focus... Well, it did better in the corner.  I had managed to slightly back focus that lens, and since this branch is behind the pig the image benefitted for this particular crop.  It's arguably the best of the images, but I think it's just because of the point of focus.

 

The 24-90 zoom on the SL did particularly well in this set of samples.  Perhaps it has a slightly flatter field than most?  Or perhaps it actually has some reverse field curvature?  Or perhaps the AF is very slightly back focused?  Or perhaps it's just a magic lens? 

 

Net result?  I don't think you can read too much into this, but it's at least interesting to look at and think about.  I've still got the same conclusion I had earlier, though... For the vast majority of subjects I don't think I would choose any of these lenses or cameras over another based on pure image quality.  At least not at moderate apertures with plenty of light.  Make your selections based on:

 

- Angle of view required

- Narrowness of the DOF desired

- Availability of light (larger chips of a given megapixel count are generally going to do better at higher ISO's)

- Requirement for autofocus and overall shooting speed

- Requirement for image stabilization

 

The only exceptions I can think of are kind of specialized.  For example, if shooting architecture I would want the lens with the least distortion at a given focal length (preferably least distortion straight out of the lens, not after distortion correction since that does somewhat reduce corner sharpness in particular, and I might care about the corners in a landscape shot).  There are also subjects where the character of the bokeh is important, so I might choose one lens over another.  

 

Basically, though, I wouldn't be afraid to use any of the combinations I tested even with the highest technical image quality as a desired outcome.  I'd rather spend the time worrying about where to stand, what perspective I want, what moment I hope to capture, how to handle colors, how to handle exposure, what the light is doing, where I want the edges of the frame, where I want the subject, what story I'm trying to convey or what emotion I am trying to express... All the things that actually make a photograph good.

Share this post


Link to post
Share on other sites

Jared ...... you have my deepest sympathy....

 

I have carried out this exercise for a number of Leica cameras in the past and it is a thankless task...... achieving absolute repeatability and valid comparisons is very difficult.

 

My findings were that there was little difference between similar resolution sensors ...... certainly not enough to show up in normal usage.

 

The older 28mm and 35mm (2 & 2.8) perform poorly wide open in the peripheries on digital sensors not specifically designed for M lenses (ie. everything outside the M stable). 

 

Everything you have found is what I would have expected from my tinkering with M, SL, TL2 and CL. 

 

Yeah, controlling the variables in this sort of testing is nearly impossible.  I'd need to be shooting a brick wall, for example, rather than a 3D scene if I really wanted to figure out how the camera/lens handled corner sharpness.  I'd have to use a flat frame if I wanted to correct for differences in vignetting (or purposely decide not to).  I'd have to use more controlled lighting rather than variable cloud cover.  Sheesh.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...