Sean Reid has just published an extensive review of SL 50

[tom0511]

Advertisement (gone after registration)

I wonder how much influence the slightly soft outside outside corners have for typical subjects shot with a 50/1.4 lens. IMO for such a lens its quite important to  have good resolution in the center wide open, otherwise we dont really need an f1.4 lens.

[Advertisement]

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

[jaapv]

That is precisely the wrong method.  You're suggesting photographers apply corrections on top of corrections, thus losing even more of the image they composed and and even more resolution near the image border.  That's like wearing a swimsuit over a tuxedo.  Or wearing red lipstick on top of green lipstick.  These are very sub-optimal methods, making a simple thing more complicated, and giving a worse result.  Not the Leica way, IMO.

you have it wrong; Leica's corrections are integrated in the lens design. Throwing those out will deteriorate the file. The optimal solution would be Jared's, but I don't see that happening in a hurry, mine is second best, assuming the photographer takes the post processing into account when taking the shot, worst is throwing out part the designed-in lens corrections.

[ramarren]

Take a look at the best "driver's cars" or what those of us who love driving qualify as such and you'll find exactly these types of requests. Power steering is not present in the Alfa 4C for example. ABS is great for controlled braking but non-existent for race cars and wouldn't be present if I was designing a pure driver's car. Catalytic converters come off for any performance application as they offer nothing but emission control.

I would prefer LR offer an option to remove corrections but it's not a big deal to me at this point. I would also prefer the cost of a lens be less if the underlying optics are not producing the perceived quality. An automotive example would be I prefer a great sounding and well designed engine in a car with excellent acoustic tuning to engine noise pumped in through the speakers (now present in many cars because the engines just don't actually sound great).

I did some experiments. If you go into the LR with an SL24-90 image and check the Lens Corrections panel, the default state is corrections off and a little notifier tells you that the DNG file has corrections built in that have already been applied. If you turn on profile corrections, that indicator goes off but LR has no correction file for the SL24-90 lens ... I wonder if it is still applying the lens corrections. I didn't have time to hunt up a 24mm frame to check against, where they would be most apparent.

[zlatkob]

you have it wrong; Leica's corrections are integrated in the lens design. Throwing those out will deteriorate the file. The optimal solution would be Jared's, but I don't see that happening in a hurry, mine is second best, assuming the photographer takes the post processing into account when taking the shot, worst is throwing out part the designed-in lens corrections.

 

You agree that the "optimal solution would be Jared's".  That is exactly what I've been saying.  That is Sean Reid's solution too.  Quoting Jared above:  "Personally, I think what Sean proposed would be optimal--have the distortion correction, vignetting correction, and chromatic aberration correction be applied by default when imported into the software, but be something I could turn off one-by-one if my particular image would benefit from it."

[zlatkob]

I did some experiments. If you go into the LR with an SL24-90 image and check the Lens Corrections panel, the default state is corrections off and a little notifier tells you that the DNG file has corrections built in that have already been applied. If you turn on profile corrections, that indicator goes off but LR has no correction file for the SL24-90 lens ... I wonder if it is still applying the lens corrections. I didn't have time to hunt up a 24mm frame to check against, where they would be most apparent.

 

The profile corrections are applied automatically.  If they weren't you would notice an immediate difference between "Enable Profile Corrections" being checked vs. unchecked.  I checked a sample DNG shot at 24mm and there is no difference at all between "Enable Profile Corrections" being checked vs. unchecked.

[jaapv]

You agree that the "optimal solution would be Jared's".  That is exactly what I've been saying.  That is Sean Reid's solution too.  Quoting Jared above:  "Personally, I think what Sean proposed would be optimal--have the distortion correction, vignetting correction, and chromatic aberration correction be applied by default when imported into the software, but be something I could turn off one-by-one if my particular image would benefit from it."

However, even if it would work in practice, it still promotes the incorrect idea that digital corrections are somehow meant to correct a flawed optical design. In reality, the lens is designed holistically, i.e. the optical corrections are made with the digital corrections in mind, thus arriving at a higher level. Without the digital part the whole lens is compromised. You would not think of removing a lens element  from a fully optically corrected lens to obtain a certain effect, why amputate part of the design now?

[meerec]

Advertisement (gone after registration)

You agree that the "optimal solution would be Jared's".  That is exactly what I've been saying.  That is Sean Reid's solution too.  Quoting Jared above:  "Personally, I think what Sean proposed would be optimal--have the distortion correction, vignetting correction, and chromatic aberration correction be applied by default when imported into the software, but be something I could turn off one-by-one if my particular image would benefit from it."

This approach is very general and you could argue all lens manufacturers should agree to open it up and should work with Adobe to have this capability exposed in Lightroom. It will never happen.

[ramarren]

The profile corrections are applied automatically.  If they weren't you would notice an immediate difference between "Enable Profile Corrections" being checked vs. unchecked.  I checked a sample DNG shot at 24mm and there is no difference at all between "Enable Profile Corrections" being checked vs. unchecked.

 

 

Yes, further tests proved this. 

 

However, I processed 24 and 50 mm exposures with the same lens in both Lightroom and in dcraw, which does not honor any corrections at all. The 24mm frame showed obvious rectilinear corrections, as expected for an ultrawide FoV, and the corner stretching makes the image look correct and identical to a similar exposure made with an Elmar-M 24mm f/3.8, without the SL24-90's uncorrected vignetting and extreme barrel distortion at that setting. The 50mm capture shows virtually no rectilinear corrections at all: there is no stretching or distortion of circular objects at the edges of the frame, there's only a small hint of flattening in the very center of the frame. More important in both exposures between uncorrected and corrected outputs are corrections to chromatic aberration and evenness of illumination at the corners and edges.

 

So again, I conclude there no point to all the bellyaching. Leaving a frame uncorrected is identical to leaving in aberrations and does not stretch or smear an image with a 50mm focal length. Corrections that do such things are evident with much shorter focal lengths, and the results there simply produce the usual flat-field 2D distortion of a well-corrected wide angle lens, which is dependent upon the geometry of the capture's projection, not the rectilinear aberrations in a given lens. 

[thighslapper]

Well I have to say there is frankly some odd logic and arguments being used being in this thread. 

 

Lenses are a means to an end ...... not an end in themselves, so the MTF charts, aberrations and how they are corrected are all of entirely secondary importance.

 

And as usual, the conditions of usage of the lens are ignored ....... 

 

Central contrast, clarity and resolution is what most people want most of the time ..... just what are you routinely photographing with a 50mm lens at f1.4 that needs pin sharp edge resolution ..... or helpful barrel distortion for some numpty lurking at the edge of the frame ?

 

Some folk here seem to think that if Leica cannot produce a lens that is perfect in even the most unusual photographic scenarios then it is some way deficient.

 

With the SL you are spoilt for choice with a huge range of R and M series manual lenses ....... so if AF is not your prime consideration there is plenty other 50's you can use.

 

Also, you have to consider that this is an AF LENS ...... which requires small light optical focussing elements that can be moved accurately and fast, which considerably complicates the optical optimisation and aberration correction compared to 'simple' MF lenses. 

 

I've been sufficiently irritated by this that you have cost me $39 to Read the Reid and see what he actually says. From my usage of the lens I would agree with most of his conclusions. His opinions about 'enforced' digital corrections are purely that ..... opinions, and like many above I think offering 'options' are fairly impractical and of minimal practical benefit. I am poorer .... and marginally better informed. At least his website is not the usability nightmare it was before .......

[ramarren]

You pique my curiosity to go read his site with that "not the usability nightmare is was before" comment. But I resist paying $39 for something that seems to have little actual value to me.

 

I went out and took a walk beside our nearly flooded Guadalupe River instead. Only brought the iPhone but got at least one decent photo...

[Jared]

Well I have to say there is frankly some odd logic and arguments being used being in this thread. 

 

Lenses are a means to an end ...... not an end in themselves, so the MTF charts, aberrations and how they are corrected are all of entirely secondary importance.

 

And as usual, the conditions of usage of the lens are ignored ....... 

 

Central contrast, clarity and resolution is what most people want most of the time ..... just what are to routinely photographing with a 50mm lens at f1.4 that needs pin sharp edge resolution ..... or helpful barrel distortion for some numpty lurking at the edge of the frame ?

 

Some folk here seem to think that if Leica cannot produce a lens that is perfect in even the most unusual photographic scenarios then it is some way deficient.

 

With the SL you are spoilt for choice with a huge range of R and M series manual lenses ....... so if AF is not your prime consideration there is plenty other 50's you can use.

 

Also, you have to consider that this is an AF LENS ...... which requires small light optical focussing elements that can be moved accurately and fast, which considerably complicates the optical optimisation and aberration correction compared to 'simple' MF lenses. 

 

I've been sufficiently irritated by this that you have cost me $39 to Read the Reid and see what he actually says. From my usage of the lens I would agree with most of his conclusions. His opinions about 'enforced' digital corrections are purely that ..... opinions, and like many above I think offering 'options' are fairly impractical and of minimal practical benefit. I am poorer .... and marginally better informed. At least his website is not the usability nightmare it was before .......

 

 

I agree that the optical performance looks like it is quite impressive.  I don't own one yet, though... On a waiting list.  That's why I'm resorting to MTF charts, not because I have a love for them.  And I certainly don't crave tack sharp performance in the corners with the lens wide open--that would be silly since depth of field at f/1.4 is going to erode sharpness away from the main point of focus long before lens quality would.  I am a little surprised that the MTF charts show performance falling off at f/5.6 as quickly as it does away from the center of the field.  That's not the design choice they made with any of their other 50's as far as I can tell.  I'm wondering whether Leica see this lens being used a bit differently from their other 50's?  Or perhaps it comes back to what Peter Karbe was talking about with the upcoming Summicrons SL's--the fact that they give a depth of field presentation more like an f/1.4 lens than an f/2 lens.  I wonder if the same is true for this Summilux--that the off axis astigmatism that is present even stopped down has a real-world effect similar to a shallower DOF?  I'll be interested when the data sheets on the Summicrons are released to see if they share this same design approach.

 

To your point about Leica optical quality in general and our tendency as a group to think that a lens that isn't perfect in all ways is somehow deficient... I agree.  Many of us--me included--often forget the old Voltaire idea that perfect is the enemy of good.  As a reminder, I would recommend anyone compare the MTF chart for any of the current Leica 50mm's to the best that Nikon, Canon, and Fuji have to offer.  It's not really very close.  (I'd recommend comparing images rather than MTF charts, but that's harder to arrange.)

 

I do have a question for you since you already own the 50mm Summilux SL.  How do you find the handling?  Is the autofocus quick enough for non-static situations?  How is the light fall-off at f/1.4 (after profile correction in Lightroom, presumably)?  I have heard that the manual focus is accurate and easy which makes sense for a lens with narrow depth of field wide open.  How does it balance?  Do you find it overly bulky?  Enough that you would choose a different 50mm if you didn't need AF and weather proofing?

 

Thanks for your feedback.

 

- Jared

[thighslapper]

It is shorter than the 28-90 but does not seem smaller in use ..... and is noticeably heavier. No vignetting of note at 1.4, MF is easy and accurate. 

 

AF is accurate and fast-ish in good light and contrasty subjects, but not a patch on the 24-90 that locked instantly on my posted sheep photos in fairly dense fog. 

 

I think it would be a bit hit and miss for rapidly moving subjects ..... but have yet to find anything to try it on. 

 

A lot of money to spend just for AF ....... particularly if you have other good 50mm R or M lenses. 

[tom0511]

The things is that IMO for anything other than still subjects AF can make a big difference. The faster the lens and the shallower the DOF the more usefull AF can be.

In regards of the option to switch correction off-I would appreciate the option but I understand also Leicas approach: The option would offer something where maybe 2% of users would benefit from but opens the risk that quite some people get a wrong/bad impression about the lens.

I think Leicas strategy is to offer not too many options, the same is valid for the user interface. This means less options, but more simplicity. No room for user faults etc.

[zlatkob]

Yes, further tests proved this. 

 

However, I processed 24 and 50 mm exposures with the same lens in both Lightroom and in dcraw, which does not honor any corrections at all. The 24mm frame showed obvious rectilinear corrections, as expected for an ultrawide FoV, and the corner stretching makes the image look correct and identical to a similar exposure made with an Elmar-M 24mm f/3.8, without the SL24-90's uncorrected vignetting and extreme barrel distortion at that setting. The 50mm capture shows virtually no rectilinear corrections at all: there is no stretching or distortion of circular objects at the edges of the frame, there's only a small hint of flattening in the very center of the frame. More important in both exposures between uncorrected and corrected outputs are corrections to chromatic aberration and evenness of illumination at the corners and edges.

 

So again, I conclude there no point to all the bellyaching. Leaving a frame uncorrected is identical to leaving in aberrations and does not stretch or smear an image with a 50mm focal length. Corrections that do such things are evident with much shorter focal lengths, and the results there simply produce the usual flat-field 2D distortion of a well-corrected wide angle lens, which is dependent upon the geometry of the capture's projection, not the rectilinear aberrations in a given lens. 

 

Your conclusion is based on testing the 50mm focal length of the 24-90mm lens.  But Sean's review is of the 50mm lens — a different lens.

[zlatkob]

This approach is very general and you could argue all lens manufacturers should agree to open it up and should work with Adobe to have this capability exposed in Lightroom. It will never happen.

It did happen with Canon lenses — they have no enforced correction.

[zlatkob]

However, even if it would work in practice, it still promotes the incorrect idea that digital corrections are somehow meant to correct a flawed optical design. In reality, the lens is designed holistically, i.e. the optical corrections are made with the digital corrections in mind, thus arriving at a higher level. Without the digital part the whole lens is compromised. You would not think of removing a lens element  from a fully optically corrected lens to obtain a certain effect, why amputate part of the design now?

It would work in practice.  It's not a question of "if it would work".  It works for M lenses and for lenses from various other brands.

 

The option to turn off digital corrections doesn't "promote" an idea, at least not an incorrect idea.  It merely restores a tool that some photographers find useful.  

 

A lens that loses its digital corrections doesn't become a "flawed optical design"; rather, it becomes uncorrected.  Likewise, a digitally correct lens isn't "unflawed" — it's still the lens, but with digital corrections.  If people feel uneasy about lenses lenses having some inherent flaws — well, sorry, that is a fact of life.

 

You agreed that having the option to turn of digital correction  is "the optimal solution".  If  it's optimal, then it is optimal.  Your solution is not optimal; it is worse.

 

Your assumption that a digitally corrected lens arrives at a "higher level" assumes that photographers don't photograph groups of people with some people near the edge of the frame.  In such photos, your claimed "higher level" produces a worse photograph.  

 

The uncorrected lens is not "compromised".  On the contrary, it is more suitable for certain subjects.  I.e., superior.  For those subjects, it is the digitally corrected lens that is compromised by the software.

[ramarren]

Your conclusion is based on testing the 50mm focal length of the 24-90mm lens.  But Sean's review is of the 50mm lens — a different lens.

 

 

So send me a file made with the SL50 lens and I'll see if there is any change between uncorrected vs corrected in terms of rectilinear correction using the same test mechanism.

 

It is likely that a zoom lens would need more corrections than a prime lens at this focal length. I really don't care, however: the principle is that a lens designed with software corrections to produce its best results are always best used with the intended, supplied corrections. The argument that you 'must have' the native, uncorrected output of such a lens because it is "better" is fallacious.

 

Of course, I have 50mm R lenses that Leica provides lens profiles for. I'll check and see if those profiles alter the geometric corrections at the edges of the frame.

[zlatkob]

So send me a file made with the SL50 lens and I'll see if there is any change between uncorrected vs corrected in terms of rectilinear correction using the same test mechanism.

 

It is likely that a zoom lens would need more corrections than a prime lens at this focal length. I really don't care, however: the principle is that a lens designed with software corrections to produce its best results are always best used with the intended, supplied corrections. The argument that you 'must have' the native, uncorrected output of such a lens because it is "better" is fallacious.

 

Of course, I have 50mm R lenses that Leica provides lens profiles for. I'll check and see if those profiles alter the geometric corrections at the edges of the frame.

 

Perhaps someone who has the SL50 lens can send you a sample DNG.  One review (jupitersnake.com) compares a corrected vs. uncorrected file, but doesn't offer a downloadable DNG.  The difference in barrel distortion is obvious.

 

Mid-range zooms tend to have little or no barrel distortion around the 50mm focal length.  I believe 50mm is about where they start to shift to slight pincushion distortion.  A 50mm prime lens, on the other hand, tends to have some barrel distortion. 

 

I didn't say "must have".  You said that.  I'm talking about the optimal method.  It's important and it's better (for certain types of photos), but not a "must have".  It's important enough for real world actual photography that Sean Reid thought it worth stating in his review.  And a couple of people in this thread have supported the idea.  Even Jaap called having the option to turn off corrections the optimal solution.

 

Don't bother yourself about checking the 50mm R lens, SL50 or any other lens.  If it hasn't been important to your photography, then it's not about to become important just because I said so.  As I said before, it depends on the subject matter and on one's priorities for a specific photo.  For many photographers, it will never be an issue.  I would still hope that photographers would support other photographers in having practical tools, even if they don't need a particular tool themselves.

[Advertisement]

Hallo,

Look here: Leica L-Mount

[thighslapper]

Try as I might, I cannot find a single 'real world' photo I have taken with the SL 50 that could be made available as a DNG that would show the differences between uncorrected and uncorrected images ....... I just do not take pictures of flat things that cover the whole frame at f1.4 - 2.8  ...... or other subjects with small enough aperture to make sure most of the frame is in focus ........ and even then most of it would be sky in the corners and entirely featureless.  

[ramarren]

I have located a dozen or so SL50 raw files that I've run through dcraw and Lightroom. There is a minor correction for barrel distortion evident. That said, it doesn't 'stretch' the rendering at corners and edges relative to proper projection with an uncorrected 50mm lens, like the Summicron-R 50mm f/2. It makes the uncorrected Summicron-R 50mm lens and the Summilux-SL 50mm rendering pretty much the same. 

 

I cannot for the life of me figure out all what your palaver on this topic is about. As I said before, if you don't like the way the lens renders, just buy some other lens.