Puts weighs in on Q…and more

[pgk]

Advertisement (gone after registration)

Right, the results speak for themselves and they are quite good from what I've seen online.  I'm not looking for fault with the camera.  Rather, I think the change in attitudes is interesting.  Paradigm shifts occur from time to time, and they seem to happen almost without notice.  We've slipped through a time hole — pushed, it seems, by rapid tech changes.  The talk used to be about optical craftsmanship, as recalled in post #14's "Leica lenses are superior to any in the market".  Now the talk has shifted to "All that matters is the results".

Its the camera. A new camera with a fixed, fast lens has to make the most of that lens. It may well be that the optical/software engineers determined that a combination of optical traits combined with software adjustments would produce the 'best results'. Take the same lens (which obviously you can't) and put it on another camera and the purely optical results may not be all that wondrous, but that's not the point, the point is that the lens is fixed and the camera produces good results.

 

With the M lenses, their inherent optical excellence and superb design might be enhanced by software, but its not a given. I would prefer to make my own decisions regarding optical correction so I'll stick to my M lenses, but I can see where lens design may be heading and no doubt the technical qualities of imagery will benefit. Some people's attitudes will shift, others will dig their heels in. I'm probably in the latter category I have to say.

[AlanG]

T

One hunch I have is that Sony skimped a little too much on its E mount, for usage with FF, which has required extreme angles for light rays at the corners. Thus highly corrected designs like the FE 35 f1.4 are massive and unusually long, as a byproduct of trying to minimise distortion and achieve corner sharpness. Even with the amount of work gone into that lens it still needs a tonne of software correction at the corners.

 

This is also a reason why Olympus and Panasonic find it fairly "straightforward" to achieve corner to corner sharpness with their system due to the generous throat size.

 

How would the "throat" size impact this? 

[zlatkob]

I dont understand why this bothers you so much. We are talking about two completely different products. A $6000 lens and a lens thats part of a system who's  optics cost 300% less....im not sure what you expected, but if you think Leica should be delivering a M quality Summilux in a camera that cost far less your expectations are not realistic. If you don't like what Leica is doing then don't buy it...its pretty simple. I think you made your point, time to move on rather than repeating the same thing over and over. Personally I could care less about the internal corrections, all I care about is the results...and on that note I am extremely pleased with my little $4000 camera and lens.

 

BTW- I also have four Summilux lenses and two M bodies...there are clearly differences in what each delivers. The Q has its advantages and disadvantages as do the M240 and Monochrome...but I love all of them.

 

Wow, you really, really didn't understand a word of what I wrote.  Not one word.  Sorry, I won't try to explain it again.  I tried and don't want to keep repeating; it will get annoying to readers who did understand the point, if anyone did.  I didn't think it was so nuanced, but you've missed the point twice now, so I give up.

Edited June 20, 2015 by zlatkob

[k-hawinkler]

Hi there zlatkob.  This thread has been more interesting than the title implied to me.  

I agree with your observation.  Thanks for stating it so clearly.

For me this implies that with the Q Leica enters direct competition with other camera makers.

So, for the Q Leica leaves their traditional rangefinder world behind, for good or for worse.

 

I am amazed though that Leica chose to accept and correct such a huge geometric lens deficiency in firmware/software.  

The point I am leading to is that lenses seem to become more specific to a particular camera, including FW/SW corrections.

Just thinking of the sensor cover glass thickness issue, amplified by the need to maintain usability of traditional M lenses.

 

But even this challenge might eventually be partially addressed through better/thinner sensor layers for WA lenses.

Also through a curved sensor that Sony seems to have in the works.

So speaking only for myself and my interest in interchangeable lenses.

I am not interested in buying such lenses that don't maintain a minimum standard for optical excellence.

In particular, if those lenses can be used on multiple systems.

[pgk]

I am amazed though that Leica chose to accept and correct such a huge geometric lens deficiency in firmware/software.  

Why? I'm curious. If it enables high image quality in a fixed lens camera at a price which they might otherwise have struggled with, what's the problem?

[mjh]

I am amazed though that Leica chose to accept and correct such a huge geometric lens deficiency in firmware/software.  

The point I am leading to is that lenses seem to become more specific to a particular camera, including FW/SW corrections.

We had the same debate a year ago when the T camera system was introduced and people recoiled in horror when they realised that Leica did rely on digital corrections. And now to everyone’s surprise it turns out they did the same with the Q. So yes, Leica does things differently when the lenses in question are either non-interchangeable or won’t be used with any other camera.

[bernie.lcf]

Advertisement (gone after registration)

I really don't know why everybody acts like the earth came to an end. DNG level correction has been there since the M8. With the M8 it is only corners (darkening ang magenta shift compensation), but geometric correction was introduced with the X already. 

 

Leica isn't even hiding it because if you look inside the DNG file, you will see the correction profile. It's all there in plain sight. 

 

Besides, calling the Q distortion huge is an exaggeration compared to all the other lenses out there. 

[Peter H]

Bernie, I don't think it would matter even if the distortion was unprecedentedly enormous. As so many have said, it's the quality of the images that matters. 

 

I suppose it does matter if you want to think about the uncorrected lens in non-photographic and competitive terms, but if you're interested in photographs, and the camera/lens unit as a machine for making photos,  I really can't see what possible difference it makes as long as the various parts work together well.

[pico]

So what we've come back to is what photography is about!

 

Photography is not about its technology.

.

[jonoslack]

I'm not sure how you can say "it has no distortion, unless you remove half the correction".  The correction is not a property of a lens, but a software method of addressing a problem with the lens.  It appears to have extreme distortion, requiring correction ... http://www.fredmiranda.com/forum/topic/1370077/3#13057657

HI There

It isn't a 'problem with the lens'  -  it's a design consideration. Hasselblad started this game a few years ago (with their 28 I think). If you accept some distortion (and correct it after the event) it gives you a lot of other options with respect to image quality. If you want a 'perfect' lens without the need for lens corrections then you'll likely end up with something the size and weight of the Otus

 

Clearly there are philosophical arguments as to whether this approach is a good one or not (you pays your money and you make your choice) but it's a respectable decision and a good way of designing a lens which is both small and high quality (which the Q lens certainly is - even though the distortion is considerable when uncorrected). 

[jaapv]

As I said before: lens design is about shifting aberrations. In the case of lenses with hybrid corrections the optical design is meant to transform the aberrations that can best be  handled optically into aberrations that are best corrected digitally, resulting in a lens that is better corrected  than it would be if limited to either of the methods. It would be extremely stupid to deny a lens designer part of his toolkit.

[AlanG]

Bernie, I don't think it would matter even if the distortion was unprecedentedly enormous. As so many have said, it's the quality of the images that matters. 

 

I suppose it does matter if you want to think about the uncorrected lens in non-photographic and competitive terms, but if you're interested in photographs, and the camera/lens unit as a machine for making photos,  I really can't see what possible difference it makes as long as the various parts work together well.

 

I agree. Also consider that when the designers are able to "correct" a lens using optics alone, they are also making various compromises. The benefit for us as photographers is that by correcting various things in even our cheap lenses, we get results that previously required much more expensive lenses or were even impossible. Such as a distortion free, c/a free, non vignetting 24-105 zoom.  And the 60X zoom on the Lumix that I bought for my stepson is way better than I expected due to the digital correction.

Edited June 21, 2015 by AlanG

[stefans4]

As I said before: lens design is about shifting aberrations. In the case of lenses with hybrid corrections the optical design is meant to transform the aberrations that can best be handled optically into aberrations that are best corrected digitally, resulting in a lens that is better corrected  than it would be if limited to either of the methods. It would be extremely stupid to deny a lens designer part of his toolkit.

 

I think I agree with what I think you're saying - but may not quite have understood what you're _really_ saying. I would understand better if your statement went something like: "... is meant to transform the aberrations that are difficult to handle optically into aberrations that are best corrected digitally..." If they are best handled optically they should be handled optically, shouldn't they? And those that have an excessive "cost" in size etc. if corrected optically may better be transformed into digital correction space... 

 

Just asking...

[gpwhite]

HI There

It isn't a 'problem with the lens'  -  it's a design consideration. Hasselblad started this game a few years ago (with their 28 I think). If you accept some distortion (and correct it after the event) it gives you a lot of other options with respect to image quality. If you want a 'perfect' lens without the need for lens corrections then you'll likely end up with something the size and weight of the Otus

 

Clearly there are philosophical arguments as to whether this approach is a good one or not (you pays your money and you make your choice) but it's a respectable decision and a good way of designing a lens which is both small and high quality (which the Q lens certainly is - even though the distortion is considerable when uncorrected). 

 

Jono, you have shot both the Q and the 28 SX extensively. You also have demonstrable skill as a photographer and as a reviewer.

 

OK, now that I have set you up, my concern about the Q is that the high-quality aspects of the DNG image are really in the central circle of 24x36, regardless of aperture (i.e., Leica does much better than that with M and R lenses). And the Q files look less 3D than from an M, at least in shots with my hands. I feel Leica should have delivered a more complete optic for $4,250. The edges and corners just never really come around, even at f/5.6!

 

Do you find the Q firmware essentially makes its 28 Summilux up to approximate a 28 SX on an M, even at say a pedestrian f/4?

[jaapv]

I think I agree with what I think you're saying - but may not quite have understood what you're _really_ saying. I would understand better if your statement went something like: "... is meant to transform the aberrations that are difficult to handle optically into aberrations that are best corrected digitally..." If they are best handled optically they should be handled optically, shouldn't they? And those that have an excessive "cost" in size etc. if corrected optically may better be transformed into digital correction space...

Just asking...

Well, yes. One could say can be handled by the method that will have the least negative consequences for the rest of the design, I suppose.

A thought model: A designer corrects for instance CA -which is not very easy to handle digitally- optically but this may increase for instance distortion (or other aberrations). Trying to correct distortion optically may in turn cause other aberrations - including CA- to increase. Better to correct CA with lenses and distortion digitally in that case. It retains the optimal CA correction and correcting distortion digitally will not deteriorate the rest of the design.

Not that I could do this work myself in a thousand years, just to give an idea why the two cannot judged separately.

 

http://www.astrosurf.com/luxorion/report-aberrations.htm

 

http://demonstrations.wolfram.com/WavefrontMapsAndProfilesOfSeidelAberrations/

 

You have to download a CDF player for the last one.

[imants]

Photography is not about its technology.

 

 But it sure helps

[thighslapper]

From what I can see the manipulations of the Q images are no worse than in the X-Vario ...... where W/A distortion etc. is equally quite marked and solved by profile correction and cropping.

 

You cannot circumvent the laws of optics and physics ........ you can either have a perfectly optically corrected system which will be big and expensive ...... or a more compact and cheaper one that uses other technology to get round the major issues. 

 

If a restaurant produces an excellent meal at a good price you don't moan about what's going on in the kitchen ....... 

[rosuna]

Are S lenses' distortion corrected by software or optically?

[pgk]

Clearly there are philosophical arguments as to whether this approach is a good one or not (you pays your money and you make your choice) but it's a respectable decision and a good way of designing a lens which is both small and high quality.....

Whilst I agree that its a viable way of designing lenses for fixed lens cameras or new camera systems, my concern isn't purely 'philosophical'. IF the same design parameters are eventually applied to say an M lens for some reason (or any other interchangeable lens), then the lens will perform well only if fitted to a camera which has the ability to apply appropriate corrections. So if there is a way of applying them this is fine (like 6-bit coding or manual entry of lens in use via a menu). But such lenses would clearly not be back compatible with older cameras which could not apply corrections, nor would they be useful performers on more 'manual' style bodies. So I have concerns over this type of design - if it ever gets applied to interchangeable lens systems (which it may well if its seen as a method of delivering 'better' lenses at a lower price point).

[Advertisement]

Hallo,

Look here: Leica Q

[IkarusJohn]

So, if digital corrections are bad, are optical corrections okay?  Aspherical lenses, floating elements, partial dispersion, multiple elements in groups?  Would a pinhole be the only, purely optically correct camera?