T, Leica zoom lens debacle

[Margu]

http://www.dpreview.com/previews/leica-t-typ701/6

During pre-launch briefings for the T, Leica was very keen to stress the optical quality of the new lenses. Most interestingly, we heard that they relied on optical corrections, rather than software, so as to project the best possible image onto the sensor.

So we were a little surprised to note an interesting little message when processing the camera's DNG files through Adobe Camera Raw: "Built-in lens profile applied: Vario-Elmar T 1:3.5-5.6 / 18-56 ASPH.. Click for more info." Clicking indeed gives more details:

this has to be the lamest lie in photography history that leica has commited.

 

[rscheffler]

Is there any surprise about software correction? On the day or so after the launch, there were conflicting reports about whether or not there is software correction. I believe Sean Reid (I'm not a subscriber, so going by reports from others) processed T files through at least a couple raw processors, one of which does not apply/recognize software correction, and noted a fair amount of native distortion.

This post by ChromaSoft basically says the same thing...

quote:

"The DNG version is 1.3, a higher revision that that the 1.1 that most previous Leica cameras have used. There a reason for this - DNG 1.3 allows for opcodes, which Leica use for lens correction.

In the DNG I looked at, which was shot with a 18-56 Vario-Elmar lens, lens correction is done by a single "WarpRectilinear" operation in the DNG. Other lenses (or the same lens at a different focal length), might use other codes."

Maybe this irks some... I don't really care. Take a great camera/lens combo like the RX1... it also relies on software correction for its barrel distortion.

 

[mfunnell]

I believe Sean Reid (I'm not a subscriber, so going by reports from others) processed T files through at least a couple raw processors, one of which does not apply/recognize software correction, and noted a fair amount of native distortion.

I did read that on Sean Reid's web site. I'll note that he expressed no surprise about this. Given his involvement with the development of the camera (as a source of end-user input) and exposure to prototypes of the camera, he would be in a better position than most to know about such things. My interpretation (in hindsight) is that it never occurred to him that software corrections weren't being used. I vaguely recall (I'd have to re-read the article to be sure) that he commented on the optical design of the lenses reducing the amount of software correction required.

I'm not sure where this idea that "no optical corrections are required" came from. I've not read it in statements from Leica. (Then again, I've hardly been assiduous in my reading as I've no particular interest in the camera beyond it's place in "general photography news".)

...Mike

 

[rbelyell]

i'd seen the same statement quoted fron dpreview in a number of other sites at launch time. i guess theyre all either lying or misunderstood what leica told them in their 'intensive' briefings. couldnt be leica did anything wrong here...or anywhere. and even if they did, i see the point is its our fault for believing them, right? i mean its not like they trade off of delivering a higher quality product than their competitors. just a higher polished higher priced one. me like shiney!

 

[tomtofa]

I think the contempt comes not from the software correction, but that Leica promoted the lens as being rigorously corrected optically, and charges $1800 for it, based at least in part on this 'optical' correction. When in fact it has the distortion, CA, and vignetting of a typical zoom.

 

[DougFord]

Does Leica manufacture any lens that's not recognized by the camera for optical correction purposes after the fact?

 

[tomtofa]

Does Leica manufacture any lens that's not recognized by the camera for optical correction purposes after the fact?

Again, if it's after the fact correction it's not optical correction.

Fuji made a similar claim for its 23mm: "Lens distortion has been reduced to the absolute minimum using only optical rather than digital correction, thereby delivering the highest possible picture quality." Except it seems to be true for that lens.

 

[Pablito]

... in fact it has the distortion, CA, and vignetting of a typical zoom.

a LOT of barrel distortion, worse than many much cheaper lenses...

 

[CK Dexter Haven]

I'm a disinterested viewer (i have no interest in this camera), but i'm not really sure i understand the conflict here.

"relied on optical corrections, rather than software, so as to project the best possible image onto the sensor."

That doesn't mean that software corrections would not be used to 'further enhance' image quality. The above statement says it's the optical corrections (that) project the BEST POSSIBLE image onto the sensor. That sorta indicates they've gone as far as optically possible with this formulation, not that their optical design yields perfection. It's still a zoom lens. And, in this day and age, i would assume they would have software profiles for even their most recent, ASPH, supermegamoney prime Summilux, so why not 'help' a zoom?

Were we supposed to expect that this zoom is so good that they would just forget about making a correction profile for it? Does the M correct for the 50 ASPHs?

 

[Godfrey]

Tempest in a teapot.
I don't care if Leica uses voodoo to make their lenses perform as they do.

G

 

[icebear]

Tempest in a teapot.
I don't care if Leica uses voodoo to make their lenses perform as they do.

G

+1, there is no RF, so I don't really care too much - if anything then about the tea 🙄.

 

[Pablito]

I don't care if Leica uses voodoo to make their lenses perform as they do.

G

Please don't give Vodou a bad name! 😉

 

[Aristophanes]

Tempest in a teapot.
I don't care if Leica uses voodoo to make their lenses perform as they do.

G

It's an 18-55 18-56mm F3.5-5.6 zoom on an APS-C sensor.It's going to have the same characteristics of...oh, I don't know...all the rest of the 18-50ish zooms on an APS-C sensor? Sort of a generic combo.

 

[Monochrom]

wow...too much distortion...

supeose leica vario t lens shuldn´t be a typical zoom...so 1750 usd for such common flaw is someting hard to bite...

 

[YYV_146]

That is WAY too much distortion for a slow APS-C zoom.

Not that Leica didn't claim it is optically corrected, but even if Leica said nothing - that level of distortion has got to have some impact on image quality. I'm guessing 4%-5% at the wide end and more than a stop of vignetting at the extreme corners.

 

[willie_901]

Please don't give Vodou a bad name! 😉

+1



here are some pointless characters so this post goes through

 

[Roger Hicks]

That is WAY too much distortion for a slow APS-C zoom.

Not that Leica didn't claim it is optically corrected, but even if Leica said nothing - that level of distortion has got to have some impact on image quality. I'm guessing 4%-5% at the wide end and more than a stop of vignetting at the extreme corners.

Yes, probably to improve it. Leaving bigger flaws in one area makes it easier to reduce flaws in another area, optimizing the optical/software package. Or would you rather have less distortion and vignetting (both relatively easily corrected) and less resolution (impossible to correct)?

Cheers,

R.

 

[BlackXList]

I realise that the purists won't view the DLux cameras as "Real" Leicas, but this was the case with them as well, so it's not a new thing for a product with the Leica name on.

I have less than no interest in purchasing a T, I know that similar software corrections are done on the RX100 I own, and based on the output from it, I'm entirely ok with that, because the pictures are fantastic.
(Also had no problem with it on my LX3, if that's a more direct comparison).

It could be argued that the price is a bit ridiculous for the T Zoom lens, but people tend to get a bit funny if you make comments about Leica and "value", so I won't.

 

[Godfrey]

Yes, probably to improve it. Leaving bigger flaws in one area makes it easier to reduce flaws in another area, optimizing the optical/software package. Or would you rather have less distortion and vignetting (both relatively easily corrected) and less resolution (impossible to correct)?

Exactly.

Correcting simple spherical barrel distortion and light fall-off in the corners (not vignetting ... for the pedant in me) are two of the easiest, least-lossy corrections you can make, while fixing longitudinal chromatic aberration, mustache or waved-shaped rectilinear distortion, astigmatism, coma, etc, are FAR more difficult.

Just like in many other locked-loop auto-correction feedback systems, you leave in the simple aberrations while fixing the near-impossible ones because it improves the overall quality of the imaging system. These lenses will only ever be used on the camera(s) for which they are designed, so it is only the total system that matters.

It's a wonder that all these "Leica enthusiasts" and "professional reviewers" are unwilling to give Leica the benefit of the doubt and believe that they know what they're doing with respect to imaging systems and quality, look at the results rather than infer absurd things from marketing statements.

G

 

[Vobluda]

Take any lens and any camera raw and run it through raw SW that opens the fiel as it is without any corrections.
For example I use Raw Photo Processor on Mac (SW is Mac only, freeware).
Sony A7 raw when open as it is, without HEAVY in camera processing, is a raw fun 🙂
Usually under exposed, everything way off.. The flange distance is just too short. If you look at the lens via "pure" raw, the FE 35mm 2.8 and FE 55mm 1.8 are joke of the lens, unperfect in everything.
So I dont think that it is Leica to blame, it is the only way to go with short flange distance..