What will the S2 actually DO?

[robmac]

Interesting - so lens copy-specific performance 'adjustments' vs targets are flashed to the given lens's ROM for in-camera correction at time of shot. Makes sense.

We're talking about using manual-control non-S2 glass on an S2 via adapters. The S2 glass, lacking an manual aperture control (like Contax 645) would be impractical to use on any other body.

 

[Guy Mancuso]

I think the real question being brought up is lens flaws like CA, Purple fringing, distortion, vignetting and flaws such as these and how are they truly being handled.

There are two options here

1 Build a lens that has absolutely no flaws. My take cost a absolute fotune to do and bigger than a house to do it.

2. Is correct these flaws in the raw processor in camera and wrap them in a DNG to correct automatically. Basically same as Hassy and Phase but doing it before it hits the raw processing step , another words in camera like a jpeg, only corrections are applied to the raw.

These are unanswered questions that some folks have. Not sure their is a public answer for them either.

 

[LJL]

Leica doesn't do these types of compromises, however, and reserves in-camera corrections for a lesser class of problems, focus shift and perhaps one or two others, like natural vignetting. One problem they cannot solve in this way is wallet-ache...

Carsten,
At the risk of getting the "dead horse" seal of approval from Bob .....this is what I have been asking about. Leica says it does not need to do any corrections. Then it says it does corrections for focus shift and aperture and ??? They do not talk about them as "compromises", and frankly neither does Hasselblad or now Phase, since they are trying to build in software corrections of their own.

I think a valid point that Rob brings up is that we do not know what they are doing or not doing comprehensively. The place where that really starts to matter, to me at least, is if Leica is asking nearly double what Hasselblad or Phase may be asking for certain lenses that both say need software correction, and Leica's similar lens may be doing a lot of its own correction in their system that none of us know anything about. Not saying they are or are not. Just asking what they are doing or not doing. I surely would not be a happy camper to have to pay some really high price for a lens that in reality is not delivering anything more than a much less costly version by Hassy or Phase that requires some automated software correction in Phocus or C1, both of which are provided to the customer. Leica may be doing the same sort of thing in-camera, which I think a lot of folks think would be fine, but we have no way of knowing right now, and yes, there is a huge wallet-ache that is not going away. So which would you rather have....a $3-4K lens that requires you to use some provided software correction when processing, or a $5-7K lens that may be virtually equivalent in output, but does the magic in camera because there is no other dedicated software provided?

LJ

 

[tashley]

Maybe I am naive but to me this is blindingly obvious:

If they measure each lens for focus shift in the factory at every aperture, then have that lens tell any S2 body on which it is mounted how much it needs to tweak focus from what AF has told it, at the precise moment of stop-down for exposure, then there is nothing going on with the image file itself, at all. Consequently the there is no trade off between requirements in image processing.

If there are compromises in the lens design such as those resulting in CA, distortion of field, vignetting, which need software correction (and I don't care whether those software corrections are pre-RAW or optional in RAW development) then there ARE tradeoffs in image quality during that processing. If you correct for vigging in software, you increase the noise in the corners because you are effectively adding exposure compensation in the corners. If you correct for sharpness falloff in corners you add noise in the corners because you add sharpening. If you correct for CA you distort colour and sharpness accuracy. If you correct field distortion you add interpolation, which is the same as upresing selectively.

So: if the S2 only does stuff to the lens at stop-down to compensate for focus shift, then I think that's really smart and it forces fewer other compromises on the lens design. But anything that requires resampling, sharpening, selective exposure compensation, whatever, IS costing IQ because it is trying to fool the eye into getting more from the pixel data.

If the S2 does notably less of this than Hassy et al (and as we all well know Phase does it too) then it will give a purer image and one that more accurately reflects what the lens projected onto the sensor. That's what I'd pay for because, frankly, I don't like using C1 to 'correct' lens failings such as the corner softness of the Phase 28D. The results, to my eye, look crude at larger print sizes off a P45+ and that means that I am not getting the best that can be got.

That is where the S2's killer value proposition might well be.

My 2 cents.

 

[LJL]

Tim, this is exactly one of the points, but we have no way to know for sure until actual specs and system are released, and then we also have no way to really test things, as S-lenses will only work on the S2. Leica saying that its optical calculations are such that there are no image artifacts, suggests, but does not prove that there are no other corrections needed. This is coming from the same folks that said that IR was....oh never mind on that last ;-)

LJ

 

[carstenw]

Leica says it does not need to do any corrections. Then it says it does corrections for focus shift and aperture and ???

Well, you are unsure about the exact point that I just cleared up Leica says it doesn't correct "image artifacts" in software. Now we just need to wait and see if that is really true. I guess it will be, others are skeptical. That is just the way the cookie crumbles.

Aw man, now I want a cookie, but I have none!

 

[LJL]

Carsten,
Not really....Leica says, in your quote: "The optic calculation incorporates all the factors and requirements of digital technology, meaning that software to correct image artifacts becomes superfluous."

That does not really say they do any in-camera correction or not. They simply suggest that other software to make corrections is not needed. That does not tell me whether Leica has built nearly flawless glass that hits every mark, or if they have included software correction into the camera programming and it takes its cue from the lens about what it needs to do. I guess I am just not sure of what they mean by "optic calculation".....lens design?.....or lens information handling in processing design?

LJ

 

[thomas]

Tim, this is exactly one of the points, but we have no way to know for sure until actual specs and system are released, and then we also have no way to really test things

who cares? Either the images are clear and sharp edge to edge (and hold well even when uprezzing) or they are not. Maybe Leica makes the best semi medium format lenses... and than it's fine. Or they invented the best (i.e. not visible) in camera software correction... and than that's fine as well.

I think the corrections are just focus calibration, no software correction. In current DSLRs there is micro adjustment for AF lenses. I think Leica provides that mirco adjustment ex factory, nothing else.

 

[carstenw]

Carsten,
Not really....Leica says, in your quote: "The optic calculation incorporates all the factors and requirements of digital technology, meaning that software to correct image artifacts becomes superfluous."

That does not really say they do any in-camera correction or not. They simply suggest that other software to make corrections is not needed.

Strictly speaking, they don't say that they don't use software correction for image artifacts, it just says they don't need to

But really, what it does imply is that no image artifacts, i.e. visible flaws, will be left in the image, nor will they be corrected by software. After that we can discuss if the lens is perfect, or if they use software to correct non-image-artifacts, in either case a less interesting issue.

The interesting issue is the verbal sparring with Hasselblad. The interesting comparison will eventually be the Hasselblad 28mm versus the Leica 24mm, and the battle will be fought in the corners.

My personal feeling is that the difference will be visible, but that the Hasselblad will be good enough, including its corrections, that very few people will care.

 

[LJL]

Thomas,
Guess I am just trying to understand some things that do make a difference on the cost side of things. If Leica is making flawless glass for the S2, needing no corrections except focus shift or whatever, then the very high price tags on them could be justified. On the other hand, if all they are doing is building lenses similar to others, and charging a whole lot more for each lens, when the same results might be achieved with software used outside the camera, versus done in-camera, it does make a big difference. How many copies of proprietary software run outside the camera do you need, and how much would that cost? (Hasselblad and Phase already answered that question, and it is a lot less than the possible several thousand pound/euro/dollars being tacked onto each lens that Leica is offering.)

So maybe there are some folks that do care, and the information to help them decide is not being made available. Phase and Hasselblad have been very up front about what may be needed for their lenses, such as the 28mm. Leica only says you need no extra software with their lenses on their camera. Does not tell me if the lenses are delivering flawlessly or if the camera is using software to make corrections like others are doing. The cost differential, however, is not trivial.

LJ

 

[robmac]

Also depends what you call 'software'. Like Clinton (always liked the guy) re: M.L. and 'not having sex' - define 'sex'.

As LJL states does the optic calc mean just from one end of lens to the other, or does it extend into the camera firmware? We were lead to believe it was the former, now we know it is the latter - at LEAST for shift. My guess, and we'll likely never know, is it extends or will extend for as yet released lenses beyond that.

As with Hassy or Phase, if it does, so what? End IQ is all that matters. After all, regardless of Hassy, Phase or Leica, we're talking about fine-tuning or wringing the last Nth percentage of performance out of very, very, very good lenses here.

However, what does matter is denying you do it when you do, using the supposed lack of it as a marketing tool and charging a hefty premium vs those that openly 'do it' under the guise that it's justified by the fact that you don't need to 'do it' when part of the lens proceeds are earned by the 'software' designers every time you tap the shutter.

 

[LJL]

But really, what it does imply is that no image artifacts, i.e. visible flaws, will be left in the image, nor will they be corrected by software.

Where does it say that last part in any of the published information? I must have missed that completely.

LJ

 

[cmb_]

Rob Stevens once demonstrated that there was in camera correction for vignetting with the DMR (for the 19mm if I remember correctly) by shooting the same scene (his light table) with the ROM contacts masked and unmasked.

But as Thomas states, who cares - it is all nice if there is no degradation to the image file.

 

[LJL]

Rob Stevens once demonstrated that there was in camera correction for vignetting with the DMR (for the 19mm if I remember correctly) by shooting the same scene (his light table) with the ROM contacts masked and unmasked.

But as Thomas states, who cares - it is all nice if there is no degradation to the image file.

O.K., but then how does that address the issues Tim raised about noise, extrapolation, etc., as real issues that occur when the image file is subjected to alteration to correct those things? In other words, is there "no degradation" to the image file or not?

Hey, an inquiring mind wants to know :angel: .....as it does impact the cost argument :argue:

LJ

 

[carstenw]

Where does it say that last part in any of the published information? I must have missed that completely.

We are still talking about the same quote: "superfluous" means just that, in this context.

 

[thomas]

If Leica is making flawless glass for the S2, needing no corrections except focus shift or whatever, then the very high price tags on them could be justified. On the other hand, if all they are doing is building lenses similar to others, and charging a whole lot more for each lens, when the same results might be achieved with software used outside the camera, versus done in-camera, it does make a big difference.

That's a valid standpoint. But I think it's hard to tell which particular shortcoming of a lens interfers with another. At these prices the lenses still can have some minor deficiencies... otherwise they would cost far more.
But still... if the IQ is brilliant by any account and with all lenses at all apertures than it's still unique as no other system offers that - either way how they do it.

 

[robmac]

And the M8 uses lens coding - but part of the DMR sales argument wasn't that no corrections were needed. They just did it, and if they forget to mention it, so what? Given the DMR example leads further credence to the fact that it goes beyond shift correction.

As was said earlier, who cares? All that matters is end IQ, right? So why is Solms using the 'holier than thou' approach for part of their S2 marketing-speak when all that matters is end IQ? Because it creates an impression in the minds of potential buyers.

The lenses, on average MAY be better, uncorrected than their equivalent competitors (you're certainly paying for them to be), but we'll never know how much was glass and how much was glass+ 'software'.

That said, when the time comes it will be H3DII-39 (or I guess 50 ;>) vs. Phase vs. S2 prints, processed thru Phocus, C1 and LR respectively that will determine if the premium is justified or not. Should be an interesting Fall/Winter.

 

[LJL]

That's a valid standpoint. But I think it's hard to tell which particular shortcoming of a lens interfers with another. At these prices the lenses still can have some minor deficiencies... otherwise they would cost far more.
But still... if the IQ is brilliant by any account and with all lenses at all apertures than it's still unique as no other system offers that - either way how they do it.

Agreed. We have yet to see this brilliant IQ on all lenses at all apertures from anything in the S-system, as nothing is out there yet (aside from what folks may have seen in demos of one lens, or the small JPEGs displayed by fashion shooter on the Leica site).

LJ

 

[Guy Mancuso]

Bottom line guys is how do you correct for distortion since this is one area that normally no lens on the planet can fully give 0 percent distortion with a super wide. I can and Hassy folks can correct for it outside the camera . How is leica able to correct for this outside the camera when they have no dedicated software to handle it or is there lens built with zero distortion. Same with CA their 24 lux at 6k has it until you use a program like C1 to correct for it. See where I am going here. If they say superfluous than that tells me one of two things it's already being done in the raw or I will leave the second to your imagination. Now I fully will say this could be a secret and we may never know and companies do have their secrets and that is fine . But if like LJ said I am going to put out 6k for a piece of glass than what truly makes that different than others and what am I buying for 6k.

 

[thomas]

As Leica claims so self-assured the very best IQ the comparision should include a 39MP back on a view camera with Rodenstock HRs and Schneider Digitars.