rscheffler wrote:
IIRC TheSuede slammed Leica/Kodak for their use of a very cheap quality cover glass/IR filter on the M9's sensor and that apparently the M240's sensor has a much higher quality one. This, combined with the redesigned microlenses, may correlate to the reports from early M240 users of better edge sharpness with various wide angle lenses. Therefore my interpretation is the opposite - Leica changed the M240's sensor topping stack for the better.
Your interpretation about the sensor topping stack might be true, but it remains a sad fact that leica sacrificed the microlens design of the M in order to make it more compatible with R telephoto lenses.
Matt Grum wrote:
I believe from posts earlier in this thread that the IR absorption layer is between the two layers of the OLPF, and that the second layer of the OLPF doubles as the sensor coverglass. Removing it leaves a very delicate sensor surface underneath.
You would have to then source a replacement IR filter, but if you did this then yes you would reduce smearing with symmetrical lenses, but at the expense of performance with lenses designed for digital. The refractive properties of the filter stack cause a degree of field curvature that lens designers now take into account (attempting to induce the opposite curvature in the lens itself)....Show more →
Yes the sandwiching of the IR filter appears to be a common according to Canon / Nikon teardowns and literature..
That leaves two approaches to fixing smearing. Either remove all the layers except the first birefringent/sensor cover and then use a regular IR lens-filter on the front of the lens the same way that M8 shooters do. This is non-destructive and called a 'full-spectrum' conversion by IR guys like lifePixel who offer it as a commercial service.
Alternatively try to substitute that first OLPF layer which is bonded to the sensor carrier with the IR absorption glass which is a far more delicate and non-reversible action. theSuade says it can be removed using a sharp knife and experience!
rscheffler wrote:
It will be interesting to see what combination of features and performance Zeiss conjures up for their upcoming line of manual focus lenses for this platform... Will they lean more towards their historical tendency of 'lens relaxation' through somewhat larger optical formulas, or will they aim to take advantage of the lack of AF components and truly shrink lens size (length) down to something akin to RF lenses? (I suspect girth will remain constant with the lens mount size.)
High performance wides are pretty big on Nikon and Canon
Do we think that Zeiss could build a solid manual focus, electronically connected, FE class 24/2 or 24/2.8 in this size:
Too much to ask for?
If these new lenses do rival RF lenses for size and offer optical performance along the lines of the two FE lenses we've seen thus far, I suspect it will make RF lenses on this platform much less relevant. There seems to be a big opportunity here if Zeiss can find the magical combination of size and performance to satisfy a large enough segment of this fairly niche market. I also expect we'll see something from Voigtlander.
Zeiss has a lot to gain and nothing to lose unless FE class bodies go no where for Sony, which seems unlikely to me. Zeiss must have a fairly good idea of the potential given their close relationship with Sony (which will include royalty data coming to them).
At the moment, they have zero direct competition in this segment. There is competition in the rangefinder lens area.
Makten wrote:
What we have yet too see is a real A/B comparison between the cameras with the same lenses. I don't want to buy the a7R just to find out that the smearing is exactly the same as with the a7. I have zero interest in 36 megapixels.
Agreed that we need to see an A/B comparison, but I would venture and predict the A7 will have less smearing due to the lower pixel density.
rscheffler wrote:
IIRC TheSuede slammed Leica/Kodak for their use of a very cheap quality cover glass/IR filter on the M9's sensor and that apparently the M240's sensor has a much higher quality one. This, combined with the redesigned microlenses, may correlate to the reports from early M240 users of better edge sharpness with various wide angle lenses. Therefore my interpretation is the opposite - Leica changed the M240's sensor topping stack for the better.
edwardkaraa wrote:
Your interpretation about the sensor topping stack might be true, but it remains a sad fact that leica sacrificed the microlens design of the M in order to make it more compatible with R telephoto lenses.
That's definitely true, especially in light of these Sony cameras being better R solutions than the M240. Leica should have pulled out the stops for RF wides, since this is where the system is primarily used (50mm and wider).
Toothwalker wrote:
What if future generations of sensors require a different filter stack for improved performance?
^ This.
Lenses are made for the longer run. Film RF lenses were made for the physics around film, and that changed, so formulas need to change. E mount + high resolution digital is a given for the next N years; who runs the optics show - the lens department or the sensor department? Do the lens guys lay out some basic parameters - here's how we'll design them - you sensor guys make it fit to work? Or is it the other way around?
Zeiss has a lot to gain and nothing to lose unless FE class bodies go no where for Sony, which seems unlikely to me. Zeiss must have a fairly good idea of the potential given their close relationship with Sony (which will include royalty data coming to them).
At the moment, they have zero direct competition in this segment. There is competition in the rangefinder lens area.
Nice to see that people are getting it. I have no doubt that Zeiss will be giving Leica users a run for their money. The body from a value and technical perspective already surpasses the M240. I have no doubt that Zeiss will deliver. The native mounts produced by Zeiss will be all the rage within the next 12 months. Leica has already lost its edge. But I'm sure most refuse to believe that, their tune will change once the first MF Zeiss lenses show up.
Matt Grum wrote:
, but at the expense of performance with lenses designed for digital. The refractive properties of the filter stack cause a degree of field curvature that lens designers now take into account (attempting to induce the opposite curvature in the lens itself).
Is it certain that the optical stack induces field curvature in modern lenses that are of telecentric design like those designed for digital? My understanding is that the field-curvature effect is limited to non-telecentric rays taking a path that means they encounter more glass in the corners due to their steep-angle.
edwardkaraa wrote:
Agreed that we need to see an A/B comparison, but I would venture and predict the A7 will have less smearing due to the lower pixel density.
Why would pixel density have anything to do with smearing?
Toothwalker wrote:
What if future generations of sensors require a different filter stack for improved performance?
I doubt that it will have to be thicker, so just laminate it with clear glass to make the properties the same, if that's possible.
Makten wrote:
Why would pixel density have anything to do with smearing?
Well, sensor toppings having the same thickness, smearing should be equal in a given area, except that the 36mp sensor has more pixels to record the smearing in that area.
edwardkaraa wrote:
Well, sensor toppings having the same thickness, smearing should be equal in a given area, except that the 36mp sensor has more pixels to record the smearing in that area.
That's exactly why I questioned that you said smearing would be worse. It's like saying that more pixels demand better lenses.
michaelwatkins wrote:
High performance wides are pretty big on Nikon and Canon
+1
The ZE/ZF 21/2.8 isn't exactly tiny either.
+1 @ Ron, i.e. where the sliding scale of size vs. quality design will land in vying for the discretionary $$$. I suspect the future native mounts for FF mirrorless can be smaller than current SLR offerings (which range from Oly 21/3.5 to ZE 21/2.8, et al) ... but just comparing the Oly vs. the Zeiss, we can clearly see that the latitude of optical design is a matter of choices/compromises as they both serve the same 24x36 with the same mirror impediment.
With the mirror out of the way, it does open up some additional latitude ... but 24x36 remains unchanged. and the closer you get to the film plane and the shorter your FL, the steeper your angles of incidence ... as the tenets of the trig will never change. Only the choices at how to incorporate a designed approach at which compromises do/don't get incorporated do.
Even the mighty Zeiss with its magic mustache can only do so much to "bend" the angles of incidence to try and "cheat" the trigonometry involved. I have every confidence that Zeiss has the wherewithal to give us a good offering ... but trig is trig and the laws of physics will ultimately trump our wants & desires, no matter how much we plead our case for some magic glass to be uber-diminutive and still cover 24x36 properly (and play well with micro-lenses and stacks of glass) for pixel peeping critical IQ performance.
rscheffler wrote:
IIRC TheSuede slammed Leica/Kodak for their use of a very cheap quality cover glass/IR filter on the M9's sensor and that apparently the M240's sensor has a much higher quality one. This, combined with the redesigned microlenses, may correlate to the reports from early M240 users of better edge sharpness with various wide angle lenses. Therefore my interpretation is the opposite - Leica changed the M240's sensor topping stack for the better.
That may very well be the case. I misunderstood Edward's post and thought that he was implying that the edge sharpness of the M240 wasn't quite as good as the M9's, which apparently isn't the case. I haven't kept up on the m240 lately to know any better.
edwardkaraa wrote:
Your interpretation about the sensor topping stack might be true, but it remains a sad fact that leica sacrificed the microlens design of the M in order to make it more compatible with R telephoto lenses.
I'm not sure it is that they sacrificed anything for R lenses, but, rather, they just wanted to improve noise performance. Apparently, better micro lenses improve the overall amount of light hitting the pixel well, but the edge light rays cause worse color shift with better micro lenses. That's also apparently why MFDBs with micro lenses (most don't have them,) have better noise performance, but worse color shift. I'd also imagine that is what is going on with the A7R. It's a trade off that we're not used to seeing with DSLRs, because their light rays are more tele centric and color shift is rarely an issue.
Zeiss has a lot to gain and nothing to lose unless FE class bodies go no where for Sony, which seems unlikely to me. Zeiss must have a fairly good idea of the potential given their close relationship with Sony (which will include royalty data coming to them).
At the moment, they have zero direct competition in this segment. There is competition in the rangefinder lens area.
Judging by the size of the Toiut lenses, I can't imagine that the Zeiss lenses for FE will be small, unless they're slow. Sure, they could maybe get rid of some of the diameter by removing the AF motor, but I doubt they'll be tiny. The ZF/ZE lenses aren't really any smaller than their Nikon and Canon counterparts that have AF motors.
douglasf13 wrote:
I'm not sure it is that they sacrificed anything for R lenses, but, rather, they just wanted to improve noise performance. Apparently, better micro lenses improve the overall amount of light hitting the pixel well, but the edge light rays cause worse color shift with better micro lenses. That's also apparently why MFDBs with micro lenses (most don't have them,) have better noise performance, but worse color shift. I'd also imagine that is what is going on with the A7R. It's a trade off that we're not used to seeing with DSLRs, because their light rays are more tele centric and color shift is rarely an issue....Show more →
But that was mentioned by Leica people themselves. Offset microlenses like those in the M9 would cause problems with R telephotos, so they came up with the solution of oblong microlenses that in theory do not require to be offset. From Leica's perspective, the worse color shift is corrected with lens profiles anyway, and guys like me who don't use Leica M glass are not really their problem.
edwardkaraa wrote:
But that was mentioned by Leica people themselves. Offset microlenses like those in the M9 would cause problems with R telephotos, so they came up with the solution of oblong microlenses that in theory do not require to be offset. From Leica's perspective, the worse color shift is corrected with lens profiles anyway, and guys like me who don't use Leica M glass are not really their problem.
I'm not sure that I totally buy it, and it would depend on which Leica person said such a thing. There could potentially be a slight vignetting increase with R teles, using an M lens micro lens setup, but, from what I understand from theSuede, it wouldn't be much.
Either way, I'm not surprised that a new micro lens design was used, considering that it is a new sensor tech and company, but it does stink for non-Leica M lens users.
It's like saying that more pixels demand better lenses.