I fear the tough e mount could make disaster more likely with large lenses not being carefully supported: the lens or body could break. Sony's two piece metal and plastic arrangement can't be cheaper to make than a one piece affair like the replacement. I'm pretty sure it's there for a reason: the plastic layer will snap off in an emergency, leaving lens and body relatively unscathed and easy to repair. A small amount of play (so long as it's all in the same plane) is a small amount to pay for that...
John TF: How close would GigaPan + DSLR approximate these images?
For critical work when stiching have lots of overlap, and then use masks in PTgui to mask out the corners and extreme edges; that way you get not only extra pixels but are effectively using only the best part of the lens.
dbm305: Hmm. What are the chances of ZE/ZF versions? a 15-30F2 right? It covers the full frame. And he hugely expensive 'compact' primes are the same optics as the ZE/ZF lenses....
F2.8? That makes T2.9 pretty impressive given how many elements there are in it (my guess was actually F2.5; F2 was a typo..)Well the compact primes well for about five times the price of the same stills versions here, so that would make it $US5K if the same ratio held. So about the price of a relatively fancy Leica M rangefinder lens. I think some might pay.... Not me!
Hmm. What are the chances of ZE/ZF versions? a 15-30F2 right? It covers the full frame. And he hugely expensive 'compact' primes are the same optics as the ZE/ZF lenses....
Garth Wood: Maybe I've missed it somewhere in the review, but can anyone tell me definitively what the raw image bit-depth is? Discussion around the Intertubes seems to waver between 12-bit (which I expected) and 14-bit (a surprise), with at least one commenter claiming to have heard directly from Sony that it's 14-bit depth.
I own this little gem, and it's wonderful. But I don't have a solid idea of how much post-processing I can do before the image starts to break down or posterize; and the manual that came with the camera (and other official docs from Sony) is/are strangely silent on image bit-depth.
It'll be the usual Sony cRAW; it's very slightly lossily compressed, so it's hard to say what the bit depth really is: in practice it varies from area to area of the image. If I were to say average 12 that might be a bit meaningless, but maybe it helps. But it starts with 14 before compression. If you mess around with RAW digger or similar tools you might be directed to bits of the image at 100% where you can see the effects, but at the whole image level you get a smaller file size for no visible loss.
dbm305: What does "support for fast AF mean"? Is it something separate from support for the new 70-200? Could it be a general AF speed upgrade, or perhaps enabling AF in the faster burst mode?
Perhaps then it's support for the focus limiting buttons. But are you sure? I know it was listed in the same line as mention of support for the 70-200, but it seemed pretty ambiguous to me and the translations are not that terrific!
What does "support for fast AF mean"? Is it something separate from support for the new 70-200? Could it be a general AF speed upgrade, or perhaps enabling AF in the faster burst mode?
moimoi: Shutter vibration that makes blurry images, have you forgotten dpreview?
It is a HUGE con.
Hmm. I suppose if the shutter speed needed to combat camera shake is also high enough to eliminate shutter shock, then that's a way of thinking that for practical purposes it's the same issue with the same fix...
What Shawn says about shooting it like an MF camera is fair enough when it comes to camera shake; but shutter shock is another matter - it's a flaw. There are lots of potential ways of fixing it: electronic first curtain, a shutter that only goes to 1/4000 and which might not as much vibration. Maybe they couldn't do it yet, but it does count as a 'con', and a distinct con from the general fact that a very high res camera needs careful shooting. I'm still tempted though.
highwave: It looks so good on the black OM-D
Both looks and spec wise, it's more fittingly the kit lens for the OM-D than the 12-50mm ever will be
E = MCsquared?
topstuff: All very lovely, but this is'nt really an F2.8 lens is it, compared to APSC or FF?
What would be the equivalent APSC lens? A 17-50 I guess? And if my understanding is correct, would this 4/3 lens have to be a F1.4 to have the same light gathering as a F2.8 in larger sensors?
As said somewhere below, I don't think it is a simple as simply doubling the focal length to get a "FF equivalent" and keeping the max aperture the same.
I am not sure if this really a f2.8 as we know it. I am sure the people at DPR can clarify !
A lovely thing though I am sure, especially with an OMD EM5.
Of course the exposures will be identical. But the poster is still correct that four times the total light (not light per area) will be collected on an FF sensor in the larger image circle.
Of course that's what gives the better noise characteristics of the FF: total light.
So what needs to be disentangled are three things:
Depth of field: function of aperture, distance from subject, angle of view of lens and used image circle.
Total light: (and hence noise level): function of aperture, and used image circle (holding light levels and length of exposure fixed)
Exposure: function of aperture (holding light levels and length of exposure fixed).
Much of the discussion of these matters conflates the three.
ZAnton: There is already Samyang 14mm f/2.8 for almost 1/10 the price.Only real crazy techno-geeks will buy Zeiss.
Yes: which tells you something about Ken Rockwell!The Samyang is cheap; yes. It has wild but correctible distortion, yes. But it's sharper than the N 14-24 or canon 14 2.8. Less CA. Better flare control than the N. It's not just me: read photozone or Lenstip.
lucavascon: Aaaahhh... Zeiss!It looks quite a deal if compared to 2nd hand 15mm f3.5 in Contax mount!!!However I do prefer old Contax mount primes adapted to 5D. And new ones still have better taste than Nikon and Canon primes (specially Canon primes, that are disappointing me so much).For those who do not appreciate the difference between the Zeiss and other lenses, if you don't see it, it means your way, taste, and preferred light condition in taking pictures is not putting your lens in crisis. No "absolute"better in lenses and tools There is simply a tool that "fits better" in your hand for that situation
@ZAnton, I tried the Samyang 14, quite unusable for me. Far prefer using the good 15mm fisheye Sigma and than correct it with PanoTools
@thx1138, you got the point. I do not like zooms, I usually prefer zeiss and do not use filters (although sometimes, I'd like a polarizing one). But I've tried that 14-24 and I can only say that it is a fantastic, German-feeling ultrawide, I'd go for it.
What problem do you have with the Samyang? It has crazy distortion, but one of the good things about it's distortion pattern is that when you correct it with a LR profile you don't lose much angle of view. And it has extremely high resolution.
Oh well. I haven't downloaded and normalized the D800 RAW pics yet, but based on how they look (only slightly worse than the 5dIII at 1:1) I'm guessing it'll come out the same or slightly better normalised. And all those nice pixels. Still, I don't regret my 5D III: it's handily better than the II (2/3 stop in RAW? That's really pretty useful = especially if it cleans up better which is my experience so far) and it would take me three years to get as proficient with the Nikon interface, by which time who knows who'll have the nicest sensor? And I do love my 17mm TSE....
Um, they don't seem to be up yet....
ET2: "Sometimes reviews get delayed."
No kidding. The preview (forget the review) of Nex-7 was promised back 3 months ago?
The first S90 I got was just frighteningly decentred; the whole right side was smeared and the right hand corners were just a blur. I sent it back to Adorama (I think) and got another that was still noticeably decentred - soft in lower right at all apertures, but not so much that it was too much of a pain in normal shooting.