Nikon D7100 preview updated with a look at effect of OLPF omission

Found it!

http://www.dpreview.com/reviews/sigmasd1/19

The Leica M9 still has a bayer matrix... Gotta see it to believe it! (just make sure you look at the center of the frame, where the lens can keep up!)

JDThomas:

Why not use a Zeiss on a Nikon, they're much better in low light high ISO situations than so called ED Nikon glass? And Zeiss lenses don't show the imperfections that Nikon lenses show when mounted on the D7100. I'm sure of both points.

As for mounting those M mount Zeiss lenses on a Leica, okay, but those are different lenses, and the reason for using a D7100 instead of a M9 or M8 is that the D7100 is less audible and much better at high ISO work than either Leica.

(I assume you're not talking about mounting Zeiss M mount glass on a 35mm film rangefinder (Leica, Contax, Minolta, Konica, etc)

sorry to say that DX17-55/2.8 is no good lens.

The 17-55 WAS a good lens when paired with earlier cameras. It's STILL a good lens when stopped down to f/5.6 on the D7100.

Heck, it's not really that bad when the image is downsized to about 12MP (but then why buy a 24MP camera if you have to downsize to get the best performance).

Really my only point is that Nikon needs to redesign the 17-55 f/2.8 to keep up with their cameras. And they likely are.

24-85VR tack sharp on DX corner to corner.

Reilly Diefenbach:

Lenses can be plenty sharp, but have poor colour performance; that's like a lot of Nikon ED lenses.

"Tack"? I've never thought tacks particularly sharp, maybe carpet tacks. Why not razor sharp? (Perhaps the term refers to something else other than actual tacks, anyhow it seems kind of silly to use it when describing lens performance.)

think 17-55/2.8 is a pro-lens of good mechanical quality only.

Nikon gears can be categoried into two groups, pre-2007 are all sh!t (the company should be called Shikon) and a totally new company was born late that year.

the new Nikon is a superman 5 yo. it doesn't have a DX17-55/2.8. if they are planning their own product for DX it should be DX16-50/1.8 at about the same size and price as 24-70/2.8.

No point in even continuing this discussion. You are completely derailing it with wacky talk.

16-50 f/1.8? Haha.

I'm only saying that 17-55/2.8 is a bad lens at that high price and I think it's a good thing we put it and the old Nikon behind us.

what a beautiful world.

@HowaboutRAW
Comparing the D7100 and K5iis at iso 6400 on this site, the K5iis looks well ahead (much cleaner) to my eyes.

Regarding AA-less cameras, I apreciate the increase of detail visible on large prints at at high screen manifications, but there is also a fair amount of false detail (moire banding) that is visible even at normal viewing.... I don't like it, so I'm not persuaded! AA-less sensors seem to work better the higher the resolution of the camera.

At the end of the resolution race, we'll all be using prime lens at f/5.6 or f/8....

They do provide that option, via a battery holder you insert into the battery grip add-on.

Josh,

I don't disagree with you, but it sounds like you are just chasing specs, instead of deciding what you need.

I HAVE made quite a few prints from a 2 megapixel camera, the first I owned, a Kodak, and they are as sharp as a D800 at 4x6. They are softer at 8x10, but they are still sharp at 8x10, especially at the right viewing distance.

Here's another thing I've tried. I shot my 16 megapixel GH2 at 8 megapixels (all cameras will downsize like this, just choose your size in the menu), then up-rezzed the 8 megapixel photos and compared. THERE WAS NO DIFFERENCE. Sometimes the 16 megapixel photo was slightly better, but sometimes the 8 megapixel photo was better! And the 8 megapixel photo was always smaller and easier to handle.

Yes, YMMV, but really, with this camera compared to previous cameras, I don't see the difference.

I hate to say it to you all but like it or not, we are ramping towards an apex where diffraction will eat any further gains in resolution. Its a simple fact that the airy disks are only so small. These high density sensors are limiting at f9 or so. Not exactly a brick wall yet, but soon your resolving power will be at the limits of the lens more than anything. The D800 is already severely challenging for glass to fully resolve. Diffraction limits of f5.6 would be nearly useless on full frame. You would throw all that resolution away and be forced to have huge files with no more information in them. 35mm was a film size. Lenses only had to resolve maybe 20mp max due to the grain size in the film. Its already been noted that the gains in fine detail on the k-5 IIs are lost above f8, though overall the image stays sharper. That's on a 16mp aps-c camera, which scaled to FF equals the d800. The 7100 pictures look on the level of FF detail, so there are gains yet to be had, but clearly the MP war will become meaningless soon if you ask me.

the optical resolution limit is the first radius of Airy disc. we need at least two 2x2 pixels or 4 photosites within this radius to read it out for Bayer sensors but that's dump calculation. practically we need only 3 (this number comes from lowest/comfortable/absolute resolution in DPR's camera tests).

we can make consumer grade lenses to reach perfect at f/2.8 but f/5.6 sounds good for lenses that most of us can afford. the radius at f/5.6 for 555nm wavelength is about 3.83um for 3 photosites, that's 1.26um pitch or over 500MP on a 35mm format sensor.

there could be 2000MP for perfect f/2.8 lenses, and it could be a mad 5000MP for "violet blue". but that's only calculation. I bet there will be little difference beyound 200MP.

but still I think we will get more and more pixels well beyond 5000MP, not for resolution but other image qualities. whether we should call them pixels is open to debate but to me, a perfect pixels is one that measures a single photon only.

DPReview needs to give us a macro shot of that globe to put this to rest. If you look at the Pentax 645D and look at the globe it seems to indicate there *are* markings that represent waves or current, but then if you look at Africa you see the same waves on the actual land continent! If true I haven't been comparing real detail for a long time and just how one moire looks over another!

It's halftone (4-colour process), and the patterns are also there (and only slightly less distinct) in the D5200 crops. Without using a significantly higher or lower resolution, or blurring to the point that the dots can't be resolved at all, moiré will happen on any sensor, OLPF or not, when two grids of slightly different sizes collide (in this case, the sensor and the halftone). That will be a constant in digital photography until a pseudorandom layout of sensels happens (and someone figures out how to extract colour information from it without paying a high sensitivity tax).

Those crosshatch lines are some of the best things I've seen on that studio setup to judge optical resolution and high ISO settings. Only a few cameras can resolve those lines correctly, and only while set on RAW files, JPEG can't resolve the detail at all. The Nikon D800E shows the lines clearly on RAW, but on JPEG they disappear into a blurred mush, just like every other camera when set on JPEG. From best to worst, the D800E can resolve the lines clearly on RAW, the D800 too but slightly less clearly, the D7100 shows them but less clearly than the D800 and with lots of moire especially in the Arabian Sea, and the D600 misses most of them in the ocean except the ones near the Seychelles. The Canon 5D Mk3 does slightly better than the D600 but with moire, and the Canon 6D is prety close to the Nikon D600 but with more moire than the D600 or 5D Mk3. The Pentax 645D is very similar to the Nikon D800E, but the 645D can resolve the crosshatches around Tanzania that the D800E can't.

I still have D200. I know where you are coming from. D300 is an awesome piece of equipment. That said, the sensor tech is lightyears ahead at this point. Unless you never shoot anything over ISO100, you should get a new camera.

Yes, thank you DPR for helping instigate new improvements, innovations, experimentation and progress. Without it, everything stagnates, and stagnation becomes so... grimy and old.

So here's to marketing and innovation! With it, in just 200 years we went from pushing a plow to landing on the moon!

Let's rather talk photography here. :)
If you start with the plow and the moon, I'd say that pushing a plow makes much more sense to me than landing on the moon...

A perfect example of 'thinking inside the envelope' brendon. You need to watch a video I saw yesterday, where a car door opens underneath the car, without changing the underbody clearance. I thought it was CGI'ed. It looks impossible. But it's not. The door disappears just as slick as rolling the window down.

Following the mechanical principle of the hybrid viewfinder of Fuji X100/s and X-Pro 1, I think.

D800E image quality is absolutely fabulous. But better use the mirror prerelease even for handheld images (unless using LiveView) if highest resolution is required (and possible due to low iso).
24MP on DX is too much - the test results are no surprise.

what bout moire lucas? since you are hell bent on fabric etc, you do know that most medium format cameras do not have AA built in either, yet they are heavily used for indoor fashion photography. by the way there are a million tricks out there in the open on how to remove them - if they ever show up in 0.001% of your shots.

@Devendra

First morie is not the only artifact that can be caused by lack of an AA filter.

High end indoor fashion shoots where Medium format is used are very controlled. A DSLR is designed for a more candid, faster paced, less controlled environment where you can't always make adjustments or shoot again to get rid of moire.

As far as removing it in post goes, when it is possible it is usually very time consuming if you want to do a good job. Fine if you are a pro retoucher paid by the hour working on high end fashion stuff where sometimes 8 hours is spent on the face alone, but a deal beaker for many armatures and pretty much all pro portrait/wedding photographers who's retouching is generally done as a cost of sale rather than as a for profit service.

I've noticed a huge difference in detail on a tripod on even my older k-7. Anything less than 1/200 has the potential for mirror movement and hand shake to rob fine detail. You may have a sharp looking shot, but even a pixel or two of blur is significant. At low shutter speeds handheld is no good unless you don't have the choice. I regularly shoot at 1/10s hand held, but high iso is already robbing detail there so its a wash. You should note that many serious photographers carry tripods regardless of their choice in camera.

Penta's mirrors are not as good as Canon or Nikon which are nothing compared with Sony. sorry to say.

Sorry Yabokkie, I did not understand what your trying to say. Blue channel fail? So as long as it does not fail, you can stop down without any loss of contrast or sharpness? I think you might be thinking about the blue pill. If it lasts for more than 4 hours, please call your doctor.

@f_stops
I don't need to re-read it. I saw the difference at f3.5 .......... but what DPReview pointed out was that it was only at this aperture that significant differences could be seen. Ergo, OVERALL,there is no SIGNIFICANT difference between the two 24Mp APS-C cams, in terms of perceived detail, when one does and the other cam doesn't have an OPLF.

@ Amadou Diallo
Thanks for the reply.

(5) Is this because of those pictures were displayed on "retina" (high pixel density) display?