Lensrentals: System MTF comparison, Canon vs Nikon

Interesting article at lensrentals - Roger Cicala tested:

Canon 24-70 2.8 MK II on 5DIII

vs

Nikon 24-70 2.8 on D800E

(and the Tamron 24-70 2.8 on both.)

.

Couple interesting points, I thought:

*  New Canon 24-70 2.8 MKII bested the older Nikon 24-70 2.8 in the optical bench test (not mounted on any cameras.)

*  On the D800E, both the Nikon and the Tamron at f/2.8 resolved more than the Canon at f/4.

*  Worse lenses on a higher-res sensor is almost always going to be better (from a resolution standpoint) than great lenses on a lower-res sensor.

Link to article HERE

.

--
Here are a few of my favorite things...
---> http://www.flickr.com/photos/95095968@N00/sets/72157626171532197/

This also refutes the tired saying, "You need better glass for if you have a D800." The D800 will improve the performance of all glass. Of course, the better the glass, the better the improvement.

It also makes you wonder if a D5200 might out resolve a D700, under bright light conditions.

I never believed it when people claim you need a better lens for D800. If a lens is sharp with D700 I expect that lens to be sharper with D800 or equally as sharp as with D700. It makes no sense any other way.

Seeing where both the Canon and Nikon lenses are aimed at (press/reportage tools) those minor difference in details are almost of non importance, especially in comparison with the D800.
It's the build and focus speed that counts mostly.
Test both lenses (the Canon and Nikon) on the D4 and 1DX makes it at least a bit more logic test instead of commercial infotainment at a low level.

As myself, I'm using the Nikkor24-70mm F/2.8G now almost since it's sale appearance on the D300/D3/D3S and sometimes on the D800. And in all those years I had never ever one failure, the lens is always working flawlessly, it's blazingly fast and you can bump (and I did many times) in objects without even scratching it. The only thing I had to replace (due to temperature differences) is the zoom ring. I'm now on my third. I need a lens which I can trust for my work, the 24-70mm F2/8G absolutely is.

Michel

--
- To observe without evaluation is the highest form of human intelligence -
http://www.fotopropaganda.com
http://www.fotopropaganda.com/blog
http://www.flickr.com/photos/9240992@N05

All three lens used in the test are high quality and the Canon ($2,299) and Nikon ($1,886) lens are pro grade.  The only thing that Lensrentals looked at was resolution.  There are other factors that affect the "performance" of a lens, e.g., vignetting, barrell distortion, chromatic aberration, edge to edge sharpness.  The only way to tell if the D800 improves the performance of all glass is to put some consumer grade lens on the D800 and see how they fare with that 36 mp sensor.

Should I have said "improve the RESOLUTION performance of any lens" which is specifically what he was testing and specifically the D800's advantage.  Of course the D800 isn't going to improve the barrel distortion of a lens...or the vignetting...or the contrast.  Yes, it does improve the "sharpness", edge or center, which is another way of saying resolution.

Look at the nasty bokeh on the Tamron:

http://photographylife.com/reviews/tamron-24-70mm-f2-8

It looks like a pin wheel... I wonder if the object itself had some issue, you can see some artifacting w the Nikon at f4...

Was thinking of buying one till I saw these images... Anyone else see similar?

--
Yucel
Glamour Photographer

This is typical "onion" distortion that many multi-lens zooms suffer. It is only really noticeable when looking at defocused spheres of light. It isn't going to show up in a lot of pictures, and if it really bothered you, can be fixed in post.

Discussion of onion bokeh

I really don't like how they presented those examples (defocused).

--
Craig Bennett
http://www.craigbennettphotography.com

The whole "lens is more important than the body" mythology is gradually fading away, aided by the stellar results with various lenses on the D800, and good riddance to it.  Put any decent lens at all on a D800 and live it up!

Agreed. I've been saying this for ages.

I use two cheap lenses - the 35/1.8G (DX) and the 85/1.8D on my d800 and they are both great, especially stopped down. They still resolve a lot more than they would on a d700.

The whole idea of a lens suddenly getting "worse" when put onto a high res body is idiotic and I was so disappointed to see that idea touted so many times on the internet.

--
My travel photography blog - http://www.frescoglobe.com

D800 would magnify those flaws.  Hence the thought that you should use good to excellent lens with a D800.

Having had my D800E now for about 8 months I tend to agree with that.

Sharpness is hardly an issue with any lens. For me its more about the rendering the lens gives that I rate as important.

Coatings, ED glass, good design add up to a nice rendering or a not so nice rendering, much like bokeh is regarded by quality.

Perhaps it is somewhat subjective but I find the 24-70 gives some of the nicest renderings of any of my Nikon lenses and the 85mm F1.8g whilst super sharp can at times give a very harsh rendering which I find unpleasant. Other times it seems great.

Also I find the Samyang 24mm F1.4 gives a marginally nicer image than the 24-70. Very good for a "cheapish" lens.

Greg.

Only at the pixel peeping level...but you have to remember the D800 puts more pixels into the same area.

The point is, sized the same, the D800 will NEVER be less sharp on ANY lens.  It will always be immeasurably better, or measurably better...depending on the sharpness of the lens it sits behind.  The sharper the lens, the more the difference.  Of course, it isn't just related to lens sharpness, camera shake and DOF will have to be tended to, if you want to get the full potential out of the D800.

Newer lens beats older lens.  Whilst good news, it's what you'd hope for.

Higher resolution cameras resolves more than lower resolution camera with similar lenses.  Again, this is what I'd expect and hope for.

I don't know how many pixels it takes in a camera to exceed the resolution of these lenses, but I'm confident that the lenses still have the edge.  That being the case, then this is also the result you'd expect.

Does anyone have figures for the theoretical maximum resolution of any of these lenses (in megapixels)?

Yes, any lens will perform better on the D800 than on the D600. Think a little about it.

--
Renato. http://www.flickr.com/photos/rhlpedrosa/ OnExposure member http://www.onexposure.net/ Good shooting and good luck (after Ed Murrow)

I'd guess diffraction is the limiting factor, it'll enter the picture before any other effect for these modern lenses, at least in the central part of image.

--
Renato. http://www.flickr.com/photos/rhlpedrosa/ OnExposure member http://www.onexposure.net/ Good shooting and good luck (after Ed Murrow)

As I wrote in the other thread concerning the same Lensrentals article, it is a good time to introduce ourselves to the practical benefits of additional resolution. The 36MP of the D800 is often talked about as a strain to the lenses, when in fact it is a blessing. The fact that lens deficiencies are seen more readily at 100% is nothing but a red herring to a lot of the folks, but also a great help for raw converters that have better information to apply lens corrections on.

However, I've been wondering about one particular example for a while. The Nikkor 28mm f/1.8 was tested by Photozone.de on both the D3x and D7000, which means 24MP FX and 16MP DX sensors, respectively. Here are the resolution charts:

D3x:

And the D7000:

I can only assume they used the same lens for both tests, which is why I'm so utterly baffled by the border area results at the larger apertures. First, can someone please verify that multiplying the resolution value (line width / picture height) of the DX tests by 1.51x (that is how much longer the FX sensor is to the DX sensor) is a valid way to compare the numbers? Because ultimately, the resolution is often measured in line widths per millimetre, and the value is then simply multiplied by the length of the sensor to obtain the "per image height" value.

Then, let's assume the DX 'extreme border' corresponds to 'border' on the FX sensor. At f/1.8 the extreme border resolution on the DX sensor is round 1600 lw/ph. If we use that to extrapolate the 'border' resolution of an FX sensor with same pixel pitch (roughly 36MP) we get ~1600 * 1.51 = 2420 lw/ph. However, the 'border' resolution on the D3x is already ~2990 lw/ph. Does not compute.

Are my calculations completely off or are they simpy different samples of the lens?

Here are the links to the tests, first D3X and then D7000:

http://www.photozone.de/nikon_ff/751-nikkorafs2818ff?start=1

http://www.photozone.de/nikon--nikkor-aps-c-lens-tests/750-nikkorafs2818dx?start=1

--
regards
Janne Mankila, Finland

Of course, I agree with you 100%. But take a look at my post below.

--
regards
Janne Mankila, Finland

your answer is in the text about the resolution on the D7000. The 28 has a wavy resolutio with a drop off at about the dx edges, but pics up again closer to the fx edges. A well known phenomena and an one that quite a few of us owners of that lens aren't all to happy about.