HELP! Lens or camera soft??? 5DMkII compared to Nikon D2x

[NickJB]

I've just gone thru the process of moving over to a Canon 5D MkII from a Nikon D2X. Reason - Canon has the range of lens that make sense for my use (pro, environmental portraiture).

The 5D MkII arrived yesterday. So today I've been running tests and I'm somewhat concerned.
5D MkII with Canon 24-70mm f2.8
Nikon D2X with Nikon 17-55mm f2.8 DX lens

Both cameras on tripod (screwed into the body direct, not the grip on the 5D2). Zoomed to approx same coverage. Set to Aperture priority. Auto focus - set to center spot. MacBeth colour checker as target. In center of colour checker is a hair - perfect for testing sharpness, and there is some "fuzz" on the edge of the checker, also great for testing sharpness.
Both cameras are essentially "zeroed" (in camera sharpening, tone etc).
Nikon is zoomed to 55mm (35mm equivalent 85mm), Canon to 70mm
I do a round of apertures - 2.8 to 22.
Import to Lightroom -no auto adjustment.
Export to PS at 16 bit, with no adjustments, open at 100%.

The files from the 5D2 are noticeably a lot softer - actually quite dramatic - from the little hair in the center of the frame, to the fuzz on the edge of the checker, to the edge of the black cutouts around the colour swatches - softer from f2.8 all the way to f11. At f16 the difference is less noticeable and at f22 the 5D is slightly sharper.

Do you think I have a lens or camera problem here?
Anything I might be doing wrong?
I know that there is greater sharpness with the smaller sensor and the density of pixels on DX cameras, but this is making me question whether I have a bad lens or if there are issues with the camera.

I can't do a similar test with the Canon 70-200 lens as I don't have a comparable Nikon lens to test against.

Any help greatly appreciated as I'm meeting a buyer for the D2X tomorrow and I might have to postpone.

Many thanks!!

nick

 
 

[marcmccalmont]

I don't have a D2x to compare it to but....
manually focus the 5DII with live view zoomed in to 10X and aperture f9 and reshoot
is it still soft then?
a. if yes, maybe the difference in AA filters or a bad lens
b. if no, the lens needs a micro focus adjustment.

if a) do you have a 50mm f1.4 for comparison?

Let us know the results
Marc

You are starting with RAW files not jpegs?

[NickJB]

I don't have a D2x to compare it to but....
manually focus the 5DII with live view zoomed in to 10X and aperture f9 and reshoot
is it still soft then?
a) yes maybe the difference in AA filters or a bad lens
 no lens needs a micro focus adjustment.

if a) do you have a 50mm f1.4 for comparison?

Let us know the results
Marc

You are starting with RAW files not jpegs?

It looks like it's more than just something like AA filters (not that they cannot be significant) - especially as when closed right down the image does gets sharp.
Good idea to do a critical manual focus, just in case it's an autofocus issue.
Always RAW. I don't even shoot companion jpgs.

I'm kinda hoping some else jumps in and says they had the same experience with Canon L series glass and a full frame Canon sensor and that it's merely an issue of the sharpness of the small sensor in the D2X combined with a phenomenally sharp Nikon lens.

Wish I had a super sharp and fast prime to compare.

If this is just what I should expect from this lens and camera, I'm actually fine with it. Slightly disappointed, but there are thousands of professional images published every day with this camera and lens. I'm just hoping to find out if this is normal or if I've got a dud lens or off camera.

cheers!

nick

[stever]

to get the best out of the 5D2 you need to use the microadjustment, and although there are a variety of ways to check focus, i find that Lensalign is pretty simple and quick once you understand how to use it.  Most of my lenses, particularly L lenses are pretty close on focus  -- however my old 300 f4 required +15 (no wonder it was so disappointing on the 20D and 40D).  There is always the possiblility of a lens with focus error (compared to a couple other bodies, i think my 5D2 is pretty close on focus, but i tested a 50D a couple months ago that was significantly off and still have a concern that Canon may use microfocus as an excuse to loosen or not improve focus tolerance which becomes very critical on high resolution cameras)

i've heard some comment that Canon doesn't sharpen the image as much in-camera as Nikon, but don't know if this is true.  in testing, i use the LR default of 25 sharpening but frequently sharpen more than this

[NickJB]

to get the best out of the 5D2 you need to use the microadjustment, and although there are a variety of ways to check focus, i find that Lensalign is pretty simple and quick once you understand how to use it.  Most of my lenses, particularly L lenses are pretty close on focus  -- however my old 300 f4 required +15 (no wonder it was so disappointing on the 20D and 40D).  There is always the possiblility of a lens with focus error (compared to a couple other bodies, i think my 5D2 is pretty close on focus, but i tested a 50D a couple months ago that was significantly off and still have a concern that Canon may use microfocus as an excuse to loosen or not improve focus tolerance which becomes very critical on high resolution cameras)

i've heard some comment that Canon doesn't sharpen the image as much in-camera as Nikon, but don't know if this is true.  in testing, i use the LR default of 25 sharpening but frequently sharpen more than this

Isn't that interesting!
I never even knew there was a microadjustment. Makes you wonder. I'll have to take a look into this. I'm not that happy about having to drop another $140 to see if Canon are sloppy, but if it solves a problem.....
I'm using a brand new 24-70 f2.8 lens on the 5D2, so you would hope tolerances would be tight.....

I don't think it's a in-camera sharpening issue with the Nikon. It is adjustable and I have it dialed down to it's lowest setting.

nick

[marcmccalmont]

written more clearly

manually focus the 5DII with live view zoomed in to 10X and aperture f9 and reshoot
is it still soft then?
a. if yes, maybe the difference in AA filters or a bad lens
b. if no, the lens needs a micro focus adjustment.

Marc

[erick.boileau]

...

[Geoff Wittig]

The files from the 5D2 are noticeably a lot softer - actually quite dramatic - from the little hair in the center of the frame, to the fuzz on the edge of the checker, to the edge of the black cutouts around the colour swatches - softer from f2.8 all the way to f11. At f16 the difference is less noticeable and at f22 the 5D is slightly sharper.

Do you think I have a lens or camera problem here?
Anything I might be doing wrong?
nick

You need to do some kind of sharpness test using another lens to get an independent variable. There is some variability in Canon's quality control by all accounts. My copy of the 24-70 f:2.8 L is insanely sharp at all focal lengths and apertures, with the exception of some softness on the extreme left side at wider than 30 mm at middle distances and apertures, but other folks have noted problems with softness. I would test for sharpness using a known very sharp lens, to see if it's the camera/sensor rather than the lens. It's also worth locking down the camera on a rock solid tripod and focus using live view magnified to be absolutely certain it's not just a focus accuracy issue.

At least in my experience, Canon's full frame cameras can indeed provide files that look just a little soft initially due to the AA filter, but they sharpen up beautifully with modest settings in ACR (or using plain USM). If they're looking really soft, something is amiss.

[NickJB]

Just to follow up for anyone interested.
I microadjusted the focus for this particular lens. You don't need a fancy tool to do this. Just a reasonable target that allows the autofocus to actually focus on a point, then a tiny white spot that is very easy to see as it softens when either forward or back focussed. My lens was off by +4. Not much on a scale of +/- 20, but has a huge effect if you are being critical (I can be - especially after dropping $1200 on a lens...).
On a tripod, the canon 5D MkII with 24-70 f2.8 versus the Nikon D2X w/17-55 f2.8 was absolutely softer. Not a doubt in my mind. Some people have suggested to me that it could be Nikon's in-camera sharpening. I don't believe it. This was not a sharpening effect. Others have suggested the AA filter on the Canon is very strong, partly to lessen noise at higher ISO's. I can believe that and I do think that is one of the contributing factors. I also think Nikon might have a better AF than Canon (at least the 5D2), but that is pure supposition on my part. Personally, I think - opinion only here - that its a combo of strong AA filter on the 5D2, exceptionally sharp lens on the Nikon (yes, I do think Nikon glass is better, even though I juts moved over to Canon) and the effect of tighter pixel packing on the smaller D2X chip.
Interestingly, a lot of this evaporates when shooting handheld. Suddenly the larger sensor on the full frame canon does not pick up hand vibration as much as the small sensor on the Nikon. Which goes to prove the point that all pro's make - it doesn't matter what the tests and charts say, it's how the camera works when you use it in real world shooting that counts!

Thanks for all your suggestions guys. Really was a big help. I feel much more comfortable with my decision to jump to the 5D2 and Canon lenses and my designer friend now owns an incredibly sharp D2X and is very happy about it!

nick