Did a simple test in the backyard with my Zeiss 28/2.8 and Canon 28-70L (both set to f/8), and noticed that the DOF varied between the lenses. The Zeiss looked sharper to me, but the Canon had noticeably deeper DOF (see fence finials and white corner post in the shots below). Is the DOF variance a result of the lens adapter (I suspect that it is)? If so, how does shooting at the hyperfocal distance come into play with an alt lens...lots of trial and error?

Original image:

Mike,

I know you're a intelligent & careful fellow, so I don't doubt these results, but I find this VERY surprising.

Isn't DOF (theoretically) purely a function of aperture & distance to the subject? If not, then this would become a very serious topic for discussion.


Instead of a DOF difference, could we be seeing here a difference in bokeh?

In theory, yes, but optical differences will affect DOF from time to time. I have also noticed that my Zeiss 50 f/1.7 appears to have DOF equivalent to an f/1.4 lens, rather than f/1.8.

Cableaddict wrote:
Mike,

I know you're a intelligent & careful fellow, so I don't doubt these results, but I find this VERY surprising.

Isn't DOF (theoretically) purely a function of aperture & distance to the subject? If not, then this would become a very serious topic for discussion.


Instead of a DOF difference, could we be seeing here a difference in bokeh?

The shots were taken with a tripod, about 4-5 feet from the subject (large planter). The only variables were the lenses, which were swapped without moving the tripod. I don't know if the differences are a function of the adapter thickness or, as you suggest, maybe bokeh (I didn't think of that).

On a related note, when trying to use my alt lenses at the hyperfocal distances, the results aren't what one would expect...again, adapter issue?

I suspect this is related to the '3D effect' of the Zeiss lens and the stronger delineation between in-focus and out-of-focus areas that is at least partially responsible for that. Actual DoF is probably very close but the perceived DoF is larger on the zoom with its smoother transition to oof as you are seeing some of the slightly oof areas as in-focus.

According to the 2nd and 3rd paragraphs under the Lens Performance section of this review (http://www.pebbleplace.com/Personal/Contax_100mm_Planar.html) the DOF can vary based on the lens design. He mentions that DOF tables are only an estimate.

Most dof variances are due to factors like CA and tend to be very minimal after stopping down. If test conditions were nearly identical there's something wrong with the zeiss or adapter.

I've seen weird DOF when testing lenses and noticed it a while back when doing extensive tests on 28mm lenses. Take the OL28 3.5 and 2.8. The 2.8 had very little DOF behind the point of focus whereas the 3.5 and CZ 2.8 did the opposite- this was also when stopped down to F8 and F11 and I had criticaly adjusted each lens for perfect infinity registration.

What I wrote on another thread about the subject:

1. Samuli Vahonen wrote: Zeiss has pretty narrow depth of field even at f/11
2. To which Daniel Buck responded: lens brand has nothing to do with how shallow the DOF is, it's just focal length, aperture, focus distance and film size
3. To which Samuli Vahonen responded: You are right, brand has nothing to do about it, but has a lot to do with lens design!!!! - Have you tested it yourself? I have shoot many times Carl Zeiss Makro-Planar T* 2/100 ZF and Canon EF100mm f/2.8 USM (and other comparable pairs of lenses) from same position while I have been comparing the lenses and it seems quite obvious to me that Zeiss has narrower depth of field at same aperture.

Here is one thread which has some answers:
http://www.fredmiranda.com/forum/topic/666807/0

Also some other people seem to have found the same on other threads as well:
http://www.fredmiranda.com/forum/topic/684609/0#6157256
http://www.fredmiranda.com/forum/topic/574990/0#5909507
http://www.fredmiranda.com/forum/topic/517227/0#4468592

Also if you look at:
http://en.wikipedia.org/wiki/Depth_of_field#Derivation_of_the_DOF_formulas
You will find out that the DOF formulas are made for single lens element symmetrical lenses, which are very far away from modern lens photographic lens designs. Also if you go to depths to this issue I would propose following document as well:
http://www.largeformatphotography.info/articles/DoFinDepth.pdf

Also like fourfa says above there is difference between mathematical MTF DoF and apparent/perceived DoF. Which I'm referring, I have never tested my lenses with artificial targets, I prefer real outdoor shooting instead... Somehow Zeiss lens designs seems to concentrate the contrast to focal plane and has then less contrast on areas which are not in focal plane. Hard to define on my 2nd language this kind of things

--
Samuli Vahonen
http://www.vahonen.com

dcmiller wrote:
Most dof variances are due to factors like CA and tend to be very minimal after stopping down. If test conditions were nearly identical there's something wrong with the zeiss or adapter.


Presently, the adapter is a happypagehk pro AF-confirm. I bought the lens from the B/S forum here, and it came with a standard, unnamed non AF-confirm adapter (neglected to ask the seller if he had made any adjustments to the lens). With either adapter in place, the lens achieves infinity focus before the stop is reached. Other than this situation, the lens is very sharp, even in the corners. But since I shoot primarily landscapes, this piqued my curiosity. I'll try a bit more testing, and then maybe send it up to Mark Hama for a look-see.

The Zeiss lens seem to be misfocused but it is hard to determine of this has something to do with how far away the DOF goes behind the fence.

I'm with Samuli on this.

Thinking about it, what on earth ca the adapter have to do with the DOF? A lens can most often be considered to be a chunk of glass elements. At some position between the sensor and the target the subject is rendered as in focus. With an adapter having the correct thickness we can focus on anything between infinity and the minimum focus distance as par the lens designer's decision.

I can think of an adapter problem affecting everything including the DOF and OOF blur if the adapter is severely skewed - but that is seldom the case.

So, surely it has something to do with the optical formula.

You guys suggesting this is an adapter problem, what am I missing?

(As a sidenote I think I prefer the bokeh from the Canon zoom in this particular case.)

DOF can't defy the relationship between aperture and subject distance, but there are lots of effects that can muddy the issue. Excluding mechanical aberrations like a sticky iris causing one lens's f/8 to not truly be f/8, or misaligned components tilting the focal plane, there is also the possibility of field curvature bringing more (or less) of the background into common focus with the subject, or, assuming you focused wide open, of the diminishment of spherical aberration shifting the centre of your focal field backward (or forward) as you stop down.

The Canon crop looks noticably sharper than the Zeiss crop to me so I will hazzard a guess that the Zeiss was simply focussed on a closer point than the Canon was.
The 28/2.8 Distagon is not at its best when stopped down to f8 anyway and its better at wider apertures.

Alf Beharie wrote:
The Canon crop looks noticably sharper than the Zeiss crop to me so I will hazzard a guess that the Zeiss was simply focussed on a closer point than the Canon was.
The 28/2.8 Distagon is not at its best when stopped down to f8 anyway and its better at wider apertures. On the CZ 28 the corners are soft below F8 and never get realy sharp above it. Bokeh can look harsh or soft depending on the lens design between 2 identical focal length lenses at the same aperture so I have no problem beliveing that optical design has an effect on how DOF is percieved.

Have to agree that Zeiss shot was definitely out of focus.

In a test series of images with several 50mm lenses I got a misfocused EF50/1.2L image. The focal plane was not far off but the rendering (amount of blur) in the background was very different from the correctly focused image.

If the OP thinks these two lenses have different DOF characteristica I suggest retaking the images making sure both lenses are focused correctly (perhaps some focus bracketing even).

Please show us crops from the focal plane in each shot. They may be focussed at slightly different distances. Different lenses can produce differing d.o.f., but nominal apertures often seem wrong too. What was the shutter speed in each shot? Were they identically exposed?

brainiac wrote:
Please show us crops from the focal plane in each shot. They may be focussed at slightly different distances. Different lenses can produce differing d.o.f., but nominal apertures often seem wrong too. What was the shutter speed in each shot? Were they identically exposed?

The exposure varied very slightly between the two shots. The Zeiss was shot at 1/400 with 0EV, the Canon was shot at 1/320 and 0EV. Focus point was the protruding portion of the planter (since the shot was taken vertically, I used the bottom-most AF point on the 5D)...I set up the shot with the Canon AF point first so that I wouldn't have to do any focus/recompose movements. The crops above are from a straight conversion from RAW to jpeg, with sharpening set to zero in ACR, then converted to sRGB. The overall reference shot above was taken with the Zeiss.

Can't do crops of the focal plane...once I converted and cropped, I deleted the original files. May have to wait for the next sunny day...due to get rain here fairly soon.

Mike Ganz wrote:
The exposure varied very slightly between the two shots. The Zeiss was shot at 1/400 with 0EV, the Canon was shot at 1/320 and 0EV. Focus point was the protruding portion of the planter (since the shot was taken vertically, I used the bottom-most AF point on the 5D)...I set up the shot with the Canon AF point first so that I wouldn't have to do any focus/recompose movements. The crops above are from a straight conversion from RAW to jpeg, with sharpening set to zero in ACR, then converted to sRGB. The overall reference shot above was taken with the Zeiss.


So it sounds like the iris is working right. You could be seeing field curvature, or continuing to assume that you focused wide open, it could be focus shift due to spherical aberration. Crops of the focus plane wouldn't hurt, but ideally construct a new test with bits of detail ranging all the way from in front of your expected DOF to well beyond it. At intermediate range and f/8, this can require quite a few planters...

I would think that aperture ratings are not 100% precise, but are more of a "closest whole number". That, and your focusing may be what you are seeing. it looks like the Zeiss was maybe not focused out as far as the Canon was, the fence looks slightly out of focus on the Zeiss.