FX/DX depth of field

[nfoto]

The basic hypothesis here is that there is no such thing as a DOF "advantage" between the FX and DX formats, if focal length, perspective, and detail magnification is kept equal. Most assertions of different DOF entails one or several of these underlying assumptions of equality be violated.

We need to prove the basic hypothesis before any further analysis can be performed.

The comparison below shows the extremely limited DOF of the Heligon 50/0.75 lens on FX (left) and DX (right). Focal length, perspective, detail magnification all are identical. It should be pretty obvious that so is the DOF (note that the DOF shows the field curvature very nicely as well).

[annedi]

[...didn't we already know this from using the crop mode on the D2X ?]

This is a great example, Bjørn. Thank you.
Do you have any objections to my saving a copy of this with your permission ??

[nfoto]

Go ahead, Andrea. You're most welcome.

And no, people need to be shown the obvious before they believe. Besides, as a scientist one starts with the basic foundation.

[Nikkor]

Just to clarify things for the less experienced readers of this thread:

The DOF issue arises because we photogs use a certain lens (eg 105mm) to shoot a certain subject (eg portrait) so when using a certain format (DX or  FX) we tend to frame the same way, which means the camera-to-subject distance changes which changes the DOF

In nfotos example the camera-to-subject distance does not change so the DOF stays the same.

[nfoto]

It is not that simple. Getting the "same" field of view is not mandatory at all. Each formst has its benefitds and drawbacks. Use it in an optimal fashion.

What I'm addressing is the myth that DX has, by default, the bigger DOF compared to FX. Higher magnification  changes DOF, but as such is not a direct function of the format.

[Nikkor]

Isn't the DOF a characteristic of the lens only? If so, then no matter what you hook up on the end of it will have the same DOF, no?

[nfoto]

In a sense, yes. But DOF is not an inherent property of the lens, it is a perception of the human mind and the circumstances under which the print is viewed.

The general argument goes like this: DX is smaller so it must be magnified more. Hence the DOF will be greater because blur discs are magnified more.

Now, in th example I showed, the scales in the viewer's domain *are* similar. The picture details appear to have equal scale. Accordingly, if the format-specific hypothsis is true, DX should have the greater DOF.

[Nikkor]

So in your example did you magnify the D300 shot to match the less magnification of the D3?

[nfoto]

Read the specifications. Everything initally was similar. That was the basic requirement for a true "null" hypothesis.

The shots are shown here at the same apparent scale. Like in a print made to a specific size.

[Leon]

The basic hypothesis here is that there is no such thing as a DOF "advantage" between the FX and DX formats, if focal length, perspective, and detail magnification is kept equal. Most assertions of different DOF entails one or several of these underlying assumptions of equality be violated.

We need to prove the basic hypothesis before any further analysis can be performed.

Of course with the fixed rules you wrote, the DOF is just the same. The shot / image / enlargement of "subject" is equal for the part it is seen. Every DOF calculator shall justify this statement as a result.

But in practice mostly we have another starting point to compare in such a way, mostly one or more of the written assumptions shall be violated beforehand. Or as a result from another starting point, there is an "advantage" (or dis-advantage) between the FX and DX formats 

[Nikkor]

OK, let me recap,

Same lens, same subject, same magnification (camera-to-subject distance changes due to different formats DX v FX) and same print output (so DX is magnified).

DOF stays the same, yes?

[nfoto]

Leon: Well, since the CoC  criterion is made dependent on the format, the DOF calaculators will indicate that the formats lead to different DOF.

My point here is to make the basic and testable assertion from observed data, not from a model, since if you start to "prove" using a model a circular reasoning comes into play.

After this starting point, we can go on to assessing the ramifications of the various formats.


Nikkor: your recap contains a contradiction.

[annedi]

Nikkor, you changed the magnification when you changed the camera-to-subject distance.
So in your example, DoF is not the same.
*********

Where do they make CoC criterion dependent on the format ??
It's just geometry....
The CoC can be projected onto any sized background (focal plane or sensor) you want,
and it is not going to change size because of the differences in the size of the background (focal plane or sensor)
as long as the plane of the background is not moved forward or backward.

[Leon]

Same lens, same subject, same magnification (camera-to-subject distance changes due to different formats DX v FX) and same print output (so DX is magnified).

No.
Same lens, same subject, same perspective, same magnification --> camera to subject is just the same.
So image is just equal for the part of the subject it is seen.
Only DX is just a smaller cut, from the entire image as seen with FX.
Print output is not involved within this example on screen, but even if you do the same "magnification", it is true.
Only for FX you shall have a bigger print as for DX in this particular example.
And as important, you have to watch the different print sizes, at the same distance in this example.

[nfoto]

Yes, it is geometry and not a tangible entity.

Have a look at DOFMaster or other online calculators of this kind and see for yourself that CoC is defined as being format dependent.

[Nikkor]

Nikkor, you changed the magnification when you changed the camera-to-subject distance.
So in your example, DoF is not the same.
*********

Where do they make CoC criterion dependent on the format ??
It's just geometry....

The camera-to-subject distance changes to get the same magnification on the different formats

Wow, I think i need to go to bed

[nfoto]

Andrea (annedi) correctly pointed out the errror of your reasoning. Careful reading of the setting up for the null hypothesis should do likewise.

[Leon]

Leon: Well, since the CoC  criterion is made dependent on the format, the DOF calaculators will indicate that the formats lead to different DOF.

Yes, it is geometry and not a tangible entity.

Have a look at DOFMaster or other online calculators of this kind and see for yourself that CoC is defined as being format dependent.

The format Coc is not involved within your example as you are not comparing formats but magnifications.
You can use every Coc, as long if you use the same Coc for different calcualtions within your example.

[annedi]

Nikkor, the camera-to-subject distance is changed to get the same angle-of-view.
In changing that distance you change the magnification.

[nfoto]

Leon: Of course, this is true in the literal sense. But we need to keep the CoC in mind in the next steps when more variables undergo changes. As I stated several times, I'm establishing the "null" hypothesis only. For this purpose, we don't need to introduce the concept of CoC at all. The point here is that we don't *calculate* DOF and test that against observation, we try to see if two pictures match in appearance. CoC is a modelling concept.