[ElJamoquio]

Physics is physics but whats interesting is the physics of exposure, not the physics of DOF measuring devices where the final image forming process and its resuts is kept out from the question; not to mention the whys, if's and buts to goes into the creative image forming process.
A DOF equivalent lens is no more equivalent than a exposure equivalent lens. Neither is equivalent but they are equally so.

This has been explained to you a multitude of times, please don't mis-inform others.

FF SNR "X" / DOF "Y" / Shutter Speed "Z" = APS-C SNR "X" / DOF "Y" / Shutter Speed "Z"

You cannot correctly argue against that.

[ElJamoquio]

Well, in a way. But I think his analysis presupposes that it is necessary or desirable for an m43 lens to achieve the same shallow depth of field as the FF equivalent.

There's nothing 'shallow' about the DOF assumption.

As a photographer, you want DOF "X". You can get it with 4/3, APS-C, or FF. Which system is better for you? What's the best tradeoff between size (usually larger formats are smaller for F/4 and faster lenses), cost (usually FF has more expensive cameras and less expensive lenses), etc.?

The most appropriate way to make that comparison is with camera systems that can take the same pictures.

[ElJamoquio]

[Edgar_in_Indy]

There's nothing 'shallow' about the DOF assumption.

As a photographer, you want DOF "X". You can get it with 4/3, APS-C, or FF. Which system is better for you?.

Um, I'm not sure if you're agreeing or disagreeing with me, but I think that's pretty much what I was saying. But even so, you can objectively say that the depth of field of a FF camera is more shallow than a 4/3 camera at the same aperture.

---

Last edited by Edgar_in_Indy; 06-15-2013 at 08:19 PM.

[ElJamoquio]

Um, I'm not sure if you're agreeing or disagreeing with me, but I think that's pretty much what I was saying. But even so, you can objectively say that the depth of field of a FF camera is more shallow than a 4/3 camera at the same aperture.

Not to get too argumentative (way past that point already?) but it depends on how you define aperture.

...and (less argumentative), you're probably assuming the same FOV, right?

If you use the common definition of aperture, and set it to a number 'X', and use a camera with a different size of sensor, and set it to the number 'X', the DOF will be different.

But you've already changed the number 'Y' of the focal length of the lens, right?

[Rmel26]

Obviously, this lens exists only for the DOF possibilities. A wide lens does not require a lot of shutter speed, so combined with K5's excellent high ISO abilities, even a constant 2.8 is good enough for low light criteria. This would really allow those waiting for FF to now have a real good alternative. Cost alone, FF would run in the 2-3 grand area. And I'm not sure Pentax could make it as small as the K5 (one of my requirements).

Me think, me will sell Tamron 28-75, Sigma 17-70 and Sigma 28 F1.8. Get this lens and attach it to my K5ii. Then bring along the K01 with the DA 70.

Question to the board: TO PREORDER OR WAIT.

[Edgar_in_Indy]

Obviously, this lens exists only for the DOF possibilities. A wide lens does not require a lot of shutter speed, so combined with K5's excellent high ISO abilities, even a constant 2.8 is good enough for low light criteria.

I'm actually looking at it primarily for its ability to keep my ISO down and/or my shutter speed up. The ability to get a more shallow depth of field is just a bonus, but would not be my primary motivator.

[bossa]

People seem to be overlooking the importance of getting more DOF. Whilst FF may offer opportunities for shallow DOF it can be shown that APS-C offers greater DOF for landscape and Macro. Diffraction obviously plays a part in helping to decide whether to use APS-C or FF for extended DOF in which-ever circumstance.

[Rondec]

These discussions are fairly futile, but it is understood (I think) that photography is not about depth of field as much as it is about light, subject, and composition. For this reason, good photographers are not limited to one format, even though it is clear that full frame offers more shallow depth of field lenses than APS-C of four thirds.

For many types of photography (landscape, macro, wildlife), photographers are often struggling to get a little more depth of field rather than less. Throughout history, wider apertures have been more about shooting with faster shutters speeds because of the poor quality of high iso film, than they have been about getting narrow depth of field.

It is imperative that photographers choose a format that allows them to realize their vision.

[Class A]

It's such a disappointment that m43 has never delivered on its promise of fast, compact, affordable zoom lenses. ... A smaller sensor means smaller lens elements, and smaller overall lens size, and reduced production cost. But instead of affordable f2 zooms, m43 has only gotten ridiculously over-priced f/2.8 zooms.

A smaller sensor also implies a higher enlargement factor.

Hence, the precision requirements regarding optical alignment, auto focus mechanics and auto focus control are much higher with a small format such as m43, compared to FF.

So while it is possible to save a bit of money on the size of the rear elements, the overall cost is mainly determined by how fast the lens is (front group size) and the precision requirements.

I believe that the high m43 lens prices are not inflated but a result of the attempt to achieve high image quality with a small sensor.

BTW, one must not confuse "same aperture" with "same f-ratio". The former (referring to the diameter of the lens diaphragm) controls the total amount of light whereas the latter (referring to the ratio of focal-length/aperture) controls the exposure (amount per square millimetre). The same amount of total light achieves the same image quality. The same exposure does not. It is all fine that "f/1.8" achieves the "same exposure" on all formats but an "f/1.8" lens for the (old) Q is just an f/10 in FF-terms and that holds true for the total amount of light, not just for DOF.

Whilst FF may offer opportunities for shallow DOF it can be shown that APS-C offers greater DOF for landscape and Macro.

That is only true when you consider the extreme end, i.e., beyond the f/22 that most lenses support.
Going beyond such extreme f-ratios rarely makes sense due to the diffraction causing resolution to suffer considerably.

The solution for macro photography is focus stacking and for landscape photography it is tilt and shift lenses. I don't believe anyone serious about either genre would be well advised to use a smaller sensor format just because the respective lenses allow to imply higher amounts of diffraction than a f/22 lens on FF.

[ElJamoquio]

People seem to be overlooking the importance of getting more DOF. Whilst FF may offer opportunities for shallow DOF it can be shown that APS-C offers greater DOF for landscape and Macro. Diffraction obviously plays a part in helping to decide whether to use APS-C or FF for extended DOF in which-ever circumstance.

It works out to be the same diffraction for the same DOF.

I think all of the primes I own go to F/22 on FF (equivalent diffraction to F/16 on APS-C) and I have some cheapo-zooms that'll do up to F/45 IIRC. I'm not worried about DOF at F/22, personally.

In that rare case (well under 1% for me) when you need more DOF than F/22 on FF, you can use the same focal length lens from a APS-C camera on the FF camera and crop.

[thibs]

It works out to be the same diffraction for the same DOF.

I think all of the primes I own go to F/22 on FF (equivalent diffraction to F/16 on APS-C) and I have some cheapo-zooms that'll do up to F/45 IIRC. I'm not worried about DOF at F/22, personally.

In that rare case (well under 1% for me) when you need more DOF than F/22 on FF, you can use the same focal length lens from a APS-C camera on the FF camera and crop.

Crop when doing Macro?
That's approaching heresy to me

[falconeye]

I will not again enter all that equivalent discussion. There's only one correct answer and it has been given already.

However, within the context of this thread abd after recent comments, I like to repeat a detail I mentioned earlier in the thread:

The 18-35/1.8 has a minimum aperture of F16. While the full frame 24-70/2.8 from the same Sigma has F22. So, even the minimum aperture equivalence holds true in this particular case (the size of the smallest aperture hole the lenses can form is the same).

This renders arguments of deep DoF for landscape void. Not that it matters anyway. Equivalent F22 creates soft images with recent cameras and is to be avoided for landscape anyway (I'd not go beyond equivalent F11 for andscape).

[lytrytyr]

I will not again enter all that equivalent discussion. There's only one correct answer and it has been given already.

Actually, looking at the "White Paper",
there is one aspect of the answer which does not yet seem to have been clarified,
as far as I have been able to find.

You describe the following 4 parameters as fundamental:
Exposure time t, aperture diameter d, field of view FoV, and number #MP of pixels.

Then as I understand equivalence,
you may have two different choices of parameters,
say < t, d, FoV, #MP > and < t', d', FoV', #MP'>,
that are equivalent.

In the "White Paper" and other discussion,
you give examples of equivalence,
but these examples are generally in terms of some scaling c.

However, if I have understood the equivalence idea correctly,
there should be some function C (for "Camera number"),
say C(t, d, FoV, #MP),
such that the two choices < t, d, FoV, #MP > and < t', d', FoV', #MP'>
are equivalent precisely when C(t, d, FoV, #MP) = C(t', d', FoV', #MP').

Are you able to give us such a "camera number" function
that is more general than just involving a scaling c?

I believe that seeing such a "camera number" function
would help people reach a fuller understanding of equivalence.

[falconeye]

Actually, looking at the "White Paper",
...
Then as I understand equivalence,
you may have two different choices of parameters,
say < t, d, FoV, #MP > and < t', d', FoV', #MP'>,
that are equivalent.

Even if I don't discuss, I'll always try to answer precise questions like this one.

I think the "White Paper" lost you somewhere which is certainly my fault. The above sentence shows it.

Equivalence basically says that a given parameter < t, d, FoV, #MP > defines an image completely, that all images shot using this parameter are indistinguishable (not just DoF, but wrt all aspects including noise and diffraction) whatever be the sensor format (of course always assuming a proper adjustment of brightness, e.g. by iso). So, two different choices of parameters aren't equivalent, your function C is the delta-function... Changes in #MP are a bit of a special case which I discuss too.

If you are a mathematician (and ignoring #MP for a second like DxO does too), then I would state it as follows: The set of all images taken with a given parameter < t, d, FoV > of the same subject and with iso set to produce the same brightness form one equivalency class. Within one equivalency class, you find images for varying sensor sizes but except for exif data, you won't be able to deduce which camera produced which images (lens artefacts like corner aberration or bokeh may allow to detect a lens, so camera equivalency is a law of physics applying to ideal lenses). Therefore, it doesn't make any sense artistically, to give sensor-dependant data to describe an image's technical parameters. Equivalence allows a photographer to hide (read: ignore) the camera format and still reveal ALL relevant photographic parameters like focal length, iso and f-stop in a meaningful way. I.e., to determine all artistically relevant parameters w/o even having to know inner camera details like the sensor diagonal.

The "scaling function" does nothing but express t, d, FoV and brightness normalization in terms of a given format's focal length, F-stop and iso. It is all really very simple.

To summarize for the sake of this thread:

The Sigma 18-35/1.8 lens + K-5II (ISO 80) camera is equivalent to a 28-50/2.8 + FF (16MP, ISO 160) camera (numbers rounded).

The images from either combo are indistinguishable from one another. This is why this Sigma lens does matter.

Should there remain any differences, then they come from a difference in MP, difference in AF precision, difference in avoidable lens aberrations, difference in available zoom range, or difference in available low iso. Equivalence allows to discuss those and avoid discussing the obvious.