What is the biggest, perfectly sharp format you can get from your sharpest negatives?

[Diapositivo]

In reading the interesting comments to the tests carried on by the site theonlinephotographer, quoted by Lee, I ended on Wikipedia, normally not the best place where to look for reliable information admittedly, and I found this at the voice "Depth of field"

A subject distance is decreased, the subject magnification increases, and eventually becomes large in comparison with the hyperfocal magnification. Thus the effect of focal length is greatest near the hyperfocal distance, and decreases as subject distance is decreased. However, the near/far perspective will differ for different focal lengths, so the difference in DOF may not be readily apparent.

[formulas here, that don't cut-and-paste well]

so that for a given magnification, DOF is essentially independent of focal length. Stated otherwise, for the same subject magnification and the same f-number, all focal lengths for a given image format give approximately the same DOF. This statement is true only when the subject distance is small in comparison with the hyperfocal distance, however.

This somehow justifies the images published by Ralph with the apparent inconsistency of the principle in the macro world. The "advantage" of the higher DoF would exist only from somewhere near the hyperfocal distance on.

In the macro field, lenses work at distances that are much below the hyperfocal distance and, in that working range, the DoF advantage of shorter focal lengths disappears.

Hence the success of 100mm macro for field work, and let's say the fact that the 4:3 camera are not considered more suitable for macro work than the 135 cameras. If the DoF advantage of shorter focal lengths had applied to all focusing distances, 4:3 users would have enjoyed a "free meal"* so to speak, and the other formats would be at strong disadvantage, and macro photography would be a selling point for 4:3 material.

Fabrizio

* That's because the small decay of resolution due to a smaller sensor is probably much less important, in macrophotography, than an increased DoF which is the constant worry and trouble of macro phogographers.

 

[RalphLambrecht]

Although I agree that as an explanation it can be just a very partial one, what I mean is not that Photoshop added details to the 24mm, but that with an "optically" made enlargement the CoCs would have grown "just like the enlargement", while with a digital enlargement the enlarged pixels are, in fact, "calculated" so the enlargement is less "confused". We don't have the "linear defocusing" of an optical enlargement so to speak. ...

What would happen in the digital world if the 24mm image was not enlarged but the 85mm image was reduced instead?

 

[removed account4]

thank you ralph !
as they say, a picture is worth 1000 words

 

[cowanw]

Bill, I don't know if I get you right, but when people say that DoF depends on aperture (among other factors) they usually mean f/value, not absolute aperture in mm. That f/value already takes into account that for different focal lengths, the absolute aperture in mm that results in a certain f/value is different.

I can't help thinking where is the "catch" here...

I was merely pointing out that the fstop is a human construct.
As such to say that Depth of field is variable to this is misleading in and of itself.
If fstop is a function of focal length and aperture, which are real measurments, and, as such, not human constructs, then the relationship of Depth of field has tio be seen in relation of the two physical /optical things that make up fstop.
Aperture is actually easy. f8 at at 28mm focal length is a much smaller hole than F8 at 200mm. The larger hole will give less acceptable spread of focus.
Focal length as Ralph has shown also affects DOF.
So we loose two ways.
I am trying to up;oad the pages from time-life the camera which helped me.

 

[cowanw]

Try again
image on the left 28 mm lens
Image on the right 135mm lens
The hole in the film plae schematic represents the COC

 

[Athiril]

From a recent test of mine.. this is Ilford FP4+ in Rodinal. 6x7cm.. but well the same level of detail applies to other formats, just you will have smaller or bigger prints at the same enlargement ratio.

The crop on the bottom right of the first pic I've calculated at about 51 lp/mm, top right is about 23 lp/mm. Just 51 lp/mm gives you a lot of detail. The crop almost looks like it's getting into macro.

Second pic, is a crop from the same image, should be again about the same, about ~51 lp/mm.

 

[Athiril]

No need to be patronizing. After presenting the math for my argument and offering pictorial evidence, you may want to put your cards on the table. Show us the math that shows how print magnification changes the DoF. I'm really interested.

Q.G.

Still waiting for your evidence. Math, data, pictorial samples. Where is it?

My claim: DoF is locked in after exposure. Print magnification cannot change it.
My proof: All published DoF equations include nothing but image-taking parameters.

Your claim: DoF can be changed with print magnification.
Your proof: ?

How? Just tell us. Really like to see it.

Sorry, but DoF is what appears acceptably sharp.

Your math and proof isn't flawed, but your claim is.

Your claim that DoF is static after exposure is flawed defeated by the very definition of DoF.

What you're talking about is what's actually in focus and what's not. You're not talking about what appears acceptably sharp (but isn't necessarily in focus), which is DoF.

 

[Diapositivo]

Ralph:

I am not sure what would happen, because when you reduce an image normally you apply some form of sharpening (is the reducing algorithm itself that is more or less sharpening), that is deemed necessary in case of reduction, but this has something to do that with pixels we have to do with "finite" quantities, you can't zoom the dimension of the single pixel while in the analogue world we are, in a sense, zooming the pixel dimension. Digital is much more "thought upon" than analogue. The general answer is that after a reduction things appear to be more "in focus" and you appear to have more DoF. This was tested by me and by several other contributors to a stock Agency, Alamy, which does not make any selection for content, but used to be very (very) picky about technical details, focus placement in particular. The "trick" that people used when dealing with "suspicious" scans was to reduce them to the minimum admitted size. That visibly improved definition when the image was inspected at 100% in Photoshop, even if there certainly was a loss of information. (Alamy have relaxed their quality criteria in the meanwhile).

Greg, I agree with you, the digital "pixel stuffing" does not explain on its own the DoF difference that can be seen in Ralph's images;

Bill, in fact, the "absolute aperture" theme was brought also in the theonlinephotographer blog by some commentator. While it might be that the "law" of indifference of focal length applies if we consider "real" apertures (in mm) that would be in any case a demonstration that there is an advantage in DoF while using smaller formats because you have to close "more" (in f/x terms) to regain the DoF which you "lost" while using a larger format. Which is quite surprising, at least to me.

I was the one who, replying to A49, pointed out that choosing a smaller format would not lead to a DoF gain if final magnification and composition had to be equal, and Ralph test showed I was wrong (until somebody points out some fault in his test).

In your cemetery shots, Bill, I don't understand where the focus point is. Is it in both cases on the "in memory of a beloved husband" tablet?

Fabrizio

PS Athirill, that is what I was thinking myself, but the pictures posted by Ralph showed that even after enlarging the difference in "out-of-focusness" is quite apparent. One would expect that, after enlarging a section of the wide-angle picture, the CoCs would be similar to the ones of the tele-lens picture. Ralph, the way I see it, did bring the "equal magnification" factor into the test, and the DoF still showed a difference.

 

[Diapositivo]

Oh, a further observation.

In his text, the curator of cambridgeincolour states that comparison of DoF should be made within regions that actually have "focus" as defined by a certain criterion. Regions "out of focus" tend to worsen focus, so to speak, more rapidly with tele lens than with wide-angle lens because of the different magnification involved.

While there is probably some truth in this reasoning, the final, perceptual result is that we do perceive "slightly out of focus" as different from "quite out of focus" in practical terms. In the case shown by Ralph, the far plane of the wide-angle image is out of focus. Nonetheless, it is quite clearly "less" so than the corresponding section in the tele picture. At each possible enlargement, if the back plane is not "quite in focus" in both images, it is going to be "less out of focus" in the cropped and enlarged wide-angle image than in the tele image.

This is an important practical difference because in photography we often deal with "degrees" of blurriness anyway.

Fabrizio

 

[Athiril]

The more pronounced narrow focus you have.. ie: the less in focus something is, ie: the less resolved something is, the smaller/further away you have to view it for it to be 'accpeptably sharp' to meet the limit of your eye.

My 3200 dpi film scans on my V500 are anything but acceptably sharp, by doing several tests of reducing down the size in Photoshop then back up to 3200 dpi until I notice any change or loss in detail starting to occur, I find the point at which it becomes acceptably sharp (and the real detail level, which happens to be ~1200 dpi, the test done without any sharpening at any point, otherwise you'll throw the test off by perceiving a loss of acutance and noise introduced by sharpening).

 

[Athiril]

Ralph's 85mm f/5.6 image is now acceptable sharp from front to back, this fulfills the criteria of having a DoF from the foreground to background. The DoF has changed with the level enlargement, or rather, the physical size in your field of vision you're viewing the image at.

It is unacceptably small however, though that is not a criteria in DoF. You can of course have images with much smaller differences in focus/out of focus area, IE: a focus distance set to further away, and something on the edge of the plane/zone of focus rather well and truly outside it.

Anyway, let's blow it back up.

Not acceptably sharp anyway, the entire thing looks out of focus, but I can't discern a focus difference in the image there.

 

[Athiril]

Experiment time.

This one is a crop since I dont want to attach huge images.

But you can see here, when this is sized down to the same as the first one the actual difference.

 

[RalphLambrecht]

... PS Athirill, that is what I was thinking myself, but the pictures posted by Ralph showed that even after enlarging the difference in "out-of-focusness" is quite apparent. One would expect that, after enlarging a section of the wide-angle picture, the CoCs would be similar to the ones of the tele-lens picture. Ralph, the way I see it, did bring the "equal magnification" factor into the test, and the DoF still showed a difference.

I should have provided the math for these images:

24 mm 85 mm
-------------------------------------------
focus distance 2 m 2 m
f/stop 5.6 5.6
CoC 0.022 mm 0.022 mm
DoF front 1.41 m 1.94 m
DoF rear 3.46 m 2.07 m
total DoF 2.06 m 0.13 m

Strictly speaking, the back wall is out of focus in both images (so is everything in front or behind the tail lamp). The point is, the DoF is much larger with the shorter lens (see the numbers above). Using such a lens, one has a better chance to get much more acceptable detail (definition of DoF) into the image. The math hints at it, and the images prove it.

One could claim that after enlarging the 24mm image to fit the 85mm image frame, the CoC should be corrected accordingly and thereby reducing the DoF. Well, the enlarging factor was roughly 3.5, so, let's reduce the CoC by it and let's see what happens:

24 mm 85 mm
-------------------------------------------
focus distance 2 m 2 m
f/stop 5.6 5.6
CoC 0.006 mm 0.022 mm
DoF front 1.79 m 1.94 m
DoF rear 2.26 m 2.07 m
total DoF 0.47 m 0.13 m

The DoF for the 24mm image is now significantly reduced but still almost 4-times larger than the DoF of the 85mm image!

No matter how you slice it, the statement that there is no DoF benefit from going to a shorter lens is incorrect, and every owner of a miniature-format camera has experienced it. As they can tell you, this benefit can be a disadvantage.

 

[Athiril]

"the statement that there is no DoF benefit from going to a shorter lens is incorrect"

You should have just linked 35mm DoF adapters for video cameras.


Your test is also shot at the same f-stop, instead of the same aperture.

 

[RalphLambrecht]

... Your test is also shot at the same f-stop, instead of the same aperture.

Now, I wonder what I'm getting when I set my camera to 'aperture priority'.

 

[Bill Burk]

This makes me think a lot of different things.

Confirms that wide angle lenses get more DoF, and it's not just because the image is smaller.
There is a real difference between formats.
For more bokeh, a larger format will give you more - for the same f/stop.
Explains why you get f/16 on 35mm format lenses and f/64 on large format lenses - for the same DoF you may need it.

 

[Athiril]

"The point is, the DoF is much larger with the shorter lens (see the numbers above). Using such a lens, one has a better chance to get much more acceptable detail (definition of DoF) into the image."

I'm going to use 35mm with 50mm lens vs a 72x48mm frame with 100mm lens as an example, because I am very tired and want nicely dividable math.

If we want the same acceptable amount of detail from them both (same lpph), all things equal, lets say we choose f/16 for 35mm, then we can use f/32 for the 72x48mm frame.

So the 35mm frame is diffraction limited to around 57 lp/mm, while the 72x48mm frame is diffraction limited to around 28.5 lp/mm, (more likely to hit diffraction limit on a lens @ f/32, but we need to remove imperfect differences from the equation which does contribute).

100mm enlarges/amplifies camera movement/shake, by a factor of 2 (in whatever fashion, up/down, left/right, rotation, etc), but the tolerance is greater (if we aren't going to use tripods for the test) as the area in maximum point of detail size (diffraction limit) is 4x larger in area, so the tolerance is greater by a factor of 2, so for the same shutter speed, we're all good, we just need the extra exposure now.

But on top of that we don't need as fine a grain film, nor as sharp, we can use one that has 4x larger grain (2x RMS), and half the lp/mm, in fact, lp/mm of the film is not a concern @ 28.5 max lens lp/mm, so we can forego that.


Lets say Fujicolor Reala 100 for the 35mm. Then given are requirements according to the datasheet, we can use Superia 1600 which has just under double the RMS grain rating, of course thats 35mm only.. but there is 800Z, which is ~1.6x larger area grain than Reala (according to Fuji), sensitivity advantage to the larger frame.. on top of that we still have the benefit of more camera movement tolerance for slower shutter speed, but here we can already use a faster shutter speed. There are plenty of other film combinations.

EDIT: Here I want to point out that Reala is 63 lp/mm, and 800Z is 50 lp/mm, so this drops our system resolution for 35mm down to 42.3 lp/mm and 24.7 lp/mm for the 72x48mm frame, so in landscape orientation, our lpph for 35mm becomes 1015.2, and for 72x48mm frame it becomes 1185.6

If we could use Superia 1600 in our imaginary format, we have +4 stops to play with instead of +3 with 800Z, but even 800Z gives us a +1 stop sensitivity advantage over the 35mm format. This is a moot point though.






Now given the depth of field is the same for f/16 50mm and f/32 100mm, the drop off rate for hot out of focus, or rather the curve at which lp/mm drops on either side should be the same, so the shape of the lp/mm vs distance curve (where the focus point is half way along the distance graph) should be the same for both. So in terms of acceptable detail, they have the same detail level, and the same resolution fall off which translates into DoF when it can be perceived.


So, where we use a tripod, there is no DoF advantage, as we can get the same DoF, with the same level of detail.

In the case where we need to shoot hand held and maintain a shutter speed or for some other reason, there is no DoF advantage, because we are able to stop down to the same DoF, with the same detail and shutter speed.


You don't have a better chance of more acceptable detail imo given the above, the DoF isn't larger with a shorter lens, DoF is only larger (when the full resolution gradient is perceivable) with a difference in aperture, not focal length, focal length doesn't change DoF.

It even seems like the bigger format even when stopping down to the same DoF has a finer detail, finer grain, and greater speed advantages on top of that. That's just one specific example though, but film appears to scale upwards in favour.

 

[Athiril]

This makes me think a lot of different things.

Confirms that wide angle lenses get more DoF, and it's not just because the image is smaller.
There is a real difference between formats.
For more bokeh, a larger format will give you more - for the same f/stop.
Explains why you get f/16 on 35mm format lenses and f/64 on large format lenses - for the same DoF you may need it.

Angle of view doesnt not affect DoF at all.

Well 65mm is wide angle on 6x7cm format, and 90-127mm is wide too on 4x5"

90mm f/8 is 90mm f/8, same DoF, regardless if its on 4x5" or 35mm.. consider you can mount that 90mm lens on a 35mm to begin with.. ow the angle of view is different.. lens is identical.. how does what is in and what is not in focus change? It doesn't.


Focal length is also irrelevent to DoF. Aperture and focus distance determine DoF.

Mounting that 90mm on a 35mm camera, is like cutting out/cropping your 4x5" shot to a 36x24mm area.. it doesnt change what is in or out of focus.. 90mm is certainly not wide on 35mm format.. wide angle doesn't get more DoF.

 

[RalphLambrecht]

This makes me think a lot of different things.

Confirms that wide angle lenses get more DoF, and it's not just because the image is smaller.
There is a real difference between formats.
For more bokeh, a larger format will give you more - for the same f/stop.
Explains why you get f/16 on 35mm format lenses and f/64 on large format lenses - for the same DoF you may need it.

Yes, I agree!

 

[RalphLambrecht]

... Focal length is also irrelevent to DoF. Aperture and focus distance determine DoF.

Nonsense. You better check the math!

See the focal length in there? Squared!

 

[23mjm]

Oh DOF, it's the Rubik's cube of photography for many!!! A quick rule of thumb for DOF everything matters!!!

 

[cowanw]

Focal length is also irrelevent to DoF. Aperture and focus distance determine DoF.

Can you help me explain the schematics in post 159 then?

 

[Bill Burk]

I made an oversimplified statement saying "wide angle", for example of course 90mm is wide on LF. I should have said shorter focal length.

I appreciate your creative thoughts in post 171. The examples make sense and support the thought that film favors scaling up.

Your 1cm square thumbnail of Ralph's 85mm is clever but overemphasizes a point that isn't disputed - I think we all agree perception of sharpness varies directly with print viewing distance. If I may oversimplify again, we agreed to an arbitrary circle of confusion that corroborates what is engraved on all our lenses (but we reserve the right to say that isn't good enough).

In your example you describe using f/16 at 35mm format and f/32 at the larger format, and you are guessing the DoF will be the same. I think you are on-track and that is the very point that Ralph's photographs proved.

I don't think your conclusion that focal length doesn't affect DoF lines up with the evidence, and I don't understand why you think Ralph's conclusion is wrong.

Maybe you're thinking of the literal diameter of the aperture blade opening instead of f/stops?

And maybe that is part of the puzzle. If we used diameter to calibrate apertures instead of stops, maybe then DoF would not vary as much with different focal lengths?

I tend to imagine "diffraction-limited" happening when the aperture blades are so physically close to each other that light is bending around them. With longer focal length lenses, the diameter of the aperture blade opening is larger. Can it also be that longer focal length lenses are not diffraction limited until higher f/stops?

And another thought that occurs to me from Ralph's tests: I generally prefer f/2 lenses for 35mm camera for selective focus (converse of DoF). 135mm f/4 probably delivers equivalent selective focus as 40mm f/2.

 

[RalphLambrecht]

Can you help me explain the schematics in post 159 then?

... or the attached equations?

 

[RalphLambrecht]

... Maybe you're thinking of the literal diameter of the aperture blade opening instead of f/stops?

And maybe that is part of the puzzle. If we used diameter to calibrate apertures instead of stops, maybe then DoF would not vary as much with different focal lengths? ...

Unfortunately, that won't get the focal length out of the equation either. (see previous post).

A 50mm lens at f/8 has the same aperture dimension as a 100mm lens at f/16 (6.3 mm), but the 50mm lens has more than double the DoF at that setting. Indeed, you have to stop down to f/32 with the 100mm lens to match the 50mm lens at f/8.

It does however work when the two lenses are used on different film formats:
For example, a medium-format 80mm lens at f/8 and a small-format 40mm lens at f/4 have the same opening (10 mm) and produce the same DoF, because the medium-format CoC is twice as large as the other.

However, there is no need to make it that complicated. The above equations work well and show the influence of subject distance, focal length, aperture (setting) and film format. Arguing that one or the other variable does not matter if compensated by another is just a stealth way of arguing for argument's sake.

We heard it all now: 'same enlargement, same DoF' or 'focal length does not affect DoF'. Yet, the math and the pictures tell a different story.