Aperture, how high is too high?

C

Chumpalot

Messages
669
Edit My Images
Yes
I've been playing around with the Tamron 90mm today and the different apertures. Whilst doing this I noted it went up to f/51. I thought that this was incredibly high and I don't think I've seen a picture on here that uses such a small aperture.

I took some shots of a remote control I have lying here next to my laptop. From f/4 to f/50 and the pictures are identical in terms of their exposure etc, the only difference being the ever increasing DOF.

As some of you might have guessed before, I'm incredibly keen on macro photography. What would be the highest aperture I'd need when taking pictures of bug and flowers etc..? From looking at the images I took I'm thinking somewhere around f/25 perhaps?

What sort of occasion would one need an aperture of f/51 for?

Thanks for your help

Dave
 
Thank you for the reply and for the links. Having never heard of diffraction before, I can quite clearly see the sharpness loss throughout the range on the picture of the watch. Very interesting indeed.

I guess it's just a matter of experimenting and seeing what works (y)
 
Macro photography is a bit of a special case, due to the close focusing distances and consequent shallow DOF, so stopping down is often a necessary evil. The problems this brings about are (a) slower shutter speeds, which may make camera shake or subject movement a problem; (b) noise from increased ISO to combat the long exposures; (c) diffraction, which will soften the whole image even though your DOF will increase.

For non-macro photography I rarely stop down smaller than f/8 with a crop body camera and pretty much never smaller than f/11. I don't really do "macro" but I guess you should try to stay with f/16 if you can, or perhaps f/22 as the smallest aperture if you really have to squeeze more DOF from your scene. I think that beyond f/22 (f/16 really) you are robbing Peter to pay Paul. Of course, diffraction looks really bad when you view at 100%, but at more usual sizes it may not appear so bad, so make sure you judge your results based on the final product rather than pixel peeping alone.
 
(y) Macro is something I really want to get into and it's something I'm genuinely interested in. I love the 'small' things in life and seeing the level of detail you can get from things so small and so colourful amazes me no end.

You've been a great help. Thank you very much.
 
how did you get f51 out of a lens that only goes to f32 ?
 
must try that with some of my lenses and see what i can get out of them .it maybe that the lens(es )go just that bit further than the last notch
 
most of my macro shots are taken with an aperture of between F/11-16.with most being shot within 4-5 inches of the lens,this is the sort of DOF you can expect...

DSC02302.jpg


fly3.jpg


obviously,if you want more DOF..stop down some more.how much of an insect will be in focus depends on the angle you take the shot..
 
Chumpalot said:
Actually, doing it again, wide open in Aperture priority it goes up to f64!
Click to expand...

Try reading this - http://www.kenrockwell.com/tech/macro.htm - especially the section on "Aperture". I came across it with a quick Google. I didn't read the whole thing but it is true that "macro" lenses are not as fast as their max aperture suggests, when used at or near 1:1 macro distances.

You will also note brief comments about diffraction and the advice to limit yourself to maybe f/16 or f/22 for "digital", by which he means cameras with cropped sensors rather than full frame.

By the way, I'm no great fan of Ken's writing so I don't know whether the rest of the article is solid stuff or a pile of tongue in cheek nonsense, which sometimes it is, by his own admission,
 
Stan the Man - wonderful shots indeed. I'm extremely jealous :crying:

I think f/16 - 20 is as far as I dare to go. It makes no sense to start risking the sharpness of the image, especially when (if) you're doing macro work.
 
donutagain said:
must try that with some of my lenses and see what i can get out of them .it maybe that the lens(es )go just that bit further than the last notch
Click to expand...

Basically, when a lens focuses very near, extension is added to the focal length (other things happen in macro lenses, the image zooms, the base focal length changes, etc.)

Because the focal length is now longer and the physical size of the aperture is the same the f number changes. Remember, the f number is the ratio of the focal length to the physical size of the aperture.

For example:
A 105mm lens may have an aperture of 10.5mm across to give an aperture of f/10. It may extend do 210mm for close up photography, the physical aperture will still be 10.5mm, so the f number will be f/20.

Magic!
 
Chumpalot said:
Stan the Man - wonderful shots indeed. I'm extremely jealous :crying:

I think f/16 - 20 is as far as I dare to go. It makes no sense to start risking the sharpness of the image, especially when (if) you're doing macro work.
Click to expand...

thanks mate..practice and more practice,you'll get there eventually.these are nowhere as good as most in the macro section....take a look at the likes of ajophotog's,alliec and dogfish magnets work to name a few...
 
stan the man said:
thanks mate..practice and more practice,you'll get there eventually.these are nowhere as good as most in the macro section....take a look at the likes of ajophotog's,alliec and dogfish magnets work to name a few...
Click to expand...

They're amongst the best I've seen without a doubt. Especially when you consider the sort of kit others use. Canon MPEs and such like....
 
The trade off between diffraction and depth of field has been argued for as long as photography has existed.
There was a whole school of photography led by Edward Weston that called itself the F64 school.

It has only seemed to be a problem since the arrival of lenses that were critically sharp over the entire field.

Diffraction only really sets the upper limit you can expect from a lens... Depth of field continues to expand till well after this point. In fact in pinhole photography Depth of field could be considered infinite whilst diffraction is all too obvious.

In the real world, like atmospheric landscape shots. diffraction is rarely an issue at all
nor would it be for street photography.

Even in Macro photography the depth of field can be of over all greater importance than softening due to diffraction.

In any event, diffraction only effects a proportion of the light forming the image. the focussed portion is still as sharp as ever what ever the aperture.

The portion diffracted by the edges of the diaphragm becomes progressively greater the smaller the area of the aperture passing the light.

In the days when lenses scattered more light anyway, by having less efficient coating, if any at all, Diffraction was of far less importance.
In these circumstances photographers resorted to developing to a somewhat higher contrast, whilst "expert professionals" were able to mitigate the edge softening by making a positive unsharp mask.

This unsharp mask technique has been imitated with software and is used today to correct the deliberate softening of digital capture caused by the screen in front of the sensor. The effect of Diffraction can equally be mitigated in this way, just as was done in the past.
 
f/64 may be no problem on an 8x10 view camera but that doesn't automatically make it a wise choice for an APS-C or DX sized sensor where the image blur will be a significantly greater proportion of the total image area.
 
tdodd said:
f/64 may be no problem on an 8x10 view camera but that doesn't automatically make it a wise choice for an APS-C or DX sized sensor where the image blur will be a significantly greater proportion of the total image area.
Click to expand...

Quite true... as I said it is the proportion of the light passing the blades of the aperture compared to the total light that matters.

That is why a canon G series limits the minimum to F8. However the image is still very usefully sharp at f8 and the depth of field at that aperture is vast; so I doubt there would be a useful trade off going any smaller.

Many 35mm lenses go down to F22, and in past years as far as F 32. as did most medium format lenses. Many large format studio lenses I used in the 50's went down to F128.
 
I think this is right, but comments welcome.

Diffraction is a function of the diameter of the lens aperture. This varies according to focal length. For example, f/22 on a 10-22mm super-wide set at 10mm is a tiny 0.45mm diameter, and diffraction is a big problem. But f/22 on a 100mm lens is a 4.5mm diameter, which is not so much of a problem at all.

My own tests bear this out, from which I conclude: f/22 on a 10-20mm lens - don't go there, and don't go beyond f/11, preferably f/8. With standard zooms 18-55mm, don't use f/22, but f/16 is usually acceptable, f/11 for preference. Longer lenses 100mm plus, diffraction is unlikely to be a problem.

The f/number increases in macro photography, as you focus closer. When the image is 1:1 magnification, the effective f/number is two stops higher than it is when focused on infinity. Eg f/4 at infinity focus effectively becomes f/8. This is due to the inverse square law (double the distance = one quarter the light). It is not related to any change in focal length which is a relatively insignificant side effect of internal focusing.

I use Canon, but am I right in saying that Canon and Nikon show the f/number in macro differently? With my Canon 100mm f/2.8 macro, the f/number remains constant regardless of focusing distance. However, the image obviously gets darker so the indicated shutter speed moves to two stops slower at 1:1 max macro. Eg 1/500sec at f/4 at infinity focus becomes 1/125sec at f/4 at 1:1. Am I right in thinking that a Nikon would show this differently, using the effective f/number, as well as a longer shutter speed? Both cameras would give an identical exposure of course, just show the settings differently. This would maybe explain the possibility for f/64 effective f/number?

Thanks for any replies :)
 
HoppyUK said:
I think this is right, but comments welcome.

Diffraction is a function of the diameter of the lens aperture. This varies according to focal length. For example, f/22 on a 10-22mm super-wide set at 10mm is a tiny 0.45mm diameter, and diffraction is a big problem. But f/22 on a 100mm lens is a 4.5mm diameter, which is not so much of a problem at all.

My own tests bear this out, from which I conclude: f/22 on a 10-20mm lens - don't go there, and don't go beyond f/11, preferably f/8. With standard zooms 18-55mm, don't use f/22, but f/16 is usually acceptable, f/11 for preference. Longer lenses 100mm plus, diffraction is unlikely to be a problem.

The f/number increases in macro photography, as you focus closer. When the image is 1:1 magnification, the effective f/number is two stops higher than it is when focused on infinity. Eg f/4 at infinity focus effectively becomes f/8. This is due to the inverse square law (double the distance = one quarter the light). It is not related to any change in focal length which is a relatively insignificant side effect of internal focusing.
Click to expand...

It really is mainly that you're changing the focal length.

The effect of the diffraction is actually not related to the lens focal length. Yes, the aperture is bigger, but it's also further away from the image plane so the circles of confusion spread out more by the time they reach the image plane. In practice the two equal each other. If you write the equations out you'll find that focal length will cancel itself out.

I may have to do a little more testing for ultrawide retrofocus lenses on SLRs, but I think it'll have a negligible difference.
 
HoppyUK said:
I think this is right, but comments welcome.

Diffraction is a function of the diameter of the lens aperture. This varies according to focal length. For example, f/22 on a 10-22mm super-wide set at 10mm is a tiny 0.45mm diameter, and diffraction is a big problem. But f/22 on a 100mm lens is a 4.5mm diameter, which is not so much of a problem at all.

My own tests bear this out, from which I conclude: f/22 on a 10-20mm lens - don't go there, and don't go beyond f/11, preferably f/8. With standard zooms 18-55mm, don't use f/22, but f/16 is usually acceptable, f/11 for preference. Longer lenses 100mm plus, diffraction is unlikely to be a problem.
Click to expand...

My own tests with my 100-400 at 100mm and focused ~25m-30m away, which I did link to earlier, do clearly show diffraction softening on a cropper at f/22. Even f/16 is visibly softer than f/11. Here are 100% crops at f/11, f/16, f/22, f/32....

f/11 :
20080218_114847_2834_DPP_crop.JPG


f/16 :
20080218_114857_2835_DPP_crop.JPG


f/22 :
20080218_114904_2836_DPP_crop.JPG


f/32 :
20080218_114913_2837_DPP_crop.JPG


These were shot with a 40D. I'm sure the results would look worse with a 50D, for which those who have done the maths reckon the sensor can pick up the effects of diffraction beyond f/7.1. Perhaps I'll run off some tests with my nice, sharp 85/1.8 and my 50D and see what happens. I'll also try my 10-22 as well.
 
Tim, yes I saw those. And the point is well made, but relatively speaking, the effects that I have found are far worse with shorter lenses at f/22. The difference between f/8 and f/22 on my 10-20mm Canon at 10mm are night and day.

It will be very interesting to see your 10-22mm shots ;)
 
Blapto said:
It really is mainly that you're changing the focal length.

The effect of the diffraction is actually not related to the lens focal length. Yes, the aperture is bigger, but it's also further away from the image plane so the circles of confusion spread out more by the time they reach the image plane. In practice the two equal each other. If you write the equations out you'll find that focal length will cancel itself out.

I may have to do a little more testing for ultrawide retrofocus lenses on SLRs, but I think it'll have a negligible difference.
Click to expand...

Be very interesting to see how you get on. As I've said, my own tests show that wide lenses, and their correspondingly small diameter apertures for the same f/number, are much harder hit by diffraction than longer lenses.

This supports the theory as I understand it, but I admit I've not done the maths. And tests can also be a bit misleading as the position of the aperture makes a difference, and so I guess does the exact profile of the diaphragm blades :thinking:

I can't post my own results as my imaging PC is down and I can't do a 100% crop :(
 
I have run my new series of test with my 50D. Firstly I present the results with my 85/1.8 lens as 100% crops. The subject distance was approx 50X focal length. i.e. 4.25m for the 85mm lens. These were shot raw and cropped and output from Lightroom with no adjustments from default settings. I'll start with the f/4 shot as the 10-22 can't manage much better. Hopefully I can squeeze pairs of images side by side.

f/4 and f/5.6 :
20090530_164611_6379_LR.jpg
20090530_164616_6380_LR.jpg


f/8 and f/11 :
20090530_164620_6381_LR.jpg
20090530_164624_6382_LR.jpg


f/16 and f/22 :
20090530_164628_6383_LR.jpg
20090530_164632_6384_LR.jpg


I'd say f/5.6 and f/8 results are the best. f/11 is good enough (considering we know the limitations), f/16 is useable but losing it and f/22 is very poor.

10mm results in the next post.
 
Now results from the 10-22 at 10mm with subject distance adjusted to maintain approximately the same subject size. I didn't do as well as I had hoped at keeping the subject the same size in the frame, but it's not a million miles away. Subject distance for these was approx 0.5m (50X focal length).

f/4 and f/5.6 :
20090530_165330_6393_LR.jpg
20090530_165335_6394_LR.jpg


f/8 and f/11 :
20090530_165339_6395_LR.jpg
20090530_165343_6396_LR.jpg


f/16 and f/22 :
20090530_165347_6397_LR.jpg
20090530_165351_6398_LR.jpg


To my eyes the results look fairly similar. I think the f/5.6 image may be the sharpest, just pipping f/8. The f/8 image is pretty good, f/11 is just good enough (certainly not realms apart from the 85mm f/11 results, f/16 is a bit iffy and f/22 is crap.

It seems to me the issue of diffraction is not so much related to the physical size of the aperture but to the relative size of the aperture with respect to the focal length. i.e. the aperture f/stop value.

Anyway, I shall continue to stick with apertures of f/8 and wider, regardless of lens, unless there is a bloody good reason to stop down more, like shooting with flash on a bright sunny day and needing to stay within max sync speed.
 
Brilliant pics Tim :)

10-22mm pics look just like mine - rubbish as f/22. And I agree that your 85mm shots also show significant deterioration. BUT, and this is where it gets tricky comparing different lenses, the 10-22mm images are much crisper than the 85mm, so, given that the 10-22mm starts off better, would you not say that the wide-angle pictures have suffered a relatively greater level of deterioriation?

The tests I did were with Canons 10-22 and 100 macro, but at around f/5.6 to f/8, they were pretty much identical. But at f/22 the wide zoom had turned to smudge whereas the macro was still holding up, at least to a tolerable level if you were pushed.

Bottom line for me was that the 10-22 was unusable at f/22, but the macro was merely bad. What I can't get my head around is that my findings do bear out my understand of diffraction being related to the physical diameter of the aperture. If this wasn't the case, why are the very short lenses of compact lenses restricted to f/5.6 and f/8? I appreciate that the format is different, but I don't get how this effects diffraction in the same way as aperture diameter :thinking:
 
HoppyUK said:
What I can't get my head around is that my findings do bear out my understand of diffraction being related to the physical diameter of the aperture. If this wasn't the case, why are the very short lenses of compact lenses restricted to f/5.6 and f/8? I appreciate that the format is different, but I don't get how this effects diffraction in the same way as aperture diameter :thinking:
Click to expand...
I think you mean compact cameras rather than lenses. If so, consider just how small the sensors on those cameras are, and how much smaller their silly little pixels are. Whatever the amount of diffraction for any given focal length and f/stop, think how much more of the sensor in percentage terms, or how many absolute pixels, are affected by the blur.

Let's take an example where the lens is causing an edge to blur by 0.1mm on the sensor, due to diffraction. On a 10"x8" (254x203mm) view camera that blur represents only 0.1/254X100% = 0.04% of the whole image length : barely noticeable. If you wanted a 10x8 print, magnifying the image captured by a factor of 1X, you may hardly notice it, if at all.

Now consider you have the same amount of lens blur captured on a 35mm sensor, with dimensions of 36mmx24mm. Now that 0.1mm blur occupies 0.1/36*100% = 0.27% of the image length. To print at 10x8 (approx) you would need to magnify the captured image by 254/36 = 7X. Now that original lens blur of 0.1mm will be magnified to 0.7mm in print. Very noticeable!

If we keep going with this, it follows that the smaller the sensor, the more of a concern any softness will become. If we do the maths for an APS-C sensor at 22x15mm then a 10x8 print would require a magnificatoin of 254/22 = 11.5X and the blur in print would be over 1mm.

Now, finally, if we look at the sensor from a compact camera, I'm not sure of the exact dimensions but probably something less than 8x6mm. That 0.1mm blur now occupies more than 1% of the total image length and, printed at 10x8 size the blur would be 254/8*0.1 = 3.2mm wide.

That is why diffraction softening is so significant with smaller sensors and why APS-C or DX cameras should not be treated in the same way as the old school 35mm cameras, much less medium and large format when it comes to DOF and aperture considerations.

If you want to take the discussion to the pixel level then consider 12MP cameras such as the D3 with an FX sensor, D300 with DX and any old 12MP compact. If you had a lens that caused an edge blur to be 2 pixels wide on the FX sensor, that same lens would produce a blur 3 pixels wide on the DX sensor and around 10 pixels wide on the compact. Whether viewed at 100% in an image viewer, or printed at whatever size you like, the image from the larger sensor will look (much) sharper.
 
tdodd said:
I think you mean compact cameras rather than lenses. If so, consider just how small the sensors on those cameras are, and how much smaller their silly little pixels are. Whatever the amount of diffraction for any given focal length and f/stop, think how much more of the sensor in percentage terms, or how many absolute pixels, are affected by the blur.

Let's take an example where the lens is causing an edge to blur by 0.1mm on the sensor, due to diffraction. On a 10"x8" (254x203mm) view camera that blur represents only 0.1/254X100% = 0.04% of the whole image length : barely noticeable. If you wanted a 10x8 print, magnifying the image captured by a factor of 1X, you may hardly notice it, if at all.

Now consider you have the same amount of lens blur captured on a 35mm sensor, with dimensions of 36mmx24mm. Now that 0.1mm blur occupies 0.1/36*100% = 0.27% of the image length. To print at 10x8 (approx) you would need to magnify the captured image by 254/36 = 7X. Now that original lens blur of 0.1mm will be magnified to 0.7mm in print. Very noticeable!

If we keep going with this, it follows that the smaller the sensor, the more of a concern any softness will become. If we do the maths for an APS-C sensor at 22x15mm then a 10x8 print would require a magnificatoin of 254/22 = 11.5X and the blur in print would be over 1mm.

Now, finally, if we look at the sensor from a compact camera, I'm not sure of the exact dimensions but probably something less than 8x6mm. That 0.1mm blur now occupies more than 1% of the total image length and, printed at 10x8 size the blur would be 254/8*0.1 = 3.2mm wide.

That is why diffraction softening is so significant with smaller sensors and why APS-C or DX cameras should not be treated in the same way as the old school 35mm cameras, much less medium and large format when it comes to DOF and aperture considerations.
Click to expand...

Thanks for that Tim :) It makes sense.

I mentally had kind of factored in the difference in format, but that didn't seem to fit with the massive size differential between the aperture diameters of a short lens, say 10mm, and a longer one of 100mm or so.

But my practical tests do not bear this out - the difference should be dramatic and obvious if it were so - whereas yours are consistent with f/number alone being the deciding factor, relative to format.

Cheers (y)
 
buckas said:
great topic, some good info
Click to expand...

I was about to say the same thing!

It's funny this thread was made...as I had a quesion which is along the same lines.

How come if a lens is, for example, f/4.5-5.6...how come the images they can produce go way over?

(i do apologize if that is a rediculous question :p )
 
chrisuk said:
I was about to say the same thing!

It's funny this thread was made...as I had a quesion which is along the same lines.

How come if a lens is, for example, f/4.5-5.6...how come the images they can produce go way over?

(i do apologize if that is a rediculous question :p )
Click to expand...

f4.5-5.6 refers to the WIDEST aperture at each end of the zoom. the smallest aperture of them is f22 or 32 or whatever.
Hope that makes sense!
 
Janice said:
f4.5-5.6 refers to the WIDEST aperture at each end of the zoom. the smallest aperture of them is f22 or 32 or whatever.
Hope that makes sense!
Click to expand...

It does yes, thanks. And when you put it that way it's rather obvious isn;t it :LOL:

Cheers
 
Great info on this thread guys,thanks........(y)
 
You must log in or register to reply here.
Back
Top