Too narrow an aperture = bad?

futureal33

futureal33

Messages
2,390
Name
Nick
Edit My Images
Yes
Last night, experimenting, I took 3 pictures on a tripod of the exact same thing in my garden..

All @ 18mm, settings were

1) 1/60sec F5.6 ISO 800
2) 1/30sec F11 ISO 800
3) 0.6sec F29 ISO 800

Now, to my untrained eye, using LightRoom I struggled to see any difference in DOF or IQ between the first 2 shots (F5.6 & F11) but on the F29 shot the IQ was significantly reduced, sharpness especially was reduced and the whole image looked pretty rubbish compared to the previous 2...

Now it is my understanding that increasing the F value generally increases the sharpness of a given photo...

But can you go too high ?

Thanks :)
 
f/16 is the max I tend to use, with the optimum being bwtween f/8 and f/11 depending on the lens.
 
Most lenses tend to have an optimum operating aperture of f8-f11. At f29 you are way out of the lens's comfort zone.

Steve
 
I found a picture of mine at f/56 yesterday!!!

What the hell was I thinking?
It was from 2005 though, I didnt have a clue back then.
 
futureal33 said:
Last night, experimenting, I took 3 pictures on a tripod of the exact same thing in my garden..

All @ 18mm, settings were

1) 1/60sec F5.6 ISO 800
2) 1/30sec F11 ISO 800
3) 0.6sec F29 ISO 800

Now, to my untrained eye, using LightRoom I struggled to see any difference in DOF or IQ between the first 2 shots (F5.6 & F11) but on the F29 shot the IQ was significantly reduced, sharpness especially was reduced and the whole image looked pretty rubbish compared to the previous 2...

Now it is my understanding that increasing the F value generally increases the sharpness of a given photo...

But can you go too high ?

Thanks :)
Click to expand...

At 18mm, unless you're below about 1m distance to subject you're hitting the hyperfocal distances (check the DoF Table).

Not a good choice of focal length for a depth of field test, but a very good demonstration why you don't need to close-up the aperture very far when shooting UWA landscapes on a crop sensor.
 
As a simple rule of thumb for wide angle landscape photography try picking an aperture equivalent to the focal length divided by 2. e.g.....

At 10mm use f/5
At 22mm use f/11
At 35mm use f/16 or f/18 (although on a cropper that is no longer "wide", but fine on full frame).

It's simple and it works and it doesn't matter which sensor size you have either. Check out the DOF you will get on the DOFMaster website by following this guideline. It should be enough for most needs.

If you don't need truckloads of DOF then try dividing focal length by 3 and you should still get a good result.

As for apertures at which diffraction begins to become visible at 100% viewing, there is a table listing several cameras here - http://www.the-digital-picture.com/reviews/canon-eos-40d-digital-slr-camera-review.aspx. For the 40D the DLA is f/9.3, which means that if you want loads of DOF you can freely stop down to f/9 without a care. Beyond that you will be buying increased DOF at the expense of absolute sharpness at the focused distance. Nonetheless, if you need the DOF then don't be afraid to go a bit further if you need to, because when viewed at more traditional enlargement sizes, like in print or fit to screen, the diffraction effects up to say f/16, should be all but imperceptible. Going beyond f/16 on a crop body is something to be used with more caution.
 
Last edited:
Awesome reply, thanks for that :)

DOF isnt something Im used to achieving, typically I am more into very shallow DOF type pics - hence why using anything above f3.5 is pretty new to me!

That DOFMaster chart has really confused me... could you explain what it is telling me??

For instance, at 18mm on a 40D in feet and inches measurement.... If we assume I was 14ft from the focus point, at F5.6 I am already past the hyperfocal point - so everything past 5' 97 will be in focus? Whereas if I'd gone for F2.8 everything past 8' 2.5 until 47'5 would be in focus... then beyond that blurred?

Thanks for any help :)
 
Rather than "beyond that blurred" it's more a case of "beyond that becoming increasingly blurred".
 
The only truly sharp plane should be at the exact distance the lens focused. Before and beyond that distance things gradually become softer and softer. DOF is a calculation that determines how far before and behind the focus point most people would find the image acceptable sharp. There is no hard barrier at which suddenly things fall apart. It's simply that things become just a bit too soft to look great once you get too far from the focused point.
 
tdodd said:
As a simple rule of thumb for wide angle landscape photography try picking an aperture equivalent to the focal length divided by 2. e.g.....

At 10mm use f/5
At 22mm use f/11
At 35mm use f/16 or f/18 (although on a cropper that is no longer "wide", but fine on full frame).

It's simple and it works and it doesn't matter which sensor size you have either. Check out the DOF you will get on the DOFMaster website by following this guideline. It should be enough for most needs.

If you don't need truckloads of DOF then try dividing focal length by 3 and you should still get a good result.

As for apertures at which diffraction begins to become visible at 100% viewing, there is a table listing several cameras here - http://www.the-digital-picture.com/reviews/canon-eos-40d-digital-slr-camera-review.aspx. For the 40D the DLA is f/9.3, which means that if you want loads of DOF you can freely stop down to f/9 without a care. Beyond that you will be buying increased DOF at the expense of absolute sharpness at the focused distance. Nonetheless, if you need the DOF then don't be afraid to go a bit further if you need to, because when viewed at more traditional enlargement sizes, like in print or fit to screen, the diffraction effects up to say f/16, should be all but imperceptible. Going beyond f/16 on a crop body is something to be used with more caution.
Click to expand...

I'm always baffled by those DLA calculations, which take pixel density into the equation. It's a theory that doesn't stand up in practise.

In reality, a high quality lens will peak at f/5.6 on a crop format camera, and f/8 on full frame. The overiding factor is the format (and the degree of subsequent magnification related to it). A super-special lens might peak at a slightly lower f/number, and a less good lens at a slightly higher one. Sometimes you might get centre and edges peaking at different apertures, but that's unusual.

If you click on the test chart images posted in the DigitaPicture link above (short cut here http://www.the-digital-picture.com/...mera=474&CameraComp=474&Sample=0&SampleComp=0 ) the lens in question - an extremely lovely 200L f/2 on a crop format Canon 50D, is quite clearly sharpest at f/5.6 :shrug: As predicted. You have to look very closely, but things gets worse after that.
 
The pixel density matters when you examine 100% crops. The DLA is relevant if you want to examine your files closely, crop hard or print huge.

If you don't want to do that then the DLA has no relevance when comparing results between cameras of equal sensor size. For the same enlargement factor the diffraction will be the same.

But if you bought a 7D, for example, because you wanted to squeeze every last drop of detail from a subject then it is worth understanding that beyond f/6.3 diffraction might just begin to chip away at your results - unless you have a crappy lens that needs stopping down further to hit its IQ peak.

Bob Atkins has a pretty pragmatic approach to this topic, which on one hand means the DLA is not that important, certainly not as important as obtaining the DOF you need if that's where your priorities lie, but ultimately you should be cautioned against stopping down more than you need to in the pursuit of the holy grail of sharpness.

http://www.bobatkins.com/photography/technical/diffraction.html
 
My lens only goes up to f/16 anyway and I'd use it if things got really bright (not likely at this time of year). I always thought f/8 was the "holy grail" of apertures.
 
I've often read that a lens is at is sharpest when stopped down a few stops.

Is there any reason why a lens isn't at is maximum sharpness at its widest aperture? It would seem the reason folk lay down large sums of cash on these lenses is to use it at this maximum aperture.

I can understand if manufacturers chose to place optimum sharpness in the middle of the aperture range so that you have a larger range to use at acceptible sharpness. However if diffraction is going to start eating into this I feel it is not the optimum use of the glass or am I way off on this one:bonk:
 
AndyB1976 said:
I've often read that a lens is at is sharpest when stopped down a few stops.

Is there any reason why a lens isn't at is maximum sharpness at its widest aperture? It would seem the reason folk lay down large sums of cash on these lenses is to use it at this maximum aperture.....
Click to expand...

If I'm shooting wide open then things like the "bokeh" quality matter just as much, if not more, than sharpness.
 
boliston said:
If I'm shooting wide open then things like the "bokeh" quality matter just as much, if not more, than sharpness.
Click to expand...

Yeah but surely sharpness is primary? Bokeh is more a property of the aperture blade design than optical quality.

Stopping down to achieve sharpness on fast glass is going to reduce the CoC reducing the effective blur (which we may have bought the lens for) wont effect the bokeh quality. I am suggesting that the number of aperture blades and their shape determine the bokeh quality and not the actual aperture of the lens itself.

I was merely questioning if there was a reason that lenses aren't optimised at their largest aperture for sharpness, considering you presumably pay for that property and diffraction is going to start effecting quality at around f11/f16.


Sorry for dragging this out, I am a newbie but very interested in optical and technical principles:bonk:
 
AndyB1976 said:
Yeah but surely sharpness is primary?
Click to expand...
At the expense of all other considerations? You are joking, right?

Most lenses are plenty sharp enough to deliver very pleasing results without needing to be stopped down to their theoretically optimum aperture. Sometimes shallow DOF is more important than ultimate lens IQ. Sometimes a fast shutter speed or lower ISO is more important. Some cameras may not even have the resolving power to test a good lens even when used at a sub-optimum aperture for sharpness.

Bokeh (not to be confused with shallow DOF) is a complex thing, and while stopping down a Canon 50/1.8 might well sharpen the lens up it will do nothing at all to improve the bokeh.

Oh, and if you're going to present photos at "normal" viewing sizes as apposed to pixel peeping them to death then you can normally afford a little fuzz at the pixel level.
 
Last edited:
tdodd said:
The pixel density matters when you examine 100% crops. The DLA is relevant if you want to examine your files closely, crop hard or print huge.

If you don't want to do that then the DLA has no relevance when comparing results between cameras of equal sensor size. For the same enlargement factor the diffraction will be the same.

But if you bought a 7D, for example, because you wanted to squeeze every last drop of detail from a subject then it is worth understanding that beyond f/6.3 diffraction might just begin to chip away at your results - unless you have a crappy lens that needs stopping down further to hit its IQ peak.

Bob Atkins has a pretty pragmatic approach to this topic, which on one hand means the DLA is not that important, certainly not as important as obtaining the DOF you need if that's where your priorities lie, but ultimately you should be cautioned against stopping down more than you need to in the pursuit of the holy grail of sharpness.

http://www.bobatkins.com/photography/technical/diffraction.html
Click to expand...

DLA is one thing - a theory - and actual images are another. The images that TheDigitalPicture links to as examples of DLA do not even support the theory - the 200L 2 lens peaks at f/5.6 on a 50D, and that is a crop format camera.

Here's another example, from DPReview, which shows a Canon 50 1.8 on a crop format 450D http://www.dpreview.com/lensreviews/canon_50_1p8_ii_c16/page3.asp It peaks at f/5.6. Then if you load up the same test on a full frame 1Ds3, the peak moves to f/8. Good quality lenses almost always behave like this.

The point is, it is the lens and the format size that makes the biggest difference in terms of where diffraction kicks in, not the pixel count.

The DPR test is also interesting because it shows how the edges peak at a slightly different f/number to the centre.

AndyB1976 said:
I've often read that a lens is at is sharpest when stopped down a few stops.

Is there any reason why a lens isn't at is maximum sharpness at its widest aperture? It would seem the reason folk lay down large sums of cash on these lenses is to use it at this maximum aperture.

I can understand if manufacturers chose to place optimum sharpness in the middle of the aperture range so that you have a larger range to use at acceptible sharpness. However if diffraction is going to start eating into this I feel it is not the optimum use of the glass or am I way off on this one:bonk:
Click to expand...

See above Andy.

In terms of diffraction, it is always lowest at lowest f/number. It is an optical characteristic, unrelated to the quality of the lens (or the sensor/film) but to the physical size of the aperture and the resulting degree of magnification needed to produce a print.

A lens reaches peak sharpness when the aberrations that are reduced by raising the f/number (ie most of them) are overtaken by the loss of sharpness caused by diffraction, which increases as the f/number is raised (and the hole gets smaller). That's when you hit the diffraction limit, and it's downhill from then on.

There are lenses which are sharpest at lowest f/number, but not for cameras. They are very specialist optics designed specifically for very particular technical applications. They do not zoom, they don't change aperture and they don't even change focus - they only do one thing, and they do it very well, eg the collimator lens on an MTF machine.
 
AndyB1976 said:
Yeah but surely sharpness is primary? Bokeh is more a property of the aperture blade design than optical quality. ....
Click to expand...

When you are shooting at f/1.4 the opening is pretty much a pure circle so the shape of the blades will not really have much effect and the pattern in the middle of the frame will be pretty much circular.

eg

14.jpg


However it's not really until you close down to f/2.8 that the shape of the blades has much "shape" as can be seen here:

28.jpg


This example uses the 50/1.4g lens which is not particularly noted for it's bokeh in the same way that the 85/1.4g (or D) is (with its "cream machine" nickname), but I'm sure bokeh wide open is primarily down to optical design rather than blade design.
 
tdodd said:
At the expense of all other considerations? You are joking, right?

Most lenses are plenty sharp enough to deliver very pleasing results without needing to be stopped down to their theoretically optimum aperture. Sometimes shallow DOF is more important than ultimate lens IQ. Sometimes a fast shutter speed or lower ISO is more important. Some cameras may not even have the resolving power to test a good lens even when used at a sub-optimum aperture for sharpness.

Bokeh (not to be confused with shallow DOF) is a complex thing, and while stopping down a Canon 50/1.8 might well sharpen the lens up it will do nothing at all to improve the bokeh.

Oh, and if you're going to present photos at "normal" viewing sizes as apposed to pixel peeping them to death then you can normally afford a little fuzz at the pixel level.
Click to expand...

No not joking:thinking:, I am only asking a simple question. Why is it not optimized at its largest aperture? You can then have all the benefits you list- I'm suprised that having a sharp lens at a large aperture is considered a joke over bokeh quality?

I am not confusing bokeh with shallow depth of field. The reason the 50mm 1.8 does not improve bokeh is not due to sharpness but aperture design. I could be having a laugh here too mind.

Anyway, I was just posing a question in Basics, nevermind.
 
The larger the aperture the more complex the optical and mechanical engineering necessary to gather and bend the light to where it belongs. Light from the lens edges follows a particularly tortuous path, whereas that passing straight down the middle has a comparatively easy ride. By stopping down you eliminate some of the more tricky light from the image, and certain aberrations are reduced/removed. The fact is that with the best will in the world it is extremely unlikely that a lens will be sharper at max aperture than stopped down a bit, even a good one.

So, stop down a bit and you eliminate the dodgiest light from the image, thus sharpening up the picture, but go too far and diffraction will start to become visible and send things the other way. That's why optimum sharpness is usually seen somewhere around f/5.6-f/8. If a lens needs to be stopped down beyond f/8 for the best results then I guess it probably is not a very good lens.
 
Last edited:
boliston said:
When you are shooting at f/1.4 the opening is pretty much a pure circle so the shape of the blades will not really have much effect and the pattern in the middle of the frame will be pretty much circular.

eg

However it's not really until you close down to f/2.8 that the shape of the blades has much "shape" as can be seen here:


This example uses the 50/1.4g lens which is not particularly noted for it's bokeh in the same way that the 85/1.4g (or D) is (with its "cream machine" nickname), but I'm sure bokeh wide open is primarily down to optical design rather than blade design.
Click to expand...


Thats fine..I accept wide open that the blades may not effect the bokeh. I also accept the bokeh is clearly affected as the blade close on the diaphram. Your example clearly shows this.:thumbs:

I always understood aperture blade design effected the bokeh and not the optical quality. If thats wrong....thats why Im in talk basics asking the question.

I am not disagreeing with you here. I was simply asking why a lens is not optimised for sharpness wide open. This would also guarantee better bokeh than having to stop down?

Anyway....I'm gonna have a read at Richards reply. Thanks for the response, much appreciated for coming back with your examples:thumbs:



Andy
 
tdodd said:
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm gives a visual example of how diffraction varies with pixel size (AT THE PIXEL LEVEL)
Click to expand...

Thanks Tim. I've seen that, it often gets quoted in these kinds of debate.

The visual examples are diagrams of the theory, not actual photographs, but that article is also honest enough to list some of the problems with the theory. One area that it only touches on however, is the anti-aliasing filter (which manufacturers don't seem to want to talk about much, I think we know why...) despite the significant but unquantified/unknown effect it has on sharpness.

The bottom line for me is this - if you take photos with a good quality lens on a crop format camera, you will get best overall sharpness at around f/5.6, pretty much regardless. Put the same lens on a full frame camera, and it will be best at f/8.

No theories involved, just reality :)
 
If I get bored tomorrow I might fire off some test shots with my 7D and 100/2.8L Macro and see how soon diffraction spoils the fun for the camera.

For now it's "Nighty night".
 
tdodd said:
The larger the aperture the more complex the optical and mechanical engineering necessary to gather and bend the light to where it belongs. Light from the lens edges follows a particularly tortuous path, whereas that passing straight down the middle has a comparatively easy ride. By stopping down you eliminate some of the more tricky light from the image, and certain aberrations are reduced/removed. The fact is that with the best will in the world it is extremely unlikely that a lens will be sharper at max aperture than stopped down a bit, even a good one.

So, stop down a bit and you eliminate the dodgiest light from the image, thus sharpening up the picture, but go too far and diffraction will start to become visible and send things the other way. That's why optimum sharpness is usually seen somewhere around f/5.6-f/8. If a lens needs to be stopped down beyond f/8 for the best results then I guess it probably is not a very good lens.
Click to expand...


Thanks for the additional response, Tim:thumbs:
 
tdodd said:
If I get bored tomorrow I might fire off some test shots with my 7D and 100/2.8L Macro and see how soon diffraction spoils the fun for the camera.

For now it's "Nighty night".
Click to expand...

Excellent :thumbs: Thanks Tim.

That particular lens is interesting. It is so mega-sharp that it hits a very high peak almost immediately and then proceeds to bounce of the Nyquist ceiling.

Should be good fun :)

Review here, also with some interesting discussion and examples of diffraction http://www.dpreview.com/lensreviews/canon_100_2p8_is_usm_c16/page3.asp
 
I think the DPReview article and demo does a pretty decent job of showing the trade off between extended DOF and ultimate sharpness at the focal point. I don't think there's much point me repeating the exercise. But, if I get really, really bored..... :)
 
tdodd said:
I think the DPReview article and demo does a pretty decent job of showing the trade off between extended DOF and ultimate sharpness at the focal point. I don't think there's much point me repeating the exercise. But, if I get really, really bored..... :)
Click to expand...

Haha ;)

The thing I'd like to see is a comparison of the same lens on two cameras of the same format with different pixel density, maybe a 7D with 18mp against a 40D with 10mp, something like that. That would show what this hypothetical pixel density theory has on the diffraction limit. I guess you could simulate it by cropped your 5D2 down to 8.2mp 1.6x format, and comparing that against the 7D. If you were really bored.

I understand the theory, but I have never seen any evidence of it in practise. In fact, the reverse. On TheDigitalPicture link posted earlier, they do actually test the same lens with many different cameras - 7D, 450D (10mp), 5D2 (21mp full frame) amongst others.

It's hard to see exactly what's going on there because that particular lens - the 200L 2 costing £5k :eek: - pretty much maxes out the sensor straight off the bat. However, if you drive yourself half blind and look closely at the centre and edge images from the 450D and 7D, to my eye they both deliver best overall sharpness at f/5.6. According to the theory, the 7D's 80% higher pixel density should mean it peaks at a lower f/number than the 450D, but it does not. I can't see any meaningful difference, whereas if it really was something worth worrying about, it should stand out pretty clearly.

However, when the camera is swapped for the full frame 5D2, the lens continues to get sharper overall up to f/8.

From which I conclude, pixel density doesn't have any significant effect on the diffraction level, but format size does. And that is really no surprise - we know that happens. It's the reason why compacts often use an internal ND filter to simulate higher f/numbers over f/5.6 becasue diffraction is so horrendous on those tiny sensors at higher f/numbers. Taking the other extreme, a large format film cameras are blisteringly sharp, sharper than any digital camera, at f/64.

Sorry to go on :D
 
AndyB1976 said:
Thats fine..I accept wide open that the blades may not effect the bokeh. I also accept the bokeh is clearly affected as the blade close on the diaphram. Your example clearly shows this.:thumbs:

I always understood aperture blade design effected the bokeh and not the optical quality. If thats wrong....thats why Im in talk basics asking the question.

I am not disagreeing with you here. I was simply asking why a lens is not optimised for sharpness wide open. This would also guarantee better bokeh than having to stop down?

Anyway....I'm gonna have a read at Richards reply. Thanks for the response, much appreciated for coming back with your examples:thumbs:



Andy
Click to expand...

Hi Andy, simple answer to your question about lens sharpness at max aperture is COST.
Take 2 lenses both having F2.8 max aperture, one cost £300 one cost £900, I will say the £900 one will be sharper at F2.8 than the other.
It's marketing really, and the old adage, ' you only get what you pay for'.
ALL except special purpose lenses will be sharper further down the F scale, it's how much down that scale before you get maximum sharpness that determines cost. The nearer to max aperture the more the cost.
 
Thanks for your reply Howard.:thumbs:

I'm happy to have a cheap lens thats not so sharp at maximum aperture. My main point was is there a reason, maybe its technical/optical* that some of the larger expensive lenses cannot be optimised at the larger aperture. My thoughts are if you are paying for such equipment then having to stop it down to acceptable sharpness consequently losing DOF, shutter speed, ISO and possibly infringing bokeh seems a bit of a waste... to me anyway.

I know these high end lenses are optimised in other ways such as build quality, weather sealing, focus motors, they shape and quality of the elements!

Thats all....

Andy
*answered by HoppyUK
 
AndyB1976 said:
Thanks for your reply Howard.:thumbs:

I'm happy to have a cheap lens thats not so sharp at maximum aperture. My main point was is there a reason, maybe its technical/optical* that some of the larger expensive lenses cannot be optimised at the larger aperture. My thoughts are if you are paying for such equipment then having to stop it down to acceptable sharpness consequently losing DOF, shutter speed, ISO and possibly infringing bokeh seems a bit of a waste... to me anyway.

I know these high end lenses are optimised in other ways such as build quality, weather sealing, focus motors, they shape and quality of the elements!

Thats all....
No problem andy, logically what you are saying is indeed what we all want.
This is where you would need to look at lens tests. You will find that there are some lenses that perform far far better at wide alertness such as f2 than other lenses at f5.6. Now! While manufacturers could restrict the aperture to let's say f2.8, where it is sharpest, but would you buy that one as opposed to one that had a wider aperture e.g f1.4 but the same sharpness at f2.8?
There are lenses that are legendary for their sharpness, but, your sensor has to be able to resolve the detail your lens is giving you. There are always going to be compromises here because of the expense. You also have to consider where the sharpness lies, usually at the centre of the sensor area, dropping off towards the edges and corners, the higher quality lenses giving good sharpness across the whole frame, this is where we have to accept compromises, I may have a lens that is sharpest at f2.8 but if I have the option of stepping down to f1.4, I will get a respectable shot in low light with a good shutter speed, even though I may be losing some sharpness.
There are lots of other factors at play here in getting a good shot, it is not always down to ultimate image sharpness. As with all things, while we may be able to measure sharpness, actually seeing the difference can be another thing.
Best not to worry about it too much, I can't see thing changing from how it is now. This hobby is no different to any other, it is based around the premis of dimminishing returns, you can go on paying vast sums of money for only a marginal increase in whatever, it is usual that we accept a level of quality v cost that we can afford.
Sorry to ramble on a bit, but I'm sure you can see where I am going here, after all we can only buy what the manufacturers offer us.

Andy
*answered by HoppyUK
Click to expand...
 
Right, there has at last been enough light to attempt a test to demonstrate diffraction at increasingly small apertures with what I hope is an acceptable level of control over the test conditions and ensuing results. The test was performed in sunshine using a tripod with legs collapsed, a lead (or similar) packed ankle weight to steady the camera and lens, mirror lockup, 2 second timer and remote release to try to negate camera shake. The target was at about 60' away, chosen to limit the effect of focus error, and focusing was performed using Live AF in Live View and confirmed with 10X magnification. Focusing was performed on the edge between the black wooden board and the pebbledash. The lens was my 100/2/8L Macro lens.

As I can only include six images per post here are the results at f/32, f/22, f/16, f/11, f/8, f/5.6. f/4 looks good and f/2.8 is a bit soft. These are 100% crops, processed in DPP using Standard picture style, sharpening = 3 and NR set to 0,0.

20110203_142834_4161_DPP.JPG
20110203_142829_4160_DPP.JPG
20110203_142823_4159_DPP.JPG
20110203_142819_4158_DPP.JPG
20110203_142814_4157_DPP.JPG
20110203_142810_4156_DPP.JPG


To my eyes it's a close call betwen f/5.6 and f/8 but I'm pretty sure that f/5.6 does look sharper, which would fall in line with the theoretical DLA being at f/6.8. By comparison f11 looks just about OK and f/16 is probably usable, if downsized. But at f/22 and f/32 the pixels absolutely suck. This is not a lens fault. It is not a camera fault. It's simply that the 7D's high pixel density really reveals how bad diffraction is at these apertures.

Perhaps the easiest way to observe the effect is by watching the shots as a slideshow. I think then that it becomes very easy to see where sharpness peaks and begins to tail off again. Here's the slideshow, starting at f/2.8 through to f/32....

http://picasaweb.google.com/EezyTig...gCISZ27bZ46OXfg#slideshow/5569474086519624466
 
Last edited:
Excellent stuff Tim :) That was the 7D? I would guess optimum sharpness is perhaps between f/4 and 5.6. Which is just what you'd expect with that particular (exeedingly sharp) lens.

It would be interesting to see it on full frame, when I would expect the peak to move to f/8.

I know that you didn't test for this Tim, but in your experiments, did you form a view on whether pixel density has any significant effect on diffraction, or the point at which it starts to hurt sharpness in a noticeable way?

BTW, f/22 and f/32 are unbelievably bad! :eek: I'm wondering if this is where the pixel-density-airy-disc-clash business shows itself? :thinking:
 
It was the 7D. I have not performed the test with my 5D2. There's just about enough light to give it a go now. Back later.....
 
You're a star! :)
 
Right, results from the 5D2 follow. Same setup as for the 7D, and of course as 100% crops the samples below have undergone less enlargement than those from the 7D so they should look sharper, if smaller, at each corresponding aperture value. Here are f/32 through to f/5.6....

20110203_152804_7052_DPP.JPG
20110203_152759_7051_DPP.JPG
20110203_152755_7050_DPP.JPG
20110203_152750_7049_DPP.JPG
20110203_152746_7048_DPP.JPG
20110203_152742_7047_DPP.JPG


Here's the slideshow in forward sequence from f/2.8 to f/32....

http://picasaweb.google.com/EezyTig...CL28qYL2qN7n7AE#slideshow/5569487553171791410
 
Last edited:
You must log in or register to reply here.
Back
Top