An example of sharpness versus DOF for different apertures

[GardenersHelper]

As part of the preparations for a video I am making I produced the following example. I thought I would share it here as it might be of interest to those who are new to close-up/macro.

Here are two images of the same dead fly. They don't quite line up because I captured them hand-held. And the colours look a bit different even though the white balance is almost identical. I have no idea why this is. However, I don't think either of these differences is relevant to the subject of this post.

The images were both captured as raw files with my FZ330 bridge camera and KX800 twin flash, using a Raynox 250 close-up lens. The scene width is about 14mm so the magnification is about 1.5:1 from the point of view of an APS-C camera of the type many people here use. Both images have been given identical batch processing in DXO Optics Pro 11 and Silkypix Developer Pro 7 using the parameters I normally use before doing any image-specific post processing. In this case they have both had only the batch processing, and no image-specific processing.

The first image was captured using f/2.8. This gives the same depth of field (DOF) as f/8 on Micro Four Thirds, f/10 on APS-C and f/16 on Full Frame. These four settings also suffer from the same amount of loss of sharpness from diffraction.

The second image was captured using f/8. This gives the same DOF as f/22 on Micro Four Thirds, f/28 on APS-C and f/45 on Full Frame. with the same loss of sharpness from diffraction in all four cases (which is a much greater loss of sharpness than in the first example).

Looking at the eye in the first image, which was captured with the larger aperture, we can see that the half of it that is in focus it is sharper than the same area in the second image. On the other hand, much more of the subject is in focus in the second image, which was captured with a very small aperture. The second image has much greater DOF. But it is not as sharp.

Which is better, a smaller amount in sharpest focus, or a larger amount in less sharp focus? I think that is a matter of personal taste and preference. (btw, with stacking you can have your cake and eat it, with lots in focus and best sharpness for all that is in focus.)


0986 1 F2.8 on FZ330 = F8 on MFT = F10 on APS-C = F16 on FF by gardenersassistant, on Flickr


0986 2 F8 on FZ330 = F22 on MFT= F28 on APS-C = F45 on FF by gardenersassistant, on Flickr

 

[dibbly dobbler]

Greater DOF winds hands down for me. Interesting as I typically restrict myself to F16 on aps-c due to worries about diffraction - might try some smaller apertures.

Thanks for sharing

 

[Brachytron]

Certainly no problem with sharpness there.

 

[Nod]

Given that the subject's static, I'd split the difference and try all 3 options - wide open, stopped right down and somewhere in the middle - say f/5.6 in this case. Wide open rarely gives the sharpest results and as you point out, diffraction rears its ugly head at f/11 or thereabouts.

 

[MattMail]

I prefer the sharper one

 

[GardenersHelper]

Greater DOF winds hands down for me. Interesting as I typically restrict myself to F16 on aps-c due to worries about diffraction - might try some smaller apertures.

Certainly no problem with sharpness there.

I prefer the sharper one

That's one of the fascinations of this stuff for me, how differently we all can react to what we see.

Given that the subject's static, I'd split the difference and try all 3 options - wide open, stopped right down and somewhere in the middle - say f/5.6 in this case. Wide open rarely gives the sharpest results and as you point out, diffraction rears its ugly head at f/11 or thereabouts.

Need to watch out for the equivalences here. This was a small sensor camera that doesn't go to f/11 - it only goes from f/2.8 to f/8, as used for these two shots. This is the equivalent of f/10 to f/28 on APS-C, so diffraction will have already kicked in even for the first shot at f/2.8 and is really destructive of sharpness and detail at f/8. But some people (I'm one of them) still prefer shots with this massive loss of sharpness/detail because of the increased DOF. In fact, all my shots of insects, spiders etc that I post here are at the equivalent of f/28 on APS-C, f/45 on full frame. I only use larger equivalent apertures for flowers (although I do go to f/32 on APS-C for flowers occasionally), and for insects etc when the subject is very small in the frame - strongly "environmental" shots.

I think most people here prefer to use around f/11 on APS-C for insects etc.

 

[Tim Illingworth]

The first one has more atmosphere to it, This would be more noticeable if the fly was alive perched on something with the back ground further away.

 

[User.81858]

Thanks Nick. Yet another interesting thread by you. I find that the illustrations really help and I prefer number 2.

 

[Terrywoodenpic]

Diffraction is related to the size of the aperture.
So here is a thought experiment ... though it is easy to do in practice.
Place a small aperture, say pinhole size, in front of a ground glass and and inspect the image. like all pinhole images it is totally diffraction limited and the image is formed by such rays as pass straight through the aperture over laid by the wave form diffraction set up by the aperture edges. This seriously softens the contrast. and give the usual pinhole look.

Then with the same pinhole aperture place over it a simple meniscus lens and bring it to focus on the ground glass. you will now find that the image is really quite good. as the small aperture has reduced many of the aberrations found in such a simple lens to insignificance. However the image is still comparatively soft, because it is affected to the exact same degree of diffraction as the pin hole was.

From this one can see that the effect of Diffraction is in some way additive. And this overlay of diffused light ( Airy rings) is limiting the sharpness in much the same way a light mist would.
There is always some improvement in apparent sharpness, if the contrast is raised slightly in processing. but it can never bring back lost detail.

Even with diffraction limited images, increasing the depth of field by reducing the aperture, will always bring more of the image into focus. but it will never increase the captured detail. It will usually still be a "Better" image. as was demonstrated by @GardenersHelper images.

As Depth of field and depth of focus are linked. Reducing the aperture also improves the back focus. as the increased depth of focus makes up for poor parallelism and placement of the sensor. This is more often the case with large format cameras, which are rarely made to such high tolerances.

 

[GardenersHelper]

The first one has more atmosphere to it, This would be more noticeable if the fly was alive perched on something with the back ground further away.

Thanks Nick. Yet another interesting thread by you. I find that the illustrations really help and I prefer number 2.

Thanks both. I love the fact that we each have our own take on these things, our own preferences.

 

[GardenersHelper]

Diffraction is related to the size of the aperture.
So here is a thought experiment ... though it is easy to do in practice.
Place a small aperture, say pinhole size, in front of a ground glass and and inspect the image. like all pinhole images it is totally diffraction limited and the image is formed by such rays as pass straight through the aperture over laid by the wave form diffraction set up by the aperture edges. This seriously softens the contrast. and give the usual pinhole look.

Then with the same pinhole aperture place over it a simple meniscus lens and bring it to focus on the ground glass. you will now find that the image is really quite good. as the small aperture has reduced many of the aberrations found in such a simple lens to insignificance. However the image is still comparatively soft, because it is affected to the exact same degree of diffraction as the pin hole was.

From this one can see that the effect of Diffraction is in some way additive. And this overlay of diffused light ( Airy rings) is limiting the sharpness in much the same way a light mist would.
There is always some improvement in apparent sharpness, if the contrast is raised slightly in processing. but it can never bring back lost detail.

Even with diffraction limited images, increasing the depth of field by reducing the aperture, will always bring more of the image into focus. but it will never increase the captured detail. It will usually still be a "Better" image. as was demonstrated by @GardenersHelper images.

As Depth of field and depth of focus are linked. Reducing the aperture also improves the back focus. as the increased depth of focus makes up for poor parallelism and placement of the sensor. This is more often the case with large format cameras, which are rarely made to such high tolerances.

Very interesting. Thanks.

I had never thought about the depth of focus aspect of this. One of those things that is obvious once you have it pointed out to you, for which, thanks again.

 

[davholla]

Interesting comparison, I think the second photo is better although I would say a tripod would have made this a fairer comparison.

 

[FishyFish]

Technically, I think number 2 is the best. It gives better detail in the subject.

Artistically however, I prefer number 1 - it looks nicer (although with maybe a tad more DOF required).

 

[GardenersHelper]

Interesting comparison, I think the second photo is better although I would say a tripod would have made this a fairer comparison.

Thanks David.

I'm not sure what a tripod would have added in terms of fairness. They were both captured at 1/1600 sec so there would be no issue of blur from them being hand-held. And in the first one the in-focus area lies between out of focus areas further from and nearer to the camera, so there isn't any question of DOF being "lost" in front of the subject, and so we are seeing the full extent of the (rather narrow) DOF. I know they don't line up exactly laterally, but I think the illustration of the difference in the extent of the DOF is fair. Or have I missed something here? (Very possible - you know how I can miss things! )

 

[GardenersHelper]

Technically, I think number 2 is the best. It gives better detail in the subject.

Artistically however, I prefer number 1 - it looks nicer (although with maybe a tad more DOF required).

Thanks. You are not alone regarding what looks good. (for example see @Tim Illingworth comment above). Then again, neither am I (My preference - with insects, spiders etc, but not so much with flowers - is towards the other end of the DOF spectrum.)

I think it's great for newcomers to close-up/macro to see from our reactions that there is no right or wrong about this, but to understand how big a difference aperture can make to the look of a close-up/macro image.

 

[Rupert67]

As someone who hasn't done much macro work in the past (just a few insects hand held with a Canon EF 180mm f/3.5L macro on my 5D III) I'm interested that you find a loss of sharpness in the second picture. Given that you usually lose some image definition when it's posted to a web page, I'd say the sharpness is pretty darn good!

The problem I always have with any attempt at macro shots is the common issue of DoF. Consequently I usually wind the ISO up a bit so that I can hand hold at f/11 or f/16 (f/22 if I can) and still get more of the insect's body in focus. I'm well aware of the so-called sweet spot of my lenses but, short of image stacking, I don't know how else to achieve a decent DoF.

 

[GardenersHelper]

As someone who hasn't done much macro work in the past (just a few insects hand held with a Canon EF 180mm f/3.5L macro on my 5D III) I'm interested that you find a loss of sharpness in the second picture. Given that you usually lose some image definition when it's posted to a web page, I'd say the sharpness is pretty darn good!

You need to look closely. Notice in the comparison below that on the left there is an area of the eye where we can see the "cells" in the eye quite clearly. The boundary between that area and the rest of the eye is rather narrow and beyond that you can't see even any hints of the cells. On the right it is much more difficult to make out the cells in that area - the eye is not as sharp. There are hints of the cells, rather low contrast and difficult to see. What is more there are faint hints outside that area, in fact right over on the other side of the eye.

So, the area that is in best focus on the left is sharper than the area that is in best focus on the right. However, more is in focus on the right, but none of what is in focus on the right it is as sharp what is in focus on the left.


Dof differences close up by gardenersassistant, on Flickr

The problem I always have with any attempt at macro shots is the common issue of DoF. Consequently I usually wind the ISO up a bit so that I can hand hold at f/11 or f/16 (f/22 if I can) and still get more of the insect's body in focus. I'm well aware of the so-called sweet spot of my lenses but, short of image stacking, I don't know how else to achieve a decent DoF.

With a large aperture, as you have been doing, but you may need to use flash to get the sharpness you would like. Working hand-held you need strong light to get a fast enough shutter speed to get a really sharp image. That issue is accentuated by using a longer focal length lens. You are probably aware of the "rule" (which obviously varies from person to person) that suggests that the shutter speed should be at least 1/focal length in order to get a sharp image when the lens doesn't have image stabilisation. As the magnification increases you need a faster shutter speed that this "rule" suggests.

Another thing you can do, especially with the kit you are using, is to take the photo from further away. You can then crop the image to get the subject back to the size it would have been if you had taken the photo from closer in. This approach - less magnification and more cropping - will give you greater DOF. The downside is that you lose detail by cropping. But with the kit you are using you should be capturing huge amount of detail and should be able to crop a lot. This does depend on the image being good and sharp though, so you may be back to using flash anyway, but it may be worth trying with natural light to explore the technique. I don't know how much post processing you do, but you may be surprised at how much you can crop and still get a nice image.

There is one more issue. How much DOF you get depends on the aperture you use. (DOF roughly doubles for every two stops decrease in aperture.) However, how much DOF you get at a particular aperture depends on the size of the sensor. For example, I use a camera with a very small sensor, with a minimum aperture of f/8. I have a micro four thirds camera which has a minimum aperture of f/22 (with the lens I use on it). I get the same DOF at f/22 on the micro four thirds camera as I do at f/8 on my small sensor camera. To get the same amount of DOF using a full frame camera like yours you would need an aperture of f/45. But the minimum aperture on your Canon EF 180mm f/3.5L is f/32. That is one stop less than f/45, so you will get 40% or so less DOF at f/32 than I get at f/8 on my small sensor camera or f/22 on my micro four thirds camera.

If you were using flash, or in very bright light, you might be able to use f/32. But there is something else to bear in mind. As you decrease the aperture your images lose sharpness and detail because of diffraction. By the time you approach the apertures I use (f/8 and f/22), and f/32 on your kit would fall into this category, there is no significant difference in resolution independent of sensor size, lens or number of pixels. So by using f/32 you will be reducing your kit to about the standard of my bridge camera with its tiny sensor and general purpose zoom lens. (It's not quite that straightforward as that, but it is pretty close).

If you look at some of my images you might think that is ok, which is fine. But you will be warned off using such very small apertures because of the impact (and it is a very real impact) of diffraction.

If you want maximally detailed images that make best use of your very good equipment, and you want lots of DOF, then I think your only alternative is to stack. On the other hand, if you would be content with lower quality images like the ones in this post over at dpreview (all of which were captured at the equivalent of f/45 for your camera), then you don't need to stack. You may need to do some other post processing though to improve the look of the images as they come out of the camera when you use very small apertures.

 

[Rupert67]

You need to look closely. Notice in the comparison below that on the left there is an area of the eye where we can see the "cells" in the eye quite clearly. The boundary between that area and the rest of the eye is rather narrow and beyond that you can't see even any hints of the cells. On the right it is much more difficult to make out the cells in that area - the eye is not as sharp. There are hints of the cells, rather low contrast and difficult to see. What is more there are faint hints outside that area, in fact right over on the other side of the eye.

So, the area that is in best focus on the left is sharper than the area that is in best focus on the right. However, more is in focus on the right, but none of what is in focus on the right it is as sharp what is in focus on the left.


Dof differences close up by gardenersassistant, on Flickr



With a large aperture, as you have been doing, but you may need to use flash to get the sharpness you would like. Working hand-held you need strong light to get a fast enough shutter speed to get a really sharp image. That issue is accentuated by using a longer focal length lens. You are probably aware of the "rule" (which obviously varies from person to person) that suggests that the shutter speed should be at least 1/focal length in order to get a sharp image when the lens doesn't have image stabilisation. As the magnification increases you need a faster shutter speed that this "rule" suggests.

Another thing you can do, especially with the kit you are using, is to take the photo from further away. You can then crop the image to get the subject back to the size it would have been if you had taken the photo from closer in. This approach - less magnification and more cropping - will give you greater DOF. The downside is that you lose detail by cropping. But with the kit you are using you should be capturing huge amount of detail and should be able to crop a lot. This does depend on the image being good and sharp though, so you may be back to using flash anyway, but it may be worth trying with natural light to explore the technique. I don't know how much post processing you do, but you may be surprised at how much you can crop and still get a nice image.

There is one more issue. How much DOF you get depends on the aperture you use. (DOF roughly doubles for every two stops decrease in aperture.) However, how much DOF you get at a particular aperture depends on the size of the sensor. For example, I use a camera with a very small sensor, with a minimum aperture of f/8. I have a micro four thirds camera which has a minimum aperture of f/22 (with the lens I use on it). I get the same DOF at f/22 on the micro four thirds camera as I do at f/8 on my small sensor camera. To get the same amount of DOF using a full frame camera like yours you would need an aperture of f/45. But the minimum aperture on your Canon EF 180mm f/3.5L is f/32. That is one stop less than f/45, so you will get 40% or so less DOF at f/32 than I get at f/8 on my small sensor camera or f/22 on my micro four thirds camera.

If you were using flash, or in very bright light, you might be able to use f/32. But there is something else to bear in mind. As you decrease the aperture your images lose sharpness and detail because of diffraction. By the time you approach the apertures I use (f/8 and f/22), and f/32 on your kit would fall into this category, there is no significant difference in resolution independent of sensor size, lens or number of pixels. So by using f/32 you will be reducing your kit to about the standard of my bridge camera with its tiny sensor and general purpose zoom lens. (It's not quite that straightforward as that, but it is pretty close).

If you look at some of my images you might think that is ok, which is fine. But you will be warned off using such very small apertures because of the impact (and it is a very real impact) of diffraction.

If you want maximally detailed images that make best use of your very good equipment, and you want lots of DOF, then I think your only alternative is to stack. On the other hand, if you would be content with lower quality images like the ones in this post over at dpreview (all of which were captured at the equivalent of f/45 for your camera), then you don't need to stack. You may need to do some other post processing though to improve the look of the images as they come out of the camera when you use very small apertures.

A really clear response to my post, Nick .. many thanks indeed. As a matter of interest I always shoot raw and post process in PSE 13.

On the subject of shooting from further away, it's an interesting point that you make. My 'other camera' (a Canon Powershot G1-X) produces the best macro shots when set on 'macro' but not zoomed right in close. I've experimented a good deal with this and, because the sensor on that little camera is a reasonable size, crops can still result in fine sharp images.

Thanks again for your comprehensive tutorial!