Problem: hyperfocal distance or lens quality?

[dbe4876]

I am experimenting with using hyperfocal distance for landscape photography. This is my first attempt with hyperfocal distance so its likely the issues I'm encountering are due to me! But, when I look at these shots in their original size view, there are areas of both pictures that look undefined and fuzzy. If someone can help me zero in on what might be the cause I would very much appreciate the learning opportunity.

Equipment: I used a Canon 30D with Tamron SP AF17-50 f/2.8 Di-II lens, on a Ynatran pro tripod with an ATH-918 power grip ball head, cable release, polarizing filter, and Cokin P120 gradual ND filter. For the two shots below I set my camera to AE mode, chose an aperture of f/22, put the lens in manual focus mode and set the focus to 7 feet.

First shot, sunrise over the marsh...

(To see the original sized image, click to go to my smugmug gallery. Then, click the "O" at the top of the page for the original image)
*Mechanics: ISO 100, 23mm, f/22, 2.5s, Canon 30D, with Tamron 17-50mm SP AF Di-II lens
*I admit to a little post-processing warming in this shot as the sun had washed out the sky and the marsh was a bit too blue.

Now, the issue I'm wondering about is the fuzziness you see in the marshes. You will need to view the image in original size to see this. But, there are a lot of areas of the marsh that look just plain fuzzy and undefined.

Second shot, sunrise over the marsh with back to the sun...

(To see the original sized image, click to go to my smugmug gallery. Then, click the "O" at the top of the page for the original image)
*Mechanics: ISO 100, 23mm, f/22, .5s, Canon 30D, with Tamron 17-50mm SP AF Di-II lens

When you view the original size of this image you can see a similar fuzziness in both the marshes and the trees in the distance. There is simply a lot of detail missing! At this point I really don't know if the problem is due to improperly using hyperfocal distance, lens quality, limitations of my 30D's senor, or other. I did note while shooting the last shot that my Cokin filter had begun to develop frost (it was a little chilly!). But that was pretty obvious in affected the images - little spots and blotches all over the place. So, I don't think the problem I am seeing here is about frost. Any ideas?

[joedecker]

So, I don't think the problem I am seeing here is about frost. Any ideas?
[a href=\"index.php?act=findpost&pid=107387\"][{POST_SNAPBACK}][/a]

A couple, but it's difficult to be certain.  I've got a laundry list of things I'd check for and experiment with, though.

First, with frost, it's possible to get a bit of that on your lens, but I tend to notice small amounts of it first more as a "veil" of lower contrast rather than as a loss of detail, so it's my guess that that's not what's going on.

Second, do you know about diffraction?   While "the more you stop down, the more depth of field you get", there's a small blurring effect that comes from light diffracting around the edges of apertures that starts creating a bit of softness when you stop down quite a bit, a bit at f/11, a little at f/16, and noticeably at f/22.  Try the same shot at f/11 and f/22 with near focus and compare objects at the distance yo focused at to get a sense of how big this effect is.

In both images you don't indicate what focal length you actually shot the images at, the right hyperfocal distance is very, very different for 17mm and 50mm--it's almost impossible to get the focus in the *wrong* place at 17mm, most of these shots would be pretty tricky at 50.  I'm wondering if you were nearer 50 in the second image, you may just not be able to get everything in focus.

I'm also wondering if a few of the reeds in the front of the first image moved during the exposure.

Also, was your tripod on something dead solid?

Finally, I don't know the lens at all, but it's always possible that it's a little to blame--but if you've had and continue to have other images that are significantly sharper at similar focal lengths, the lens is probably not to blame.

[dbe4876]

In both images you don't indicate what focal length you actually shot the images at, the right hyperfocal distance is very, very different for 17mm and 50mm--it's almost impossible to get the focus in the *wrong* place at 17mm, most of these shots would be pretty tricky at 50.  I'm wondering if you were nearer 50 in the second image, you may just not be able to get everything in focus.[a href=\"index.php?act=findpost&pid=107495\"][{POST_SNAPBACK}][/a]

Sorry. The Tamron was set to 23mm for both images. From the chart I looked at it seems hyperfocal distance would be about 1.2m at f/22. So, setting my focus to 7 feet should have been safe, shouldn't it?

[DarkPenguin]

Focus on the things you want in critical focus.  Then set your f stop to get the dof you need to pull everything else in (front and back).

[howiesmith]

Why set the focus to the hyperfocal distance.  Only objects at that distance are critically focused.  That my be in front of anything in the photo.

I would determine the closest and farthest distance of objects I want to be "in focus" and calculate the focus distance accordingly.

What CoC did you use for the hyperfocal distance calculation and why?  Why then just set the lens to f/22?

At f/22, you are likely seeing some diffraction effects.

[joedecker]

Sorry. The Tamron was set to 23mm for both images. From the chart I looked at it seems hyperfocal distance would be about 1.2m at f/22. So, setting my focus to 7 feet should have been safe, shouldn't it?
[a href=\"index.php?act=findpost&pid=107505\"][{POST_SNAPBACK}][/a]

Yeah, that should be sharp, save for diffraction, from about 2 feet out, "for some value of sharp".  (I'm using a COF of 0.03, which may not be right but it's served me well enough.)  Like you I show the true HD distance as being closer than that, which should have given you a reasonable margin of safety to bring the far trees into sharp focus.

I'm thinking it's very likely you're getting hurt by diffraction.

As far as the closest objects, how close were they?  If they were more than about 2.5 feet from the camera, you very likely would have gotten away with f/16 -- down to 2.0 feet if you nailed the hyperfocal distance.

[dbe4876]

Okay, just to throw a monkey wrench into this, I just located a couple of images that I happened to take at f/13 (all the rest were at f/22)...

(ISO: 100, 23mm, f/13, 8s)

And here's a section of the image at 100%...



You can see right off the bat that we have a difficult subject matter here. This marsh grass is going in every direction, and there's a lot of it mixed in-between the cattails. It seems that the farther out you go from the camera, the harder this stuff is go get any definition on. I did also use the Cokin ND filter on this shot, so that may be a factor however slight. So too the polarizer. But I'm inclined to think it's a combination of difficult subject matter, limits of the lens, and likewise the limits of the digital format. What do you guys think?

It'd be interesting to compare these kinds of shots with a film camera and a 23mm Canon L-lens!

[Eric Myrvaagnes]

I would suggest trying the same scene (or a very similar one) with no filters at all, at a couple of different apertures (perhaps f/8 through f/22) to see just what effect the lens opening is having, without any possible effect of filters. Then post some of them, with 100% crops, and if they still all look fuzzy, there may be a problem with the lens.

[dbe4876]

By the way, I used DOFMaster, set to a CoC of 0.019. The DOFMaster website listed that value for the Canon 30D. I set up the chart for my lens as follows:

Short lens length: 17
Long lens length: 50
Distance Units: feet
Minimum Distance: 7
Maximum Distance: 200
Minimum f-number: f/2.8
Maximum f-number: f/32
Circle of confusion (mm): 0.019.

When I find 23mm on the X-axis, at 7 feet on the Y-axis, and follow it up I see f/12.7 (pretty dern close to f/13;-)

Now, the question is - how do I use this information correctly? If everything from 1/2 the hyperfocal distance is supposed to be in focus, do you focus your lens to 3-1/2 feet? Or, do you focus it at 7 feet, and make sure the closest subject matter is no closer than 3-1/2 feet? Is the difference academic?

Ah! Just found my answer at DOFMaster...

When the lens is focused on the hyperfocal distance, the depth of field extends from half the hyperfocal distance to infinity.

Photography, Phil Davis, 1972.

[dabreeze]

some of the later posts in this previous thread may lend a hand in understanding the practical application of hyperfocal theory:
Hyperfocal Distance

[Doug J Scott]

But, when I look at these shots in their original size view, there are areas of both pictures that look undefined and fuzzy. If someone can help me zero in on what might be the cause I would very much appreciate the learning opportunity.


Dave,

Let's look at your first image only - your issue is simple. Revisit this scene, or one like it, set your lens to f/8, dial your shutter for proper exposure, focus about 1/3 of the way across the scene or at infinity, then post your sharp results for us.

I suggest f/8 because your lens is likely its sharpest there, and a smaller aperture really is not needed for this landscape. But even diffraction from your f/22 will not cause the OOF softness on the horizon in each of your scenes. It is just a simple matter of racking your focus closer to infinity.

[dbe4876]

Revisit this scene, or one like it, set your lens to f/8, dial your shutter for proper exposure, focus about 1/3 of the way across the scene or at infinity, then post your sharp results for us.
[a href=\"index.php?act=findpost&pid=107728\"][{POST_SNAPBACK}][/a]

Thanks, I'll do just that! Ah, another opportunity to get up before sunrise on a Saturday;-)  Funny how the more I get decent pictures (or results) the more I feel like rousing myself from a Saturday morning slumber. Ha!

[Jack Varney]

Dave,

If you're going to be up late tonight it would probably be fine, for sharpness purposes, to go later in the day.  

[dbe4876]

Popped over to the lake after work and captured some shots at f/8. Before I left the house, I used Dudak's Depth-of-Field Calculator, which gave me the following:

Coc: 0.019
*Lens Focal Length: 24
Lens f/stop: f/8
Focus Distance: 12.5
Near Depth: 6.233
Far Depth: Inf
Total Depth: Inf
Hperfocal Distance: 12.433

*At the lake I tried to set my lens to 23mm, but actually I ended up with 24mm. That's why I listed 24 as the lens focal length.

I set the camera to AE mode, set my f/stop to f/8, and set the lens to manual focus and tried to focus at about 12 feet (the last foot number on my lens is 7, before the next marker at infinity, so some guessing is required). The camera was also on a tripod, I used a cable release, and set the drive mode to timer, with the mirror lock-up turned on - when you press the cable switch the mirror locks up and two seconds later the camera takes the image. This is my standard procedure for landscape shooting.

Now, here's a shot from a week ago or so, at f/22...
[a href=\"http://dbe.smugmug.com/gallery/2474765/13/137531385#137530779-L-LB\" target=\"_blank\"] I left the ISO at 250 from a family shoot this last weekend. So, we do have a little ISO difference right off the bat. By the way, I did not use the Grad ND filter tonight as the cloud cover evened out the sky all by itself.

Now, for some 100% shots. First, from the f/22 shot, and from the left side of the frame...
[a href=\"http://dbe.smugmug.com/gallery/2474765/13/137531385#137530602-O-LB\" target=\"_blank\"]

[DarkPenguin]

Marginally related to this...

http://www.naturephotographers.net/article...7/rb0307-1.html

Saw that article today and this seemed like a good place to post a link to it.

[Ray]

I don't know about you guys, but I'm seeing more definition from the f/22 shot. Yes, the ISO difference may be in play, but how much? In any case, shooting at f/8 did not produce a steller improvement as was hoped. I know there is a difference in the circumstances too, so the comparison may not be all that great - it would be better to see comparisons from the same shoot with different f/stops. And, of course, there's always that "user error" factor we refer to in the IT world. Okay, I'm open for comments! 
[a href=\"index.php?act=findpost&pid=107797\"][{POST_SNAPBACK}][/a]

I suggest you download a Norman Koren test chart (sorry I haven't got a direct link at hand) and test the resolution of your lens at various f stops.

It's generally the case that 35mm lenses are sharpest at f8, but that's not a hard and fast rule by any means, particularly in the case of zooms where the sharpest aperture can vary depending on focal length selected. Some zooms at certain focal lengths are actually sharpest at f16 and only imperceptably less sharp at f22.

Lens performance can also vary at different focussing distances. I've taken shots with my 20D/100-400mm/1.4x extender combination at f22 which seem sharper than the same scene at f11 (in any part of the image). But I've also taken close-ups (using the same set-up) which are definitely sharper at f11 than f22.

Know thy lenses!

[Doug J Scott]

In any case, shooting at f/8 did not produce a steller improvement as was hoped.

Dave,
 
F/22 only looked better to you because it was adding depth of field, somewhat compensating for your incorrect front-focusing of this scene. Notice how sharp the foreground is compared to the horizon?

As I suggested before, please shoot a similar scene just by focusing further into it (the 1/3 rule) and you'll see it sharpen right up, and yes, at f/8. Your shooting at 24mm where your depth of field will extend way back toward your camera from your point of focus.

Then shoot one with the lens AF'd on the horizon - yep, at infinity. Both efforts will produce a sharp horizon and you'll likely be amazed at how far the sharpness extends toward the camera.

If I were there shooting this scene with you, I'd simply thumb the AF button on the grass about 75 feet from you, set the camera on the tripod, mirror-up and shoot at f/8, no worries.

Forget the charts and try it?

[dabreeze]

If I were there shooting this scene with you, I'd simply thumb the AF button on the grass about 75 feet from you, set the camera on the tripod, mirror-up and shoot at f/8, no worries.

Given that the hyperfocal distance for 24mm @ f/8 is aprox 12' (using a fairly standard CoC for a 20D), then any focusing distance beyond 12' (@24mm @ F/8) will be hyperfocal also (that is, acceptably sharp from half the distance to the focus point to infinity).

So, Doug, I challenge you to go out and shoot a similar scene yourself. My guess is that if you have any kind of true infinity (that is, like these sample pix, something genuinely far off), your oft-repeated but very fallible rule of 1/3 focusing will come up short.

That's b/c if you have anything in the image foreground closer to you than (.5)75' or 37.5' it will, by the very nature of hyperfocal theory, begin to lose focus.  And that's using your guestimate of 75' which, in the scene in question here, looks to be much closer than a third into the scene.

Either way, one third into a scene or 75' or so, is too imprecise to work all the time. Sure, depending on where your close objects are, what focal length you're using with what aperture, you're gonna get results some of the time.

But if you want real control over sharpness in the images you shoot, you're gonna have to look into hyperfocal distances and theory. Don't even think about shooting traditional wide angle, forced perspective, foreground-background landscape shots ala David Meunch without doing some homework as to where exactly in the image you need to be focusing.

Dave: more than likely Ray has suggested the path to proceed down here. you may simply be getting less than optimal results from a less than optimal lens that may have some calibration problems. it's amazing how easy it is to knock optical elements off kilter and go crazy afterwards. YOur on the right track vis-a-vis understanding hyperfocal theory to achieve the control you want. you just may not have the lens to do it with?

Good luck,

Derek von Briesen
Sedona, Arizona
www.pbase.com/sedonamemories

[Doug J Scott]

So, Doug, I challenge you to go out and shoot a similar scene yourself. My guess is that if you have any kind of true infinity (that is, like these sample pix, something genuinely far off), your oft-repeated but very fallible rule of 1/3 focusing will come up short.

No guessing needed, Derek - the proof is in the imagery one creates.

Dave, rather than post a sample image here, I invite you to take a brief stroll through say, the Interiors section of our web site. This section alone shows a number of wide-angle images with in-focus areas extending all the way from 4 or 5 feet to infinity. Not one of these images was focused by calculating hyperfocal.

Best wishes on your efforts. Keep it fun.

[howiesmith]

So, Doug, I challenge you to go out and shoot a similar scene yourself. My guess is that if you have any kind of true infinity (that is, like these sample pix, something genuinely far off), your oft-repeated but very fallible rule of 1/3 focusing will come up short.

www.pbase.com/sedonamemories
[a href=\"index.php?act=findpost&pid=107828\"][{POST_SNAPBACK}][/a]

The math is pretty simple here.  No experiments needed.

The near limit of DOF is:

Hu/(H+u)

where H is the hyperfocal distance and u is the focus distance.  When u = H, the equation reduces to:

near = H/2 or u/2.  This shows that for relatively large values of H (compared to the lens' focal length), the near limit is about half way to the hyperfocal or focus distance, not 1/3.

I use Ansel Adams "The Camera" as my reference.

One third is more correct for small focus distances where the lens focal length cannot be ignored.