Having trouble focusing wide angle lens for big DOF

[FlyingSquirrel]

I'm having quite some difficulty properly focusing my Canon 10-22mm lens for landscape shots with small aperture (f/22) and very near to very far subject matter (a few feet near, to infinity). My results are very inconsistent, meaning sometimes the near is in focus but not the far, and vice versa. And sometimes, it seems, in between (is that even possible?!)

Also, with f/22, I am getting diffraction which is making it slightly more difficult to judge if my focus is on point. There could also be a little noise contributing, but I doubt it, since I am using ISO as low as I can, usually between 100-400, on my 7D.

I read (and have experienced) that for wide angle lens with huge DOF as I describe, focusing is extremely unforgiving if you are off even a little. This could obviously be incorrect, but it seems to be accurate based on my results. I have been focusing manually with the distance scale on the lens barrel, so I am thinking that may be the problem as it is not a precision method.

So anyway, I guess I am hoping for some advice/guidance/help on how to figure out where to focus, at what distance, and if I should do it manually or somehow use AF, and working with near subjects out to infinity, and f/stop, etc. I have no problems with telephoto, macro, and other photography in general; only this wide angle landscape issue is being a stickler

Thanks for any assistance.

[Shadowman]

You should find this book or search for the author through youtube.

http://www.kinokuniya.com/sg/index.p...=9780770433055

He suggests focusing one third the distance into your intended composition for best deep depth of field and to use the smallest aperture (f/29 or more) your lens is capable of achieving.

[Colin Southern]

Lookup hyper-focal distance (calculator).

Generally, the DoF of an UWA lens stopped right down is so huge that it's pretty hard to miss with the focus -- so I suspect that your issue is probably more to do with sharpening, diffraction, and lens quality.

Case in point; on a 7D with a 10-22mm lens - at the hyperfocal distance of 24.260cm @F22, your DoF ranges from 18.845cm to Infinity.

[Manfred M]

A very strange problem indeed. With a lens that wide at f/22 (especially at 10mm) everything from a few inches in front of the lens to infinity should be in focus. If I look at the CiC DoF calculator at the 10mm setting everything from 23cm to infinity should be "in focus" at f/22 for a 1.6mm crop frame. At 22mm, things are not quite as forgiving, but still, anything from 3 or 4 feet out to infinity shold be fine . I shoot two super-wides; the Tokina 11-16mm on my crop frame and a Nikkor 14-24mm on my full frame and can't ever recall having a problem like the one you describe. Generally focusing can be quite sloppy when you shoot wide.

I would do exactly what John suggested on my old manual focus prime lenses as their distance scales tended to be very good; but I don't find the markings (if they even exist) on the autofocus ones can be relied on as much.

I wonder what else is in play here?

[Colin Southern]

He suggests focusing one third the distance into your intended composition for best deep depth of field and to use the smallest aperture (f/29 or more) your lens is capable of achieving.

It works OK for some situations, but not for others.

Case in point - the example above, where 24.260cm (being the mathematically correct focusing distance for maximum DoF) isn't anywhere remotely "approx 1/3 of the way in to a scene that "end" in infinity.

[HaseebM]

Can the DoF preview button work in such situations?

[Manfred M]

Can the DoF preview button work in such situations?

DoF preview was always a rough guide, but at f/22 the image in the viewfinder is going to be so dark as to be useless.

[nickjohnson]

Hello Matt. A bit of information that I hope you will find useful.

It's my understanding that the majority of modern zoom lenses are not true zooms – but actually vari-focal lenses. By vari-focal I mean that the point of focus changes as one zooms the lens to change the focal length. Accordingly, the focus scale on the lens can only be accurate for one focal length (zoom) setting. My experience is that the lens focus scale is accurate at the longest focal length setting. So using such scales in practical situations is often difficult and sometimes impossible.

So, what to do? Try the following:- Set a moderate aperture – say f11 or f8. Set the focus to the furthest thing you want sharp – usually that thing is a long way in the distance and effectively infinity. Take a picture. The result is sure to be non optimal in a technical seance – but is should be pleasingly, and apparently evenly, sharp - front to back. You may even find that the very close bits - whilst blurred - are subjectively fine for most practical purposes.

The technique described is not my own, and is only good for normal to wide angle perspective shots. Be aware that it cannot work if the foreground is very close since the reproduction ratio is starting to approach the macro / close focus range for the close objects.

Have fun.

[DanK]

I find this puzzling. The greater the DOF, the more forgiving focus is. If you look at a DOF table or DOFmaster, it shows that at 10 or 15mm and f/22, it should not make a big difference where the lens is focused, unless the near points are less than 2 ft away. I'm pretty sure that Nick is right that the 10-22 is not a parfocal lens, but with a DOF this large, the minor changes in focus that one gets with a vari-focal lens shouldn't matter much at all.

And sometimes, it seems, in between (is that even possible?!)

Nope, I don't think so, if the lens is working properly.

[RustBeltRaw]

It's possible (though probably not likely) that a lack of corner sharpness is being perceived as improper focusing. I'd recommend trying a shot set with the previously-mentioned hyperfocal calculations, centering the background in one shot, and the foreground in another. Comparing the center of each shot will let you rule out corner issues, and let you know which part is out of focus.

I find the best way to get maximum depth of field is to stop the lens down as much as I want (f2-f22, scene-dependent ), focus at infinity (or the furthest point I want sharp), then pull the focus back until that part goes soft (judged in 10x live view if possible). If the foreground's in focus, great. If not, then I decide which one I'd rather have sharp, or revisit my aperture choice. Mostly, I do this with my Tokina 11-16mm f2.8, and partially because I think its focus scale is off a tad.

[FlyingSquirrel]

Thank you guys so much for your helpful replies. I found much of the information useful and learned some things I did not know.

Simply for my own knowledge and reference, I have decided to do an extensive and thorough series of test shots. I've listed out a variety of testing scenarios and methods of carrying them out. All that remains is to actually follow through with said tests. I don't really want to do it, but I think it will reveal a lot about exactly how my specific lens performs at various "focal lengths," apertures, and focusing distances; in the long run the time and effort I will spend on the tests will probably give me some insight into how to use my gear to get the best image quality and DOF I need.

I would have to agree with Nick that it seems that the distance scale on my lens barrel is not accurate at all zoom lengths, which might explain my inconsistent photograph results. Of course, Dan has a point about the DOF covering most of that difference, but my tests should help clarify that (I hope).

Time to work myself up about the tests...woohoo, let's do this! I love doing tests! Yeah! (LOL)

[jcuknz]

The 1/3--2/3 focusing rule to be found in advice on focusing does not always apply and I suspect with a UWA lens where subject material is close to camera it may not be. In the one example I worked out from hyperfocal tables the third rule only applied from 25ft out to infinity and as you got closer the came back to 50/50 at around six feet. However last time I mentioned this somebody pointed out that it varies considerably depending on the lens and aperture. So depending on your subject matter and its distance from the camera this maybe upsetting things.
Personally I rarely go for deep focus but try to establish what the viewer is likely to want to be sharp and focus on that point. Usually anothing behind maybe soft but because it is far away that is not noticeable.
The other thing I did recently when the subject lent itself was to take two photos and use the sharp parts of each in editing ... but I have only tried this once, and it worked I think, and do not know if it is a viable approach ... I don't have a UWA lens so am not tempted to try deep focus shots

[RustBeltRaw]

The 1/3 front, 2/3 back guideline mainly applies to portrait shooting and the associated focal lengths. It's supposed to encourage one to autofocus on a subject's face, then go manual and push the focus a little toward the subject's ears to use all of the depth of field. Works reasonably well for close-to-medium distance subjects with 50-200mm lenses, but falls apart with any lens as focus approaches infinity.

At or near infinity, it's more like 1/50 front, infinite back. Yes, fellow math nerds, I know that isn't even a real ratio - just illustrating that there's very shallow depth of field between the camera and the focal plane, and a very deep field on the far side of the focal plane. At close-to-macro focus distances, 1/2 front, 1/2 back is more accurate.

[chauncey]

Matt, after looking at your flickr site I see your problem...you're way out of your comfort zone.
You do great photographs of things but, no landscapes there yet. Anyway...

As you noticed, diffraction sets in at just shy of f/16 in most canon gear,
if you must have a longer DOF you might try focus stacking using something like f/8.
I've only tried it once and wasn't careful with focusing points...looked like a POS.

I do better shooting thingys as well.

[Polar01]

Matt: when using a UWA lens as I do a lot for landscapes, here is what I do, once set up as I usually do pans, is if I am shooting at 21mm wide or equal to you 13mm x canon crop approx. 20.8mm the f-stop it f/18 I do not need any more than this, I know that if I auto focus on something 6 feet away that everything from 2.4ft to infinity well be in focus. Now I turn the auto focus off and simply shoot.
As for the distance marked off on the lens barrel, do not even bother with it, the reason it as the throw distance of autofocus cameras is so short the distance can not be judged. It is not like a manual focus lens where the throw is huge, that is a large amount of turning of the dial only makes a small adjustment to the focus.
A good general rule is at 30mm or less at f/18, manual focus on something 10ft away then everything from 4.5ft to infinity will be in sharp focus. For you with a crop Canon camera that would be anything showing under 19mm.
Check out the pan I did at Cedar Falls using this method in the "Landscapes and Architecture" threads.

Cheers:

Allan

[FlyingSquirrel]

Thanks again, guys. I was playing with the DOF Master website calculator, and found some information which I find very interesting. This may or may not be surprising to any of you, and may or may not be interesting to you. Most of you seem to have enough experience that it will not be a surprise, but I've included it anyway. And also, I have more remarks at the bottom of the post below this data:


----
7D
*10mm*
f/22

Focus at 1 ft, DOF is from 0.44 ft to infinity (DOF is "practically infinity")
Focus at 10,000 ft, DOF is from 0.76 ft to infinity (DOF is "practically infinity")
Focus at Hyperfocal Distance 0.8ft, DOF is from 0.4ft to infinity (DOF is "practically infinity")

So, at 10mm at f/22, you can focus almost anywhere and have practically infinite DOF

----------
7D
*12mm*
f/22

Focus at 1 ft, DOF is from 0.53 ft to 7.95 ft (total DOF is only 7.41 ft)
Focus at 1.2 ft, DOF is from 0.58 ft to Infinity (DOF is "practically infinity")
Focus at 10,000 ft, DOF is from 1.1 ft to infinity

So, at 12mm, 0.2 ft difference in your focus point is the difference between a DOF of only 7.41 ft vs a DOF of practically infinity
----------
7D
*22mm*
f/22

Focus on subject at 1ft, DOF is from 0.8 ft to 1.34 ft (total DOF is only 0.54 ft)
Focus at hyperfocal distance 3.77 ft, DOF is from 1.885ft to infinity (DOF is "practically infinity")

So, at 22mm, 2.77 ft difference in your focus point is the difference between a DOF of 0.54 ft vs a DOF of practically infinity


----

I feel like this could explain some of sharpness issues I seem to be having. Since I was shooting at no more than f/22 at any given focal length (between 10mm up to 22mm), I could obviously have been focusing in the wrong range, thereby giving me insufficient DOF. It's important to note, very often I try to have foreground within just feet of my lens.

I still feel like diffraction is contributing to the apparent "out of focus" appearance of the photo, but it seems to be more than that. I will do some real life testing at various apertures to compare diffraction amounts and its effect on the apparent focus.

I also think that the quality of my lens is pretty poor. In some of the photos where I've managed to have seemingly good focus and DOF, it's quite soft on the corners/edges, and has quite a bit of chromatic abberation. Not very satisfied with the lense's IQ. Shockingly, B&H reviews for it have quite high marks, and comments about the sharpness and excellence of it. It may be possible I got a lemon. It may also be possible (and likely) that those reviewers don't have as stringent of standards as I do for IQ and are considering the quality of a downsized, sharpened image, as opposed to real IQ at 100%. Maybe some day I'll look into something better quality. I wonder if Canon has L series UWA.

[jcuknz]

Since DoF calculators have to have a value to work to they can only be a guide and should be treated with caution depending on the final product. The aspect of DoF that many ignore or are not aware of becuase it is rarely mentioned is that one has to start with a given Circle of Confusion and this changes with the size of the image produced and how close it is being designed to be looks at from. The CoC size many tables are based on is/was a 10x8 inch print viewed from a distance of 13 inches. So the DoF will be less if you are making a 20x16 unless you can stop the viewer from getting as close as 13 inches

The CoC used is the size of a circle of confused image that appears as a sharp dot to the viewer ... bearing in mind that Depth of Field is an illusion and only the light from a given object at which the lens is focused is truely sharp .... light from other objects is converging or expanding circles of confusion at the sensor or filmplane but look sharp so long as they are not magnified too much.

I guess having written that I will find that there is a CiC tutorial explaning it much better

[Colin Southern]

I guess having written that I will find that there is a CiC tutorial explaning it much better

https://www.cambridgeincolour.com/tu...h-of-field.htm

[rpcrowe]

As mentioned above, you cannot use a DOF guide as effectively with today's auto focus zoom lenses as you could with manual focus prime lenses because:

Most of today's auto focus lenses do not have a comprehensive distance scale that the older manual focus prime lenses included. In fact, some of today's lenses (albeit usually the less expensive varieties) have no distance sale at all. An example of this would be the EF 50mm f/1.8 MK-ii lens. It has no distance scale while the previous Mk-i model included the scale.

A standard feature of the focus scale on a manual focus prime lens was/is a double set of f/stops as shown in his image..

The yellow line indicates the distance at which the lens is focused and then you can read the DOF (area of "acceptable" focus) from where marks for the f/stop you are using match to the close and far side of the distance scale. Placing the indicator for the f/stop you are using on the infinity mark will automatically focus the camera at the hyperfocal distance and the DOF will be from infinity to where the equivalent f/stop mark matches with the near side of the focus scale. In this case the camera is focused at the hyperfocal distance for shooting at f/22. The DOF will be from slightly less than 7-feet to infinity.

Simple, quick and easy. I frequently used this system with manual focusing prime lenses when I knew that I would not have time to manually focus

However, a zoom lens is only at the focal length indicated when it is focused at infinity. At shorter focusing distances, the focal length is different, sometimes quite different, from what is marked on the lens. If you are not sure of your focal length, using the hyperfocal distance focusing system cannot be accurate...

BTW: focusing on the hyperfocal distance will not give you gnat-sharp focus from half the hyperfocal distance to infinity. It gives you "acceptable" focus which may or may not be what you want.

That said, I have had virtually no problems auto-focusing when I use my 12-24mm f/4 Tokina, especially towards the wider end of the focal range. I usually shoot in the area of f/8 to f/11...

[DanK]

I feel like this could explain some of sharpness issues I seem to be having. Since I was shooting at no more than f/22 at any given focal length (between 10mm up to 22mm), I could obviously have been focusing in the wrong range, thereby giving me insufficient DOF. It's important to note, very often I try to have foreground within just feet of my lens.

Yes, but if you check the hyperfocal distances, you will see that as long as you don't focus very close, the point of focus should not matter much even at 22mm. At 22mm and f/22, as long as you focus at least 4 feet out, you should have DOF to infinity. so including close foreground material is not a problem unless (1) it is closer than about 2 feet, or (2) you focus on it, rather than at the hyperfocal distance.

As a couple of people said, DOF is not a range in which everything is tack sharp; its the range within which sharpness is acceptable.

I recently bought the same lens as a refurb from Canon. I put it through the usual tests, e.g., shooting a brick wall at different apertures and then pixel peeping. I could clearly see some softness in the corners, particularly wide open (not your problem). I returned it to Canon, and they sent it back saying that it was within specs. So some amount of corner softness must be normal. However, all the reviews suggest that for real uses, rather than pixel peeping, it performs well.

Re CA: I haven't tested that on mine yet. However, CA (particularly lateral CA) is very easy to correct in post. In LR, which I us, you just have to check the box for the lens correction for that lens (which it will detect from the metadata). Longitudinal CA takes more work, but LR has tools for that as well.