PDA

View Full Version : Problem with back focus.... I think



Granville
20-02-2013, 8:23pm
I was reading some threads here and on another site about sharpness in photos because I haven't been overjoyed with what has come off my camera so far and since I'm still new at this, wanted to make sure that if there wasn't anything else wrong with the camera/settings/etc, then it was probably my technique.

I came across a number of discussions on back focus, and what it meant. There were also some test that could be done to see if it's an issue on a particular camera, so I mocked up what I thought may be a good test, but I don't know enough about it to know if it's an issue for me.

I simply set my tripod up, and put amy guillotine from my study on a table outside. I then focused on the 15cm mark at various apertures to see what the results would be like. All shots are taken with my Nikon D3100, 50mm 1.4G prime, using single point focus. Sharpening has previously been set to 7 in Standard Picture Controls and is still on that setting.

This first image is at 1/500s f3.5. It seems to me that the focus area is from 15cm to 10cm. and I'm not sure that is ideal.
97708

This second photo is taken at 1/3200s, f1.4, and all the focus zone is from about 18cm to 15cm. Maybe front focus.

97709

This next one is at 1/100s, f8, and DoF is quite deep, but still, it looks like the focus zone is between maybe 21cm all the way back to 5cm.

97710

As I said, I read about this for the first time today, so I'd really like an opinion from those who know about these things to see if I have an issue. I have been really reluctant to put any photos up here as they really lack the sharpness that others have, and I haven't known how to correct it.

Thanks i advance

arthurking83
20-02-2013, 10:45pm
Two things that you need to be aware of with focusing issues .. ie. that are vital to eliminate prior to contemplating any servicing/fixing/adjustments to be made.

1. fast aperture lenses are very susceptible to focus shifting. Google focus shift and what it really means.
Briefly; this is a condition where you have your lens and subject fixed relative to each other, but as you alter the aperture value, the focus or the clearest line of focus appears to shift fore-aft compared to another aperture value. At some aperture value(dependent on lens design) this focus shift seems to cease.
So as a random example of what this could result in, is say with your 50mm, at f/1.4 focus plane seems to be at the 17cm mark(instead of 15cm) and gradually move back as you stop down to about f/2.8. Bit from f/2.8 and smaller, the focus shift has ceased, but now DOF makes it slightly harder to judge the exact plane of focus.

2. your test focus target is not ideal .. and this is the most common failing that most folks find themselves with when doing focus checks.

There are many opinions about the net as to what constitutes the best focus target, and some people are happy to pay for these targets too.

I've found that AA batteries are about the best focus check target. Any brand any type, but the size of the AA battery is quite accommodating for most camera focus systems.

Considering your preliminary results shown here, there may well be a focusing problem with your lens, but that actual extent should be determined before action is taken.

Find yourself 5 AA batteries. One of them will be set in the middle position, and the other four will be setup equally on either side of the middle battery.
these periphery batteries will be placed far enough apart so as to not intrude into the path of the focusing square of the camera when focusing on any other battery.
the two batteries directly on either side of the central one, should be set to that the front face of these two will be about half way into the depth of the middle battery. The other two batteries either side of those two again, will be set the same depth back again.

ie,. 5 batteries in a v formation(like birds flying in formation) where the two outermost batteries are placed where their front face is about level with the rearmost face of the middle battery, and the two inner side batteries are set half as deep as those two outer batteries.

The actual distances are not vital, other than they help to determine how far back or forward focus inaccuracy actually is. Don't be fooled into thinking that you want finer granularity in depth, as the actual inaccuacy may not be apparent, same deal with going too deep too. I found that each layer of battery set back half the depth of the actual diameter of the battery works best.

Of course you don't have to use batteries, any similarly sized item around the house works.
But you need to have everything set level and as perfectly square as your patience can muster a few mm of slackness this way and that way doesn't really make much difference .. it's only a general idea of the problem that you are actually looking to find.

A very important point to be sure that the focus system is given a good chance to prove itself.
Good amount of lighting must be used at the time of focus. if that means shining a bright torch on the subject, then all the better.
AND the focus square of the cameras focus points should only see one item at a time, and must be set so that the focus square is contained within that target only. It should not be set so that the square cover the subject and a little bit more! The subject you are trying to focus on, must surround the focus square and also extend itself beyond those boundaries. This is to completely eliminate other contrasty sections of a scene from diluting the focus points ability to focus where the operator actually wants it too focus on.

FWIW, while I wasn't there at the time of your test, the problem with this test is that the focus square has too much contrasty data to focus on. That is, even tho you say you focused on the 15, the focus square must surely have also been covering more that just the 15, and will certainly have also seen the small dashes, and probably some of the other numbers .. at least the 14 and 16 too .. maybe even more.
So the only certainty in a test like this, is that where the focus point was actually focusing on, is uncertain.

I don't know where you would find this info now, but the way these focus systems work is via the use of a set of lines inside those focus squares. Usually the central square is a cross type focus point .. so if you can imagine, the focus system in engineering terms looks something like this .. [+] this is the central point(cross type)

On consumer level bodies, the other focus points rely on a vertical line for focusing .. so those points will look like this ... [|]

This is a very vague rendering of how they actually look, but it gives you an idea.

the real issue is in the accuracy level of the engineering/manufacturing process, and we can only assume that the cross points or horizontal lines are going to be directly in the centre of the square.
I reckon the odds are very short that these lines that do the actual focusing, are not exactly central to the square .. and may all have varying levels of skew, or shift. Problem is that nobody knows.

But a long time back on chap did bother to check, and he found that at least one of his focus points on an early model Nikon, was actually so far out that the vertical line was half way out of the actual focus square in terms of it's vertical accuracy, and also shifted to the far right of the square too, almost to the edge of the square itself.

I think once you understand the pitfalls of doing these tests, you at least better equipped to not make fundamental errors or assumptions that lead you into an endless circular path of wasted effort.

Also note that others believe that the best way to test or check for focusing is to set the subject 20x the focal length of the ... blah blah.
I say rubbish, and I generally test for massive failures in performance at distances I'm more likely to use.

ie. 50mm .. about 1-1.5m .. and test at both those distances .. but not limited too those distances .. I have just enough time and patience to also test at 2m as well .. etc, etc.
(yeah well 1m is 20x focal length, but that's not the point! :p)
In your case here tho, there is no reason you shouldn't do the testing at 45cm .. which is far less than 20x focal length :D

.. hope this helps.

Granville
21-02-2013, 8:49am
Helps a lot. Thanks for all this. I'll digest it again and set up a test with bateries, but this time just so I can see the results, and not to draw definitive conclusions just yet.

Thanks again.

agb
21-02-2013, 9:06am
I have read that it is not good to use batteries, because they are round, and there is a chance that the camera will get confused as to what to focus on, front or side edge. It is much better to use square edged subjects, such as I show here. These old video boxes are staggered at an angle and the tape measure allows for a reasonably accurate gauge as to what is in focus and what is not. The focus point was then on the centre box.

arthurking83
21-02-2013, 4:35pm
it's the rounded nature of the batteries that swayed me too them in preference to squared off material. I used to use redheads matchboxes as well.

While there is always the possibility that the focus grid may lock onto part of the side of the battery, therefore resulting in a slight shift in the focus, the shift is only going to be a few millimeters anyhow. ie. pretty much insignificant if focus is locked onto to the actual subject. That is, DOF takes care of these very few mm of inaccuracy anyhow.

But the reality is that this has never really happened to me with my gear(that I've tested). I've had/got issues with two lenses, and the rounded shape of the battery helps to visualize how close AF fine tune is getting to the 'correct mark' as you set adjustment.
(FWIW: both lenses were not set with AF fine tune as they were zoom lenses, and setting AF fine tune at one focal length then ruins the focus accuracy at another).

But the rounded nature of batteries also helps to determine how the focus system works for ya in many real world situations, for example such as focusing on a person's eye. The human eye is rounded to begin with, and also in situations where you have your subject angled and not square on, you may have part of the ear directly behind the eye, which could also cause the focus system to get confused and focus on the 'wrong' spot.

Sometimes making the test procedure a little more complex can reveal insights as to how overall performance may work in the field.
But making the test too complex, and much more prone to errors is not recommended.

In the case of my two lenses, one of which I still have, I just leave the lens without AF fine tune, and compensate myself to account for the very slight backfocus I get at the long end of the zoom range.

AF fine tune, whilst a great feature is not the last word in focus accuracy. Lens design, the engineering within the camera's focus system, the available lighting, operator technique .. they all play a greater role in achieving better focus accuracy in a consistent manner.
I mainly use AF fine tune as an assistance to determine if there is in fact a problem, and if I had a problem that is overly annoying with any particular lens, I'd much prefer to have it serviced and checked to work better without resorting to AF fine tune anyhow.

milesy
21-02-2013, 7:51pm
I have read that it is not good to use batteries, because they are round, and there is a chance that the camera will get confused as to what to focus on, front or side edge. It is much better to use square edged subjects, such as I show here. These old video boxes are staggered at an angle and the tape measure allows for a reasonably accurate gauge as to what is in focus and what is not. The focus point was then on the centre box.

where do you find VHS tapes these days??

agb
21-02-2013, 8:50pm
where do you find VHS tapes these days??

Still got a heap in the cupboard and a player and an old TV to play them on. I have no idea whats on them except for one that is called "In Flinders Wake" I think.
My wife thinks I am some sort of hoarder, but I dispute that.

Mark L
21-02-2013, 10:20pm
where do you find VHS tapes these days??

Plenty at our place. We currently even run the PVR through the VHS player. Older TV (still) and I'm not the hoarder and old video watcher! My wife disputes that hoarder bit.