I could use some help and advice with my lens - front/back focus and Micro Adjust

C

clicktor

Messages
863
Edit My Images
Yes
Hi all,

I currently have the following kit:

1: Canon 50d
2: Sigma 18-50mm f/2.8 EX DC Macro
3: canon EF 50mm f/1.8 mkII (nifty fifty)

I have previously recognised soft focus on portrait shots and fairly close focusing tasks indoors with both lenses. Photos of my kid would come out soft and more often the tip of the nose would be more in focus than the eyes which were focused on. I performed the popular "battery test" (lining up batteries a different distances about), normal target test parrelle to camera, and also the lens chart test where you place the target at 45 degrees to the camera and focus on the FOCUS HERE line so you can clearly see the depth of field spread to see front/rear focus.

I know what you might be thinking...the lens chart tests at 45 degrees are prone to issues. Well... my findings leave me a little confused. In summary using tripod/mirror lock up etc making sure all settings are constant and good for testing, I am finding consistently that both lenses show front focus when close to minium focus distance of the lens, but when I go back the recommended 50 x the focal length, hardly any micro adjust is required. Why is this?

Why, when I use the lenses for close up targets am I seeing much more front focus? The 45 degree lens chart test shows it best, but even flat targets do as well. I am starting to get frustrated and just want to know if my gear is ok. One of the reasons I went with the 50d was because I thought my lenses front focused a little and that I could correct it, but how can I possibly even use my lenses for closer distances if when I adjust for this (+10 or so) it then makes it out at longer distances. This is not even taking into account different focal lengths of the zoom which makes it even more complex.

Is it a case of, you get what you pay for and this is to be expected?

I tried my friends 55-250mm and it's bang on at 250mm, but it's not as faster lens with more depth of field. Hmm.... I know it can get obsessive and the recommendation is to just go out and shoot, but I want to shoot at shallow DOFs and I want to know my gear is good and not holding me back.


Thoughts?
 
If both lenses behave similarly, it would seem to be body related. What happens if you put your lenses on your friends camera?
 
Shrug... I know my nifty fifty needs a little adjustment on the 5D2, but I've not noticed it needing different adjustments at different focal lengths. I'll have a 7D (and a couple more lenses) in a couple of weeks too, so will be having doing this myself soon. No experience of the Sigma though.
 
So, are you saying that the longer distance shots are actually out of focus after micro adjust vs non adjusted? If so, roughly by how much?
 
Sorry for confusion. I mean like this:

1: Distance to target is <1m but more than MFD of lens = requires +10 micro adjust (i.e. lenses front focus by about 8mm and need correcting)

2: Distance to target is 50x Focal Length so 2.5m or more @ 50mm = lenses look best at 0 micro adjust or maybe at the most +1 or +2.

I'm beginning to wonder if I should send the lenses to someone else to test as I think it might be me! :( It's 100% not the body. Could not possibly be like it on 5 different bodies.
The Sigma also seems more prone to front focus when at 50mm full zoom focal length. I could post some images I guess tonight but I mean all it will show is what I have described.
 
Aah I see......I've micro adjusted several of my own lenses but I've never noted how much it affects the focus at varying distances, though I have to say I always notice a lens front or back focussing at shorter distances to the subject which is what prompts me to tweak em.

Perhaps it's a depth of field thing? y'know like the shorter distance to subject = narrower depth of field, farer = wider DOF.......So the wider DOF needs less adjustment cos more stuff falls into focus anyway?
 
Yeah. I might retest. Seen a couple of other ways of doing MFA tests and adjustments.
 
If I had a pound for every time I've posted not to AF test at artificially close distances, I could take a nice holiday.

The nifty 50 is a pretty low grade lens mechanically and you honestly can't expect too much by way of mm accuracy at all distances. I tried three recently and none of them was either dead accurate or consistent at close range, but two of them were okay for actual picture taking (one was faulty I believe). I don't know about the Sigma but that is also a cheaper lens for the spec. But having said that, if you adjust it for realistic distance shooting, even at f/1.8 you'll probably find that either you or your subject will move further than the lens will misfocus, and that's assuming you have nailed the right spot to focus on in the first place.

To be blunt, if you are going to micro examine everything and pixel peep at 100% then you need to buying the best quality kit you can get, and even then you will not get perfect performance at every focal length and at all distances, every shot. And unless you have a pukka macro lens, things will almost always be worst at close range.
 
HoppyUK said:
If I had a pound for every time I've posted not to AF test at artificially close distances, I could take a nice holiday.
Click to expand...
If a lens focuses to a minimum distance, and it gets used at close to that distance, why is that considered artificially close? The magnification from most lenses at mfd isn't anywhere near macro values...
 
The thing is one of the reasons I bought the Sigma 18-50mm f/2.8 EX DC Macro, was because the MFD is only 0.20m, and in fact, often it focuses closer than that with no issues. This is really handy for portraits indoors but not when thet are coming out soft. I mean I see talk of using lenses at more "realistic/common" distances. Well I want to use my lens on fairly close targets indoors. I don't think that's too much to ask? How do others cope? Buy better lenses?
Hoppy I know to test lenses you suggest to use distant targets, I spoke to you about it before actualyl in other threads. Maybe I am expecting too much.
 
arad85 said:
If a lens focuses to a minimum distance, and it gets used at close to that distance, why is that considered artificially close? The magnification from most lenses at mfd isn't anywhere near macro values...
Click to expand...

By 'artificially close' I mean testing at a focusing distance that you would not normally use. Usually because the depth of field at longer range makes it very much harder to detect any errors, but that then puts the lens right at the edges of the tolerance envelope, errors creep in, you correct them with microadjust, and then the lens is out at normal range.

So you have to be realistic about it, accept that AF systems are mechanical analogue devices and not 100% perfect at all times, and set the best compromise. But also bearing in mind what I said above about keeping it real in terms of where the biggest errors are likely to come from in practise, ie subject/camera movement and user error etc.

clicktor said:
The thing is one of the reasons I bought the Sigma 18-50mm f/2.8 EX DC Macro, was because the MFD is only 0.20m, and in fact, often it focuses closer than that with no issues. This is really handy for portraits indoors but not when thet are coming out soft. I mean I see talk of using lenses at more "realistic/common" distances. Well I want to use my lens on fairly close targets indoors. I don't think that's too much to ask? How do others cope? Buy better lenses?
Hoppy I know to test lenses you suggest to use distant targets, I spoke to you about it before actualyl in other threads. Maybe I am expecting too much.
Click to expand...

By a distant target, I simply mean something realistic - which doesn't usually involve filling the frame with a sheet of A4 paper. I'm really not sure why you should want to shoot portraits at 0.2m... :thinking:

I think the problem might just be that you're unlucky with your lens choices, and then pushing them a bit too far in the direction where they're weakest. Sigma 18-50 2.8 is known not to have the best AF (see TheDigitalPicture review, amongst others) and you honestly can't expect much of the 50 1.8.
 
HoppyUK said:
By 'artificially close' I mean testing at a focusing distance that you would not normally use. Usually because the depth of field at longer range makes it very much harder to detect any errors, but that then puts the lens right at the edges of the tolerance envelope, errors creep in, you correct them with microadjust, and then the lens is out at normal range.

So you have to be realistic about it, accept that AF systems are mechanical analogue devices and not 100% perfect at all times, and set the best compromise. But also bearing in mind what I said above about keeping it real in terms of where the biggest errors are likely to come from in practise, ie subject/camera movement and user error etc.
Click to expand...
Yup... understood.

I guess that's why the more expensive lenses have full time manual focus and why large viewfinders are important. I know when we had the 450D, swapping between it and the 5D2 it was very obvious the difference in view the different viewfinders give...
 
HoppyUK said:
By 'artificially close' I mean testing at a focusing distance that you would not normally use. Usually because the depth of field at longer range makes it very much harder to detect any errors, but that then puts the lens right at the edges of the tolerance envelope, errors creep in, you correct them with microadjust, and then the lens is out at normal range.

So you have to be realistic about it, accept that AF systems are mechanical analogue devices and not 100% perfect at all times, and set the best compromise. But also bearing in mind what I said above about keeping it real in terms of where the biggest errors are likely to come from in practise, ie subject/camera movement and user error etc.



By a distant target, I simply mean something realistic - which doesn't usually involve filling the frame with a sheet of A4 paper. I'm really not sure why you should want to shoot portraits at 0.2m... :thinking:

I think the problem might just be that you're unlucky with your lens choices, and then pushing them a bit too far in the direction where they're weakest. Sigma 18-50 2.8 is known not to have the best AF (see TheDigitalPicture review, amongst others) and you honestly can't expect much of the 50 1.8.
Click to expand...

Yeah fair enough. I think I obsess too much! I'm really happy with the Sigma otherwise. The nifty fifty ironically despite it being good for low light experimentation, is best for me out in daylight shooting stopped down to 2.8 or so. But then that can be said for many lenses.

The Sigma I do not use for portraits at 0.20m all the time, but I mean sometimes I will want a close up full head filling the frame shot of my little-un, which the nifty might chooke at since it only has an MFD of 0.45m. I dunno, maybe I'll send it to you hoppy to try out! :)
 
HoppyUK said:
So you have to be realistic about it, accept that AF systems are mechanical analogue devices and not 100% perfect at all times, and set the best compromise.
Click to expand...

I reckon that pretty much nails it. A couple of times I found myself initially dialling in large microadjust values, like +/- 20, then later on backing off to 10-12 when I realised something wasn't quite right. I only have 1 lens now that actually requires a -20 setting to get consistent results. FYI....Sigma 50-150mm f2.8 EX DC HSM (Mk1).
 
clicktor said:
Yeah fair enough. I think I obsess too much! I'm really happy with the Sigma otherwise. The nifty fifty ironically despite it being good for low light experimentation, is best for me out in daylight shooting stopped down to 2.8 or so. But then that can be said for many lenses.

The Sigma I do not use for portraits at 0.20m all the time, but I mean sometimes I will want a close up full head filling the frame shot of my little-un, which the nifty might chooke at since it only has an MFD of 0.45m. I dunno, maybe I'll send it to you hoppy to try out! :)
Click to expand...

Nooo! :eek:

You could send it to Sigma and they'll make sure it's as good as it can be. Maybe it's just a bit out of spec and they can tighten it up - worth a try. There is more sophisticated adjustment available to them than you can do yourself with MA. Depends on the lens but I suspect the 18-55 has some internal adjustment, whereas I'm pretty sure the 50 1.8 does not.
 
Well I could post up pictures of charts etc but I know real life examples are best. Here is a shot I took tonight in the lounge. White balance set to tungsten, wide open at f/2.8. Yes I know shutter is very slow but I was supporting it on a cushion. The point is, look at the focus of the eyes area (where I focused on center point) and it appears some of his Tshirt and the duplo is more in focus (i.e. front focusing). I wouldn't call this "artifically close", rather an every day indoors snapshot. Yes it's also grainy at ISO 800 with no processing what so ever, just trying to show the focusing. Am I asking too much focusing on eyes in above head tungsten light? i.e. the eyes are in shadow?

IMG_4365.JPG
 
Yes, it should nail a shot like that easily enough. But I'm not sure it hasn't.

His top is sharper both in front and behind the eye line, so there must be subject movement - quite likely 1/15sec.

Why don't you just give Sigma a call and see what they can do?
 
Micro adjust aside I wouldn't stake my life on a camera's ability to auto focus constantly. Canon say on their web site that if you take three shots you'll see focus differences between them and that this is normal, so I guess mileage may vary...

If you push things to the very edge of what's possible you have to accept a level of failure and I think that f2.8 at 1/15 sec qualifies as pushing your luck a bit. If I think that I'm pushing my luck I try and maximise the chances of success and in this case I think I'd have selected a higher ISO to achieve a faster shutter speed and taken more than one shot and accepted that I might have to bin some and I might even have switched to manual focus.

I don't know if that white balance is right, I shoot RAW, sometimes with custom WB but always RAW.
 
Yeah, he was really still to be fair to him. My boy right there! :) They are very patient for Dad :)
I will shoot some more with it and see what the results are, that was a harsh test above.
Yeah I have called Sigma before and it would cost money as this was a second hand purchase.
 
You must log in or register to reply here.
Back
Top