Micro-Adjustment charts... Bunkum :)

JohnN

JohnN

Messages
6,359
Name
John
Edit My Images
Yes
Hi all,

Well I thought I would try and sort out my lenses as I felt I was getting soft images with kit that others were getting good stuff from, so off I trawled to good old Google and found these little gems:

1) http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html
2) http://regex.info/blog/photo-tech/focus-chart

Starting with the first one I spent over an hour trying to get the damned images to produce a pattern, going from -20 to +20 and not a peep (starting with little changes which is why it took so flipping long!), very frustrating - guess I just have weird monitors, so I gave that up...

Off to site 2, printed the chart, set it up on a nice low angle, this time results, aparently my test lens needed a -20 adjustment!!! To say I was shocked was an understatment, so...

Off to the garden, took some images and they were worse than before!

Sod this I thought, so I just found an old crusty fat ball that was still hanging off a post and took some snaps at -10, 0 and +10 and it was closest at 0, so -5, 0 & +5, better at +5 than -5 so, I just took images at +5, +4, +3, +2, +1 and 0 for refernce, lo and behold most were pretty good, but +2 stood out as the best.

A few shots later including a blackbird with worms in its mouth that was sharp and thats me a happy bunny.

I then went to work on all my lenses and have done the lot and guess what, yup all are much better now :)

-20 indeed :bonk:

No doubt I was doing something wrong with the charts, but still - I spend all morning and half the afternoon on charts and on screen images and an hour on my own playing about gets four lenses much better if not perfect :)

Anyway just thought I'd share.
 
Last edited:
JohnN said:
lol, could be right :D
Click to expand...


Im not a big fan of the charts to be honest. And I have only had one lens that I really thought needed adjusting for anyway. I did it pretty well like you describe in the end. I think seeing where the lens is focusing can be useful if there is a serious problem but maybe I'm just lucky and haven't had a lens that was significantly out of tolerance.

I do think that focussing problems are more often than not user error and if there is a serious issue the lens would probably need to go in for repair anyway.
 
That was my thinking which is why I didn't do it before, but after sorting the MA I got a bee in flight, the face of a butterfly, could see the eye ring of a blackbird on the roof, clearly see eyelashes on a piccy I took of my daughter.

It was a massive boon for me :)
 
I tried the first one on your list and must say, I found focusing on the text in the window that displays the image much more reliable than the image itself. The image suggested -13, yet the focus was spot on when looking at the computer screen text.
 
See this thread, post #40 http://www.talkphotography.co.uk/forums/showthread.php?t=310467&page=2

I'm not a fan of those test charts either, but a little bit of kitchen science, when applied appropriately in a real world method, works a treat :thumbs: Try the cerial box and ruler shown above - easy, dead accurate, takes two minutes once you get the hang of it.
 
HoppyUK said:
See this thread, post #40 http://www.talkphotography.co.uk/forums/showthread.php?t=310467&page=2

I'm not a fan of those test charts either, but a little bit of kitchen science, when applied appropriately in a real world method, works a treat :thumbs: Try the cerial box and ruler shown above - easy, dead accurate, takes two minutes once you get the hang of it.
Click to expand...



When I have messed with fine tuning I have used the cereal box type setup to see what's going on but used the method where you take a series of shots of a target at different fine tune adjustments and then picked the sharpest one to decide on the setting. Did the cereal box type test again just to confirm.

I have had a few ideas of how to make a simple test rig that is very accurate, will try them out and post results if I get around to trying them.
 
lensflair said:
When I have messed with fine tuning I have used the cereal box type setup to see what's going on but used the method where you take a series of shots of a target at different fine tune adjustments and then picked the sharpest one to decide on the setting. Did the cereal box type test again just to confirm.

I have had a few ideas of how to make a simple test rig that is very accurate, will try them out and post results if I get around to trying them.
Click to expand...

I think have tried every kind of target known to man (or been posted on the web at least) and gone half blind in the process. The cerial box and ruler beats them all, and if you set it up carefully is absolutely accurate - ie perfectly flat box, dead square, which is very easy to do at sensible distances which are less critical of perfect tripod positioning.

To get it mm accurate though, you need a very fast lens or depth of field conceals detail, in practise f/2 or lower. With higher f/numbers the temptation is to move closer to reduce DoF, but that's when you start to provoke or exagerate errors that are not present at normal shooting distances and if you correct for them you end up making real subjects worse.

Or to look at that another way, with most popular lenses at typical shooting range, the AF system is unlikely to be the weakest link in the chain. Certainly, I have discovered that the camera's AF is far better at focusing that I ever will be and in critical working, subject or camera distance shifts coupled with my own inability to track the focus 100% of the time, are more likely culprits. It's easy to get paranoid, no camera/lens is perfect at all times.

You can either read off the error from the rule and adjust accordingly (if you have a good LCD, you can read it straight off that) or shoot a series at different microadjustment settings and see which one is best. Works either way. It's essentially the same as the commercially available targets, some of which cost a couple of hundred quid, and is effectively identical to the LensCal http://spyder.datacolor.com/product-cb-spyderlenscal.php which is quite a nice toy for $60, but completely unnecessary.
 
System I used was simply to hook the camera up to the PC, fire up the EOS Utility software and put it in live view at 100% or 200%.

Find a decent target with high contrast, then let the camera AF in the normal way (Quick AF in live view mode), then you just use the controls in the EOS Utility to nudge the focus in or out in single steps to see if the picture gets better or worse. Do it a few times to ensure the results are repeatable. However many steps you need, relative to the Quick AF lock, to make the picture look best is how many steps of micro-adjustment you need to dial in.

The advantage of this method is you're comparing relatives rather than absolutes. All you have to determine is whether the image gets better or worse each time you nudge the focus.
 
Last edited:
Cheers, I'll give that a go (shame the images are no longer available) and see how it works out as these things seem to work out for others.

However if I need 25-50x focal length does that mean that if I were to test the 100-400mm at 400mm I would need to have the chart 20m away, in which case where on earth would you put it??
 
You must log in or register to reply here.
Back
Top