Check your microadjust - It makes a BIG difference

[etherial]

Recently I purchased a brand new 70-200F2.8L IS lens to go with my 7D. I noticed while using it during my recent trip away that I wasn't hitting the focus anywhere near what I would expect. I was really disappointed with the sharpness of most of the photos. I tried all sorts of things with the camera focus setup all yielding similar results.

So after much reading about this, I went to the micro-adjust feature on the 7D. I was a little sceptical as to how much difference with would make and being a brand new lens (and rather expensive at that!) one could surely think that it should be pretty close. In the field I settled on a +15 setting which I was rather disturbed by.

Today I have finally had a chance to do some controlled tests. I set up the camera on a tripod square to target, IS off, live view for framing and using quick AF (so it focuses through the normal AF system). AF set to centre spot with the camera levelled using the inbuilt level. Metering was as constant as I could with apperture wide open (minimum depth of field). I used the inbuilt timer (with mirror lockup) so camera movement was non existent.

The camera was set up at >10m distance with focal length of 200mm.I then went through a sequence of shots starting with the micro-adjust at -20. I would manually turn the focus ring to infinity, then make the camera auto-focus. I took three shots on each setting until I had covered -20 -10 0 +10 and +20 on the micro-adjust.

I then did the same process with the camera at 100mm focal length and just over 5m distance to subject. I also did the same with my 24-105F4L lens at 30mm and 100mm.

After transferring to my computer I selected the sharpest image of each set up 3. I have then zoomed to 100%, cropped, labelled and assembled for you below.

As for the result, well lets say that the default setting of zero is bad and it is no wonder my photos were not coming out sharp. None of them are great especially when I compare to the 24-105mmF4L (I'll post those results here too when I get a chance to collate them but the shot at 0 is sharper than all these). Obviously the sweet spot might be between two of the points I have tested at. I will need to repeat the test at closer intervals in the range which is close to narrow down the sweet spot.

In summary as you can see, the micro-adjust makes a massive difference, especially on this lens. I'm keen to hear peoples feedback about this, I posted it here for two reasons, to show the difference micro-adjust can make, and to get advice on what I should do with this lens! I am really disappointed in it overall; the F4 version I had was MUCH sharper (hope you're enjoying that Alan!). I'm 99% sure I will send it in for service (I guess I will contact store I bought it from and see what they tell me). I'm concerned also that a +10 setting which looks best at 200mm is still poor at 100mm, so will it ever be right!?!!

[arthurking83]

.... I'm concerned also that a +10 setting which looks best at 200mm is still poor at 100mm, so will it ever be right!?!!

almost certainly not!

That's always a problem with zooms, and all my zoom behave slightly differently.
You kind'a expect that though as each lens changes it's optical formula(the way in which each lens element/group is placed relative to another) for different focal lengths.

Both my real problem children, the now gone 80-200/2.8 and the 28-75/2.8 were inaccurately focusing at the longer end, but perfectly focused at the shorter end. So in the ned :p.. I left them as is and tried to compensate with the way I shot them, either with smaller apertures or slight body movement back from the subject as I shot.

BUT.. you said you used liveview to frame and the quick AF system to focus.

two things:

Did you defocus for every attempt to focus? You should defocus the lens both backwards and forwards for each attempt at focusing. That way it eliminates any slack in the focusing system or whatever else may be causing the focusing error.
What i did was 10 focus runs, where I totally defocused the lens backwards and then hit the AF button, then next time I'd defocus the lens forwards and retried focusing.
In my case both lenses are/were screw driven so there could be some slack in the screw drive-lens mechanics that could be an issue.

Also, are you sure that the AF performance is exactly the same as it is in non live view mode? Do you get perfect focus using the slower contrast detect focusing system(where the sensor actually does the focusing, not the AF system.



if the problem exists without any errors in testing as you've tested it, then the lens is probably best taken in for a service calibration with the problem clearly explained.
Looks to me as tho you've done ok tho Mic and it's not rocket science now to check lens calibration...

long live live view! :th3:

[etherial]

Hi Arthur thanks for the reply. In answer to your questions, yes I did defocus for every test. For the three shots on each setting I defocused the first in infinity, the second to minimum and the third to infinity. There was no pattern of whether the focus was better when coming from minimum or infinity.

On the second point, that is something I didn't try! Completely forgot. I do need to do it again to try and get closer to the exact setting, so I'll try that then.

I'm not feeling confident about this lens though, I am so disappointed in its results so far. :(

[arthurking83]

.....

I'm not feeling confident about this lens though, I am so disappointed in its results so far. :(

Yeah! it happens.

funny thing is.. it may only be confined to your 7D. if you still have any of your original cameras maybe you could try it with them.. or someone else's camera.

but the best thing to do is to get to the Canon service folks and get them to check/calibrate it.

[etherial]

I'm confident it isn't the camera as my 24-105mm F4L seems ok, and before I bought the 70-200 F2.8L IS I had the F4 version which I found to be brilliant on both my old 450D and the 7D.

I sold that here on AP to help fund my "upgrade" to the much hyped about F2.8L IS version.

[peterking]

Thanks for this guys. It makes me think I need to do some testing myself on my two remaining lenses.

[joele]

Another problem with adjusting under a set light condition (say inside under tungten) is AF systems can behave differently under different light sources..

Pentax k100d/k10d/k20d has that problem, a lot of people found they front focussed under tungsten lighting but then fine under sunlight or other types of lighting. So they adjusted under tungsten adn it was back focused under different lighting.

[rwg717]

Thanks for that post....might be having the same problem here too?:(
Richard

[etherial]

Well after being in contact with the retailer and sending them the images attached they have spoken to their Canon rep and agreed there is a problem. They have asked me to send the lens back to them and it will be handed over to their Canon rep.

I'll report back here how it goes. *insert fingers crossed smiley here*

[Ploddy]

Nice one etherial, a good solution.

Just so you know, I have downloaded the image tests that I did on the lens for you but havent posted yet - I should be able to do so over the weekend. But irrespective of the result its great news Canon are honouring a service for you.

[darylcheshire]

I suppose Canon introduced the micro-adjust to reduce the amount of service calls for lens adjustments.

I'm no engineer but I wonder at what's acceptable for the millions of lenses and cameras which canon sell that have inaccuracies. It is known as instrument error and using two identical measuring devices never produce identical results, and this is factored in calculations and is called the uncertainty value.

Still, it's good that the lenses can be re-calibrated. Ideally, the lenses should be calibrated when they leave the factory as part of the QC. (at least the L lenses). Still, temperature and transit issues take their toll and perhaps a lens in North America would act slightly differently to a lens in Queensland. However I'd expect the lens quality to minimise this issue.

Daryl.

[etherial]

Well I got my lens back from Canon today. The service report gave me the standard "cleaned and adjust focus", no surprises there. Who knows what they really found.

Off course the first thing I did was have a play and see if things had improved. I did a few focus tests tonight using the same method as before. The good news is it is more consistant than it was. I also went further and did some more tests using a slightly different method where you use a focus test chart at 45 degrees to the camera and it tells you if you are front or back focusing.

After about half and hour I am satisfied that it is now ok at a setting of +10. This does sound high and maybe I should get my camera checked out, but at least it was consistant through the focal length range and at different distances.

I also took some time to do my 24-105 in more detail and ended up setting that at +3.

So I'm happy now I can put that to bed for a while and get on with shooting!

[Ploddy]

Great news mate, I know it was bugging you - if it was my lens I would be the same. You pay a lot of money for these things and you expect them to be right.

I just have to get off my ass and calibrate mine now...

[etherial]

I should post the other method I just changed to, I think it is easier as it gives you a direction to adjust as you go. I did it slightly differently that in this link, I set up the paper at 45deg angle and put the tripod and camera straight. That way I could test at close and distant ranges (I went from as close as it would focus to a few meters).

http://focustestchart.com/chart.html

The chart and method is down the bottom.

[Ploddy]

Great link there, but man, I cracked up at the 45 degree debate... don't cross that guy...

[etherial]

Just a final comment to close off this thread. Last weekend I used the lens out in the field shooting at an event for the first time since I got it back from Canon. Very happy with it now, all photos extremely sharp and various focal lengths. It is now as good (if not better) than the F4 version I had to begin with.

So the lesson learnt - it does pay to check you focus with some thorough tests. If you find it to be dodgy send your lens in for calibration. And when it comes back check it again to ensure it matches up with your camera and all is good. The results are worth it!

[dche5390]

Dont. Have. Micro-ajdust. Nooooo! 70-200L seems to be focusing just fine with my 5D. Not too sure about the 50L however ...

[AdamJL]

AF Microadjust is a great feature, but blimming hard to do with a 500mm lens. You can't use the 45 degree test, and instead need to stand about 20 meters away from a printed grid stuck on a wall.

[TAYLORW619]

Hi Etherial,

I also have the 70-200 f/2.8 MK1 and I am interested in doing this 45 degree test. I understand everything that has been said for directions except consistency through all zoom ranges. How do I test consistency for all zoom ranges?

I am interested to see the results!

Thanks,
Taylor

[etherial]

AF Microadjust is a great feature, but blimming hard to do with a 500mm lens. You can't use the 45 degree test, and instead need to stand about 20 meters away from a printed grid stuck on a wall.

You can do the 45 degree test the same way I did it. Set the test paper up at 45 degrees and then have the camera set up accross the room pointing at it. Just put the test page on something rigid and stand it at 45 degrees on something like a chair. Set you camera up on a tripod, and away you go.

[etherial]

Hi Etherial,

I also have the 70-200 f/2.8 MK1 and I am interested in doing this 45 degree test. I understand everything that has been said for directions except consistency through all zoom ranges. How do I test consistency for all zoom ranges?

I am interested to see the results!

Thanks,
Taylor

Hi Taylor, you can just do the tests at a few points through the range. I did three points, 70mm, 130mm and 200mm. I found that after my lens came back from canon I got the same result at each focal length. If they are slightly different you can bias your adjustment based on what focal length you use most. Or send you lens away for calibration.

[joele]

I don't know how good the 45 degree test is, as the af point isn't exactly a tiny centred point within the small box within the viewfinder, shooting at a flat target for adjustments is better... try the method below.

http://www.birdphotographers.net/forums/showthread.php?58042-AF-microadjustment-tricks

[AdamJL]

You can do the 45 degree test the same way I did it. Set the test paper up at 45 degrees and then have the camera set up accross the room pointing at it. Just put the test page on something rigid and stand it at 45 degrees on something like a chair. Set you camera up on a tripod, and away you go.

Hi etherial. I hope you realise just how big that slider will have to be!

The centre focus point when you stand far away (and you really should be doing all Micro Adjusts at the distance you'd normally use the lens) with a 500mm lens would nearly cover half an A4 page, possibly more.

[etherial]

I don't know how good the 45 degree test is, as the af point isn't exactly a tiny centred point within the small box within the viewfinder, shooting at a flat target for adjustments is better... try the method below.

http://www.birdphotographers.net/forums/showthread.php?58042-AF-microadjustment-tricks

Yes I read that thread a few times and used that method the first time. I found though that it doesn't give you a reference on which way to adjust. For me the 45 degree method was easier as it gives you instant feedback of how far out and what direction to adjust.

There are a heap of methods out there, it's a matter of finding one that works for you.

Mic

[etherial]

Hi etherial. I hope you realise just how big that slider will have to be!

The centre focus point when you stand far away (and you really should be doing all Micro Adjusts at the distance you'd normally use the lens) with a 500mm lens would nearly cover half an A4 page, possibly more.

Hi Adam, I'm not sure what you mean by "slider". Anyway Canon recommend 50x the focal length as the test distance, so for your 500mm that would be 25m.

[etherial]

For the techo nuts out there...here is another method. I tried this too early on but found I couldn't get repeatable results. As I said before, find a method that suits you.

http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html

[TassieSnapper]

Im interested in reading your test document, the thing about pet ownership and the like. Can you please post a shot of the whole page?

[etherial]

Im interested in reading your test document, the thing about pet ownership and the like. Can you please post a shot of the whole page?

Serious?

[TassieSnapper]

Serious?

:D

[AdamJL]

Hi Adam, I'm not sure what you mean by "slider". Anyway Canon recommend 50x the focal length as the test distance, so for your 500mm that would be 25m.

Hi etherial

Apologies, by slider I meant the old 45-degree ruler thing that's commonly used. But in any case, whatever is the target, believe me, it's completely impractical to MA a 500mm lens this way. The target has to be very big so the focus doesn't cover it (and at 50m, it has to be huge).

[arthurking83]

I agree with Adam, and recently basically explained the same thing to another member.

The problem with these 45° A4 paper tests is that you are really really really close to the subject, which is not always the case in your real photos.
Getting too close can magnify a problem to the point where if you adjust for back-front focusing issues up close it may work against you at normal operating distances due to the naturally deeper DOF(at the same lens settings).

lenses(like the Nikon 80-200/2.8 in my case) CAN have different focusing characteristics at different focus distances.
Note! I did say CAN. My Tammy 28-75/2.8 is consistent in it's backfocus issue at all focus distances, where the Nikon was OK close up, but backfocused at standard portrait distances(say 5-10meters away!.. damned annoying to say the least).

I don't check using the 45° test.
The battery test is better, or the wall mounted test chart at a reasonable (if you have liveview) as another alternative.

[etherial]

Just to be clear, I don't advocate doing it the way it is pictured in the other thread as it seems way to close. I still kept to Canon's recommended minimum of 50x the focal length so when doing the 200mm I was going through two rooms of the house :D. (My wife thinks I'm nuts by the way!)