AF Microadjust

CFC1

Distinguished Member
Joined
Mar 16, 2006
Messages
15,911
Reaction score
4,457
Points
3,308
Location
Europe
I had problems a couple of years ago with AF on my nifty fifty on my Canon 7d. It was hardly ever sharp, with the focus point fluctuating between back and front focussing. I did the usual chart tests and microadjusted accordingly, but I could never get any consistency either way. I noticed that if I acquired focus and then pressed the shutter again repeatedly, the lens would judder like it was trying to find focus again. This would happen 19 times out of 20. I would assume that if it was locked on correctly, then it wouldn't judder (try to refocus) again!. On the odd time that it didn't judder, if I then took the shot, it would be perfectly sharp. I gave the lens up as a bad job and a couple of weeks ago replaced it with the 50mm STM version.

I took a couple of shots for the first time last week, but again, wasn't that happy with the sharpness. So, I did a quick AF test on a chart this evening and am still having problems. This new lens is still doing the judder thing again when I press the shutter after focus has been confirmed, although not nearly as often as the old lens ... maybe 8 times in every 20. But still, the AF chart shows a variation in focus. It looks like it's back focussing in the main, with the odd shot about right.

The test was wide open at f1.8 on a tripod. I know that there is this rule that says that 1/3 in front and 2/3 behind the focus point should be in focus. Is this true in a test at close quarters like this?

Finally, is the "juddering" that I mention above when the shutter is pressed after focus has been achieved normal? Unfortunately I haven't got another body to test it on. I'm beginning to wonder if the problem is actually with the 7d body as opposed to the lenses.

Here are 4 consecutive test shots, each refocused after defocussing with my hand at a closer distance. Number 2 looks about right to me, but the other 3 seem backfocussed. Are these within tolerences or not?

Any advice/info would be greatly appreciated.

Cheers, Calvin

1.

2.

3.

4.
 
I don't shoot Canon so can't comment on the behaviour of their lenses, however with some lenses on some brands lenses do behave like this in terms of the judder when refocussing even if it's on the same point. The 35mm f1.4 on the Fuji is the worst I've come across with this, but some Olly lenses do it to.

As for AF testing on charts like this you will find that it will vary from shot to shot, even in extreme controlled situations. Unfortunately it's the nature of phase detect and one of my biggest bug bears. Whilst you're right in that generally you should have 1/3 in front and 2/3 behind the focus point if you use live view in these test scenarios (and therefore dead accurate AF) you will find that DOF is 50/50. Since discovering this I have adjusted my phase detect AF to match this the best I can and have had the best results this way.

The other frustrating thing about phase detect is accuracy will also vary with subject distance and so it's impossible to get it set right for all scenarios. This is why the Sigma and new Tamron docks are so good, you can adjust AF fine tune at multiple settings at different focal lengths and subject distances. Nikon and Canon are really slacking in this dept, although Canon are marginally better in that you can set AF fine tune for both ends of a zoom lens, rather than just the one setting of Nikon.

From you images above I would say there is a touch of back focus, but still acceptable as the focus point is always sharp.
 
Cheers Snerks. Appreciate you commenting so thoroughly mate. I'll microadjust the focus forward a point or two and accept the limitations you outline above.
 
Not a Canon user, but I don't recall this being an issue - and yr Tests look pretty fair to me.

In real photography the subject is all over the frame at varying distances, so whether it's 50-50 or Other, probably is down to User Preference.

Yet, I'd agree that it would be a worry if the lens is hunting, but I wonder if that isn't the fault of the Test-Chart, since as soon as the focus changes, it finds another line that's similarly sharp - so it can't decide which is the one you want....
Perhaps Others here have the full story - It will be interesting to know...

Good Luck.
 
but I wonder if that isn't the fault of the Test-Chart, since as soon as the focus changes, it finds another line that's similarly sharp - so it can't decide which is the one you want....
.

I don't think so Harry. I had spot focus on that thick "FOCUS HERE" centre line. It never once focussed anywhere other than that.

Appreciate you replying, cheers.
 
I'm a canon 7D user and have the 50mm II but not the new STM model.

The original nifty fifty had a micro motor autofocus system and it was a hit or miss. I carry out micro focus adjustments on all my lenses but the nifty fifty jumped about so much from one attempt to the other that it seemed pointless.

For photography you are better with the ring USM autofocus systems. The STM was really introduced for step-less video focus.
 
My experiences exactly with the nifty fifty. The 1.4 is out because it's softer at 1.8 than the nifty fifty ... and the 1.2 is out because it is so pricey. To be honest, I'm thinking about moving to Nikon now as Canon are so slow to update their bodies in particular, and when they do, they are still not competitive specs wise as Nikon. I was looking at possible upgrades for my 7D and was maybe thinking FF with the 6d, but it's 4 years old now and it won't do my job for sports. The AF system with its 11 points (and only 1 cross type) is piss poor, TBH. The Nikon d750 is streets ahead. With Canons track record, I haven't got any confidence that the 6dmk2 will be an advance on the Nikon. Then, my other option is to update my 7d to the mk2. But that's been out for 2 years now and whilst class leading then, it is blown away by the Nikon d500 now, which is what I'm leaning towards at the minute. It's just that I have 5 Canon lenses, 2 flashes, triggers and other stuff, so the change is gonna lose me money....

Appreciate you commenting waysted.
 
My experiences exactly with the nifty fifty. The 1.4 is out because it's softer at 1.8 than the nifty fifty ... and the 1.2 is out because it is so pricey. To be honest, I'm thinking about moving to Nikon now as Canon are so slow to update their bodies in particular, and when they do, they are still not competitive specs wise as Nikon. I was looking at possible upgrades for my 7D and was maybe thinking FF with the 6d, but it's 4 years old now and it won't do my job for sports. The AF system with its 11 points (and only 1 cross type) is piss poor, TBH. The Nikon d750 is streets ahead. With Canons track record, I haven't got any confidence that the 6dmk2 will be an advance on the Nikon. Then, my other option is to update my 7d to the mk2. But that's been out for 2 years now and whilst class leading then, it is blown away by the Nikon d500 now, which is what I'm leaning towards at the minute. It's just that I have 5 Canon lenses, 2 flashes, triggers and other stuff, so the change is gonna lose me money....

Appreciate you commenting waysted.
Calvin as you may have seen from the Canon thread I almost swapped to Canon sue to issues with my D750 and Nikon's service dept and when I looked into the cost to change systems it's not as bad as folk make it out to be. If I sold my lenses and bought the equivalent used Canon ones there was very little in it. The only major difference was the cost of the body as Canon tends to overprice them imo. This was pre Brexit result and the weakness of the pound though, but I'd have thought the relative market would still be the same.
 
Yeah, I followed your woes over your d750.... Yeah, I'll do a serious costing in the next couple of days and see what I'm looking at to change. I've indeed noticed the d750 price snuck up by a couple hundred quid in the last couple of months!
 
Yeah, I followed your woes over your d750.... Yeah, I'll do a serious costing in the next couple of days and see what I'm looking at to change. I've indeed noticed the d750 price snuck up by a couple hundred quid in the last couple of months!
Yeah it has, but then so have used prices of your gear so it's swings and roundabouts.
 
Cheers Toby. I read somewhere that the Nikon 70-200 is actually only about 160mm at the long end! Is that true?
 
Cheers Toby. I read somewhere that the Nikon 70-200 is actually only about 160mm at the long end! Is that true?
Less than that, 135mm. It's not something I've ever noticed, but then I've not really tried to compare, plus it's only at close focus distance. Not great though when you pay that kind of money, although it actually works out an advantage for some of my shooting. That's only the VRII though, I don't think the VR1 suffered anywhere near as bad, and they've just released (or shortly releasing) a new one that is supposed to be the bees knees (although I'm not sure how they can make it optically better than the VRII), is a bit lighter due to their fresnel glass, and also has no focus breathing. Trouble is it's silly money :(
 
My experiences exactly with the nifty fifty. The 1.4 is out because it's softer at 1.8 than the nifty fifty ... and the 1.2 is out because it is so pricey.
From the charts on DPReview there doesn't seem to be much between the 1.4 and the 1.8 with regards to sharpness once you step down to 2.0 (center and edge better on 1.4 and 1/2 to 2/3 range better on the 1.8).

"If" the 1.4 focus is more accurate then it would certainly be better than a 1.8 that is out of focus.

I don't have the 1.4 so can't comment on real world vs charts. I do have the 28mm 1.8 and lots of people claim it's soft but mine is only soft at 1.8 and sharp by 2.0. The softness at 1.8 is actually really handy!

Agree on your thoughts about Nikon being ahead but I'm in the same boat as I have a 6D, 12 lenses and 3 flash guns. Personally feel that although Nikon are currently ahead that I can still take great photos with the Canon gear so no point in changing. If I was starting from scratch then it would be Nikon.
 
This nifty 50 is still doing my head in. It's all over the place. I'm microadjusting anywhere between -7 to +6 depending on how it's feeling at the time. I've just done some flat wall tests with a defocus in between each shot. Wide open at f1.8. These are just bigger than 1 to 1... (focus is on the "1.8")

1.

2.

3.

4.

5.

6.


:mad:
 
This nifty 50 is still doing my head in. It's all over the place. I'm microadjusting anywhere between -7 to +6 depending on how it's feeling at the time. I've just done some flat wall tests with a defocus in between each shot. Wide open at f1.8. These are just bigger than 1 to 1... (focus is on the "1.8")

1.

2.

3.

4.

5.

6.


:mad:
:(
 
Unfortunately not. My only other camera is a Fuji x100s ... oh, and a TZ10 that I use to win the weekly TZ compo... :devil:

I'll see if I can borrow one...
 
Unfortunately not. My only other camera is a Fuji x100s ... oh, and a TZ10 that I use to win the weekly TZ compo... :devil:

I'll see if I can borrow one...
Deffo worth trying on another camera, don't you have a local store? Perhaps your 7d's just getting on a bit, after all 7d's are almost as old as Jim ;)
 
I doubt that it's the camera but you never know.

You may be expecting way to much precision from the cheap 1.8 50mm lens. You have the STM version but I'm not sure if it more precise than the old version. Looking at some dpreview forums posts it seems to be more accurate.

Micro focus adjustment allows you to move the "average" result in one direction or another but it does not improve the precision of the autofocus i.e. every camera/lens combination will have an inherent precision and the micro focus adjustment has no impact on this.

So if you take 6 photos at one setting, eg -10, another 6 at 0 and 6 more at -10. Compare each set and see which set gives the best results (ie closer to correct focus). If it is 0 then repeat at -5, 0, 5.

This article gives an over view of micro focus: AF Microadjustment Tips

I never use charts to do the focus adjustment. I set-up 3 items (cans, matchboxes etc) in a diagonal, stand at the distance I normally plan to shoot at and focus on the center item. Then I compare that with the item in front and behind.
 
I doubt that it's the camera but you never know.

You may be expecting way to much precision from the cheap 1.8 50mm lens. You have the STM version but I'm not sure if it more precise than the old version. Looking at some dpreview forums posts it seems to be more accurate.

Micro focus adjustment allows you to move the "average" result in one direction or another but it does not improve the precision of the autofocus i.e. every camera/lens combination will have an inherent precision and the micro focus adjustment has no impact on this.

So if you take 6 photos at one setting, eg -10, another 6 at 0 and 6 more at -10. Compare each set and see which set gives the best results (ie closer to correct focus). If it is 0 then repeat at -5, 0, 5.

This article gives an over view of micro focus: AF Microadjustment Tips

I never use charts to do the focus adjustment. I set-up 3 items (cans, matchboxes etc) in a diagonal, stand at the distance I normally plan to shoot at and focus on the center item. Then I compare that with the item in front and behind.
Whilst there are tolerances in AF they should only be to within 1/3 of the zone of focus, and I'd not be happy with Calvin's shots tbh. But yes you are right in that changing AF fine tune just moves the average and will not help with precision. Your method of fine tuning sounds a bit long winded to me, but if it works for you :smashin:
 
Well, I borrowed a mates 6d today and did the same wall chart test as yesterday with a few lenses, inc the 1.8stm. There were very little variations between each set of 6 consecutive defocussed and then focussed shots. When doing microadjusts, if focus was out, then it was out consistently by approx. the same amount in each picture in the set... certainly not the wild fluctuations of yesterday. :mad:
 
I doubt that it's the camera but you never know.

You may be expecting way to much precision from the cheap 1.8 50mm lens. You have the STM version but I'm not sure if it more precise than the old version. Looking at some dpreview forums posts it seems to be more accurate.

Micro focus adjustment allows you to move the "average" result in one direction or another but it does not improve the precision of the autofocus i.e. every camera/lens combination will have an inherent precision and the micro focus adjustment has no impact on this.

So if you take 6 photos at one setting, eg -10, another 6 at 0 and 6 more at -10. Compare each set and see which set gives the best results (ie closer to correct focus). If it is 0 then repeat at -5, 0, 5.

This article gives an over view of micro focus: AF Microadjustment Tips

I never use charts to do the focus adjustment. I set-up 3 items (cans, matchboxes etc) in a diagonal, stand at the distance I normally plan to shoot at and focus on the center item. Then I compare that with the item in front and behind.

Cheers.:smashin: AFAIK, it's not a simple microfocus adjustment issue, but more a random variance when autofocussing. I can't pin it down to a front or back focussing issue. It might be front in one shot, back in the next and bang on the next ... and so on. And, it's not just with the nifty fifty. It's doing the same with my Canon 100mm macro and 70-200mm too. :mad:

As an example, if I am sitting 5 feet away from my wife and focus on her nearest eye with the nifty 50, maybe 1 in 6 pictures will be sharp on the eye... the rest will be either front or back focussed... to varying degrees. :(
 
Last edited:
Here are some results from todays testing to give you an idea how consistent the results were. The crops are pretty big and taken from the same distance as yesterday, but they were taken on the FF 6d today, as opposed to the crop7d yesterday. Once again, the focus was on the "1.8".

Original size...


1.

2.

3.

4.

5.

6.
 

The latest video from AVForums

TV Buying Guide - Which TV Is Best For You?
Subscribe to our YouTube channel
Back
Top Bottom