AF Micro-Adjustment tool

[Bart_van_der_Wolf]

I see what you are trying to say , but I must say your English is very technical , dare I ask what your profession is ? ^^

I'm a photographer, amongst others, with a technical inclination. As Ces said, the subject is of a technical nature, so I try to be exact in my explanations.

Let me give it a try and I'll be back if more things need explaining, thanks for the kind feedback

I'm sure you'll get it solved eventually, it's a great lens so it deserves to be optimized.

Bart

 

[Nicolas ROLLAND]

Hello,

As this is my first post here, first I would like to thank you all for this really interesting (geeky ?) thread. I am a French 1D Mark III owner (got it one week ago) and naturally as I have started to digg into the camera features, I became interested in micro-adjustments.

So far I have tried inanimate objects targets, paper focus targets (flat and 45°) and now Bart's LCD moiré target. But no luck. Whatever I do, I get really inconsistent results. I have tried to micro-adjust (@ full zoom, wide open) my Canon 17-40L, Canon 100 macro 2,8 and Sigma EX DG 70-200 2,8 and while I think I get a fair enough result at say +5 if I just try again and shoot some more tests say -10 I see no real difference with the previous +5 tests. I am probably missing something but I really don't get it.
So far the LCD moiré method seduces me most (probably because of the geek in me) but I get no better results.
Here is how I shoot : Mark III on tripod (old but fair enough Gitzo Gilux with Manfrotto ball head), mirror blocked, 2s timer shot, One shot drive, C.FnIII-8-0 and target displayed on my iMac screen (I tried with a secondary display and laptop screen as well).

So any clue or help on my micro-adjustment quest would be greatly appreciated

Thanks,

Nicolas

P.S.: please excuse my French (or at least typos and grammar mistakes from a French English language lover but forever learner)

 

[nicolas claris]

P.S.: please excuse my French (or at least typos and grammar mistakes from a French English language lover but forever learner)

LoL! tu veux dire "please excuse my English" I suppose?

Don't worry, sicnce the beginning I do my best to teach all these English mother tongue guys to understand me, and that's a challenge! LoL!

BTW, welcome on board OPF, we're not so many French around here, but quite a lot of very active European guys (though some are still in vacation).

Back to the topic, I'm sure that Bart (a Duch fellow BTW) will answer you much more accurately. But your settings seems strange to me.
To set easily my 1DS3, I used a tripod, 100 ISO, ƒ2.8 (max anyway) longer lens if a zoom lens, and afair a speed around 1/60s

With these settings I could have consistent results…

 

[Nicolas ROLLAND]

LoL! tu veux dire "please excuse my English" I suppose?

Well, no actually that was just a bad humour attempt ^^

Glad to see French people here !

 

[Nicolas ROLLAND]

No answer so may be this was really bad humour ?
Seriously, I have tried again and again I have bad results. I just feel dumb now.

 

[Cem_Usakligil]

No answer so may be this was really bad humour ?
Seriously, I have tried again and again I have bad results. I just feel dumb now.

Bonjour Nicolas,

Please don't feel like that, there really is no need for it. I am certain that Bart has been occupied with the thing we call "life" and was unable to reply. He'll probably come back to you soon.

Your humour was spot on, BTW.

Cheers,

 

[Nicolas ROLLAND]

Thanks for your kind words Cem,

I don't feel bad, just dumb as I said.
You are right I may be somehow impatient.

 

[nicolas claris]

Hi Nicolas
As I have no idea what info is missing to you, I tried to summarize Bart's suggestions
I have kept the # of each post so you may go back there to read the context…

Hope it helps (and maybe some oher OPFers too!)

Bart's posts below:

#3

What happens as one manually adjusts the focus is that at the exact optimal focus setting the background will change from uniform gray into larger (colored) aliased dots when viewed on the camera's LCD. The circles and cross hair will allow to acquire AF easily, and when calibration is optimal, moiré will be maximized on the camera's LCD.

A procedure that works for me:
- I switch to LifeView, which in its current implementation will only allow manual focus.
- Optimize manual focus by searching for maximum aliasing. This will only occur a best focus (and assuming a decent enough lens is used), otherwise the defocus will act as a low-pass filter and prevent the aliasing.
- Switch off LifeView, and watch the lens barrel's focus indicator for the next step.
- Use AutoFocus (single AF spot) to focus on the (laptop) LCD screen, and watch the direction of adjustment. That will show whether the current AF calibration setting will front or back-focus.
- Apply an adjustment via the camera menu, and repeat the procedure. Once the adjustment is optimal, there will be no difference between manual and auto-focus.

#7

Canon suggests to use the longest focal length of the zoom range

#11

LifeView should be used with aperture wide open (most shallow DOF).

#16

Reasons for not achieving the moiré can be;
- too far out of focus, but that's also what the crosshairs are for, a visual clue and as well as an AF sensor aid
- distance too close, or magnification to large (which is the same), which prevents small enough detail to cause moiré
- distance much too far, which will cause the moiré to have too low an amplitude to see
- the lens has too low resolution
- or the manual focus is too reactive to nail the exact spot, try slower manual focus adjustments

#43

Assuming you aim at the target from a tripod (otherwise you'll probably see camera movement aliasing patterns), it's best to disable IS (since not all versions switch off when tripod use is sensed). When attempting calibration with a handheld setup, IS will help to stabilize the movement artifacts.

Any front mounted filter will reduce the overall resolution, while insertion type of filters (on longer tele lenses) are an integral part of the optical optimization. Either way, when a filter is used, the resulting optical path will change and affect the optimal microadjustment. So if you more oftenly than not use a (front mounted) filter, I'd suggest to also use it for AF microadjustment, even though it might reduce the moiré amplitude a bit. In fact, the target allows to determine if there is a noticable effect with/without filter use, given a certain calibration.

Moiré manifests itself as coarser patterns (aliases) than the underlying structures giving rise to the effect. Their shape/color/both depends on the optical distortions and focus differences that are introduced by the lens. When you see 'larger/coarser' structures appearing as you focus, you're witnessing aliasing effects.

#45

Here are 2 examples of what moiré could look like:

The example on the left shows saturated color bands instead of a 50% gray background. The example on the right shows a mottled (and colored) background. Both types of pattern reach their maximum amplitude when focus is optimal. The size/frequency, and amplitude of the patterns depends on shooting distance and focal length. The higher the optical quality of the lens at its maximum aperture, the easier it will generate such patterns.

The shape and size of the aliases is hard to predict, so one should look for any background pattern that's not uniform gray. When it is hard to get a clear moiré effect, try shooting from another distance. When the camera's optical axis is perpendicular to the LCD screen, then the patterns will appear uniformly across the display. In the example above one can see a slightly higher amplitude towards the lower left, which indicates a slightly non-perpendicular shooting angle (despite the setup alignment with a mirror).

# 56

Taking the shot is not necessary for the AF Microadjustment procedure that I presented in this thread. Optimal focus can be determined without firing a single shot, just by looking at the rear LCD in 'LifeView' and maximizing the amplitude of the aliasing patterns by optimizing (auto) focus.

#58

1. Find a distance at which you can achieve obvious moiré effects.
2. Use AF to achieve focus.
3. Using Life View, and assuming sub-optimal focus (otherwise quit ;-) ), note in which direction you need to manually adjust focus to maximize moiré.
4. Depending on the amount of adjustment needed, choose an adjustment amount and adjust in the direction to compensate for the direction of adjustment needed as determined in step 3.
5. go back to step 2. and do another iteration.

And here are the links of the different targets:

First

Second

 

[Nicolas ROLLAND]

Thanks for taking the time to assemble all this. I read the thread from top to bottom, but you are right I have to try again and follow scrupulously Bart's directions. I'll let everyone know how it goes.

 

[Bart_van_der_Wolf]

Hello,

As this is my first post here, first I would like to thank you all for this really interesting (geeky ?) thread. I am a French 1D Mark III owner (got it one week ago) and naturally as I have started to digg into the camera features, I became interested in micro-adjustments.

So far I have tried inanimate objects targets, paper focus targets (flat and 45°) and now Bart's LCD moiré target. But no luck. Whatever I do, I get really inconsistent results.

Hi Nicolas,

Sorry that I didn't respond earlier. I did read this post, but I was not entirely clear on what exactly you meant by "inconsistent results". Then I got occupied with other pressing matters, and it slipped my attention.

So, back to the issue, what's inconsistent? Can't you achieve an optimal adjustment setting, or do you still get front/back focus after settling on a certain setting?

I have tried to micro-adjust (@ full zoom, wide open) my Canon 17-40L, Canon 100 macro 2,8 and Sigma EX DG 70-200 2,8 and while I think I get a fair enough result at say +5 if I just try again and shoot some more tests say -10 I see no real difference with the previous +5 tests. I am probably missing something but I really don't get it.

What do you mean with that you see no real difference? Are you referring to the front/back focus, or do you get moiré at significantly different settings? I assume you do get moiré at a specific focus position, correct?

Bart

 

[Nicolas ROLLAND]

Hi Nicolas,

Sorry that I didn't respond earlier. I did read this post, but I was not entirely clear on what exactly you meant by "inconsistent results". Then I got occupied with other pressing matters, and it slipped my attention.

Hello, no please excuse me for not being at all... You know what it is like, new toy, tests and no results can be quite disturbing

So, back to the issue, what's inconsistent? Can't you achieve an optimal adjustment setting, or do you still get front/back focus after settling on a certain setting?

While I do seem to get an optimal adjustment setting, I still get front/back focus.

What do you mean with that you see no real difference? Are you referring to the front/back focus, or do you get moiré at significantly different settings? I assume you do get moiré at a specific focus position, correct?

Well I do get moiré at several settings with a best effect at a specific setting BUT when this setting is applied, I do not get better focusing results in real life situation than with a completely opposite setting (ex: best moiré @ +2 but settings @ +2 or -10 gives nearly same sharpness at autofocus target). So to me it is inconsistent as I should get sharper focus at the best moiré setting and significantly worse focus on such a different setting. And I don't, I am no even able to tell apart the shots if not aware of the applied setting.
That is what I don't understand.

Hopes this is a better explanation as I realize it is probably as hard to understand as it is to explain...

Thanks,

Nicolas

 

[Bart_van_der_Wolf]

While I do seem to get an optimal adjustment setting, I still get front/back focus.

Okay, that helps to clarify. First, we need to realise that AF is not perfect. There will always be some spread around the optimal focus position, due to mechanical/electronic variations and specific optical input. The goal of the AF Microadjustment is to get the bias out of the equation, so we on average get a correct focus, with equal chance on front/back focus. Getting this bias correction right IMHO does help to get more reliable focus, because I notice less jitter once focus is achieved.

Well I do get moiré at several settings with a best effect at a specific setting BUT when this setting is applied, I do not get better focusing results in real life situation than with a completely opposite setting (ex: best moiré @ +2 but settings @ +2 or -10 gives nearly same sharpness at autofocus target).

Okay, but with which autofocus target (you mentioned trying several different ones)? There is also quite a difference between calibrating e.g. a lens with an f/4 widest aperture, and an f/2.8 or wider one. The wider the aperture, the more visible the shift of focusplane will be. Also, the differences in the actual focus plane at adjustment settings of 10 or so, while noticable, are not huge. The adjustments are very fine pitched.

So to me it is inconsistent as I should get sharper focus at the best moiré setting and significantly worse focus on such a different setting. And I don't, I am no even able to tell apart the shots if not aware of the applied setting.
That is what I don't understand.

That is puzzling, I agree. Maybe we should tackle one specific lens, e.g. the 100mm macro, and see what happens in practice when you go through the full range of adjustments, say in steps of 10 (so use -20, -10, 0, 10, and 20).
The first thing you should notice is that the AF will focus on increasingly more distant positions on the focus ring. If it doesn't then the microadjustments are not functioning.
Secondly, when you make actual shots at f/2.8 on a tripod, e.g. close-ups of a newspaper at a 45 degree angle, you should see a shift of the AF focus plane. Do note that an angled surface is less useful for a calibration itself, but it could be used as a test of actual performance. However, there can be shifts of the focus plane that are due to the nature of the subject, not the adjustment alone. There should however be a trend as you make more tests at various adjustment settings.

Bart

 

[Nicolas ROLLAND]

This is very clear, thanks for all this ! I HAVE to try on a paper sheet to really check focal plane evolution. This is just so obvious... It is a bit late tonight so I will have to wait a few days before I try again as I won't have access to my camera.

To answer about the target I used,, well I used both your target target and the circles one and also paper targets like these : 1 - 2 but always flat and parallel to focusplane.

Whatever I will find, I will report results here. May I post pictures to illustrate here (linked from say Flickr at medium size) ?

Cheers

Nicolas

 

[Bart_van_der_Wolf]

To answer about the target I used,, well I used both your target target and the circles one and also paper targets like these : 1 - 2 but always flat and parallel to focusplane.

Good, parallel to the focus plane at least reduces one variable for testing.

Whatever I will find, I will report results here. May I post pictures to illustrate here (linked from say Flickr at medium size) ?

Yes, upto 4 images for inline display per post are allowed, but non-resized crops are preferred when looking at detail. If you use html links to Flickr you can post the full size files there, and post the URL(s) here.

Bart

 

[Nicolas ROLLAND]

OK so I tested with my 100 /2,8 macro. First with the LCD moiré target and then checking results with a printed text sheet @ 45 degrees from focus plane. I think I got it right this time. I miss time to post a full walk through today but I surely will.
Here are the neutral and the best setting I ended with (100% crop but still large so big files, be advised):

Target on LCD @ +0

focus is on the word "subject" @ +0

Target on LCD @ +2

focus is on the word "subject" @ +2

I would say that this macro lens makes it way easier to see the moiré. I only took a few shots with my Sigma 70-200/2,8 and with this one it is really much harder to get a good feeling. I'll report on it here too anyway. Just have to take time.

 

[Ralph Honsbeek]

Hi again ,

please tell me again how to find out if a lens is back focusing of front focusing , the basics

thanks

 

[Bart_van_der_Wolf]

please tell me again how to find out if a lens is back focusing of front focusing , the basics

Hi Ralph,

Just Auto-focus on the target, enter Life View, and then observe which direction you need to turn the focus ring to improve focus (increase aliasing). If you need to focus closer, you're backfocused, if you need to focus farther, ...

Bart

 

[Ralph Honsbeek]

Gee thanks that makes a lot of sense now that you made me think about it , thank you ery much !!!

 

[Thomas Kuehne]

Dear Bart,

thanks ever so much for this technique. I had a quick play with it and it is so much more convenient than using focus charts.

Here are two questions:

1. Some AF systems apparently are influenced by the colour temperature of the target. A lens backfocusing under tungsten light, can be just fine in daylight. If the colour temperature of the target display used isn't sufficiently close to daylight, this may affect the accuracy of daylight focusing. Have you validated your calibrations with focus chart tests under daylight conditions?

2. My camera doesn't have live view, so I tried to spot the moire pattern through the viewfinder. It works, the moire can clearly be observed (I'm assuming the second "grid" is the structure on the focusing screen), but it is difficult to tell when the moire is at its maximum.

I therefore slightly tilted the target screen, so that when I focus, the moire pattern will wander up and down the screen as I adjust. Pointing the lens at the centre spot of the test pattern, I know that I have achieved optimal focus when the moire pattern is centred around the centre spot.

Do you think that is a admissable variation of your test?
I have yet to confirm my calibrations with daylight focus chart test, but didn't want to wait to thank you for your great work and get your feedback on my comments.

 

[JohanElzenga]

y camera doesn't have live view, so I tried to spot the moire pattern through the viewfinder. It works, the moire can clearly be observed (I'm assuming the second "grid" is the structure on the focusing screen), but it is difficult to tell when the moire is at its maximum.

I don't think you can use the viewfinder for this. The viewfinder is a separate optical system, so things can look perfectly in focus in the viewfinder and could still be slightly out of focus on the sensor. In fact, that is probably how back focus and front focus can occur in the first place.

 

[Thomas Kuehne]

The viewfinder is a separate optical system, so things can look perfectly in focus in the viewfinder and could still be slightly out of focus on the sensor.

That's right. This variation assumes that the focusing screen is perfectly aligned. If it isn't then manual focusing will lead to suboptimal results too.

In fact, that is probably how back focus and front focus can occur in the first place.

Front focus (FF) and back focus (BF) created by the camera occurs because the optical path length to the AF system differs from the optical path length to the sensor. Once this is fixed/calibrated, FF/BF should not occur at all with any lens. As to why some lenses will still FF/BF even with a perfectly calibrated camera is a mystery to me.

Bart already responded to my posting, but his response and my response to it got lost in a recent system problem. Responding to my mentioning of colour temperature influencing the AF system, Bart said some lenses have different focal lengths depending on the colour temperature. I was wondering whether prism-based AF systems may also be upset by a colour temperature which deviates considerably from daylight. While it is possible to calibrate screens to 6500K, many screens may not necessarily be calibrated. Also, an LCD display will always use three (RGB) components in a mix to recreate a certain stimulus in the retina. A single wavelength, say corresponding to yellow, will be emulated by a mixture of a two wavelengths, corresponding to red and green. I'm not saying this makes the test invalid, but it is food for thought. The AF system may, or may not, be fooled by the specific colour recreation process of screens.

 

[Bart_van_der_Wolf]

Front focus (FF) and back focus (BF) created by the camera occurs because the optical path length to the AF system differs from the optical path length to the sensor. Once this is fixed/calibrated, FF/BF should not occur at all with any lens. As to why some lenses will still FF/BF even with a perfectly calibrated camera is a mystery to me.

There are 2 types of calibration. One is performed on the body itself. AFAIK that is a hardware calibration, and it involves the position of the (sub-)mirror (maybe different for non Canon models). Maybe there is software/firmware involved as well. The second calibration is of the lens electronics, and involves the lens response to the commands by the camera body. It allows to fine tune the angle of rotation invoked by the focus motor, and it can take care of specific tolerances in the electronics and/or mechanics of the lens. Everything within the tolerance ranges, which are inevitable.

Bart already responded to my posting, but his response and my response to it got lost in a recent system problem. Responding to my mentioning of colour temperature influencing the AF system, Bart said some lenses have different focal lengths depending on the colour temperature. I was wondering whether prism-based AF systems may also be upset by a colour temperature which deviates considerably from daylight.

Yes, but perhaps to a lesser degree due to the spectral sensitivity of the human eye (less sensitive at the extremes of the spectum). It may be a good thing to repeat something else that got lost in the OPF system issue. Not all lenses are equally sensitive to wavelength differences, it depends on their correction for Chromatic Aberration.

I'm currently trying to calibrate a Canon EF 135mm f/2.0 L lens, which exhibits clearly visible CA when used wide open. I need to do some more testing, but I think I'm experiencing an AF position shift as the light changes from daylight to Tungsten (maybe in the order of 0.5% of the actual focus distance, which is very noticable at f/2.0 and that is also used to AF). I can only see that in practical AF verification after AF microadjustment, as my calibrated LCD screens have a constant 6500K color temperature.

This also raises a question about the spectral response of the AF sensor.

While it is possible to calibrate screens to 6500K, many screens may not necessarily be calibrated. Also, an LCD display will always use three (RGB) components in a mix to recreate a certain stimulus in the retina. A single wavelength, say corresponding to yellow, will be emulated by a mixture of a two wavelengths, corresponding to red and green. I'm not saying this makes the test invalid, but it is food for thought. The AF system may, or may not, be fooled by the specific colour recreation process of screens.

Yes, uncalibrated screens are usually quite bluish, 9500K or thereabouts. But the LCD colors R,G,and B have broad band emissions which probably overlap in most LCD screens. The AF sensor is probably also sensitive to a broad spectral band, but not necessarily equal to the human visual system or the camera's sensor, so CA may matter unless it's only green sensitive.

Bart

 

[Ron Morse]

I have to thank Bart for the target and explaining how to use it in an easy to understand way. I also have to thank Nicolas for directing me to this great thread.

I adjusted almost all of my lenses. I haven't done my 70-200 2.8 IS or the 500 yet. Most were 1 to 3 +/-. One was I think was + or - 6. Then I got a big surprise, I set up my 85/1.8 that I have never got decent pictures from on the 20D/40D and now the 5DMII. I went to - 20 and it would need from a guess about -23 to get it perfect. Right now it is pretty darn sharp but I know it needs a little more. I guess it needs a trip to canon. The lenses that I always thought were good only needed +/- 1 to 3 and now are even better.

I'm a happy camper. Thanks, Thanks, Thanks.

 

[Bart_van_der_Wolf]

I have to thank Bart for the target and explaining how to use it in an easy to understand way. I also have to thank Nicolas for directing me to this great thread.

Hi Ron,

I'm glad to have been of service.

Don't forget to also test the final settings at the actual shooting distances you use most. Depending on how the lens is calibrated internally, there may be a difference between close focusing and distant focusing.

Also, because a lens needs a microfocus adjust on one body, it doesn't mean that it requires the same or any on a different body. It's the combination that may need tweaking.

Bart

 

[Mike Shimwell]

Bart and Ron

Although it's absolutely the case that a lens may need different adjustments on different bodies I had a similar experience when I set mine up. All bar one of my lenses required no adjustment, but my 85 1.8 has a 5 adjustment on the 1Ds3. That was the lens that I found consistently slightly disappointing on the 5D - not awful, but not what I expected. Correctly adjusted it's exactly as I thought it should be.

Mike

 

[Ron Morse]

Well, I really hate to admit it but I messed up with the 85/1.8. I tried it and it was off quite a bit. When I tried it first I was stopped way down and it seemed pretty good. I tried it wide open and it was worse than before. I set up this morning as far away from the computer as I could get in the room and it became so much easier. Plus 5 and its right on the money. It was very difficult to tell if +5 or +4 was best but I think +5 was slightly better. First time with the micro-adjust but I can't believe how far off I was. I didn't realize how critical distance is.

I tried the 70-200 2.8 IS but can't get back far enough in this room to do anything. I'll set a lap-top up on the kitchen table and give it a go.

 

[Bart_van_der_Wolf]

I didn't realize how critical distance is.

Hi Ron,

That's why I suggest people to do some more verification tests at common shooting distances. That may be hard to do with an LCD screen at larger distances.

The reason behind it is that the internal calibration settings for a lens may result in a different offset at different distances. Zoom lenses even add another variable, focal lengths.

The situation even becomes more critical with wide aperture lenses. On the one side, they allow to pin-point the focal plane offset more accurately, on the other side they introduce residual lens aberrations like spherical aberration.

One of the most notorious lenses is the 50mm f/1.2, because it's widest aperture is quite a bit wider than the focus system's, which seems to be based on an f/2.8 aperture. Yet, autofocus is based on the widest aperture that a lens provides.

When I look at the Chromatic Aberrations of my 85mm f/1.2 and 135mm f/2.0 lenses wide open, I wonder which wavelengths affect the AF sensor most (the green or the magenta OOF areas, or the in-focus average). The CA is quite noticable on high contrast edges, but it's everywhere else as well. Any discrepancy between short and long focussing distances will aggravate the situation.

In theory, the official calibration at a (Canon) service centre should include all the variables, but I have my doubts about the actual accuray of their procedures. Therefore it's a good thing we have at least some control over the final offset parameter.

Bart

 

[Bob Slassor]

I have used this method to set up my 24-70L on my 1Ds Mk3. It is a big improvement to the usual 45 degree chart I have been using. I went out and took some very ordinary pics to see if there was a difference in real terms after setting up my lens with this method and there was. Brilliant. For the record, when I used the 45 degree method, I had my adjustment set to -2. Using this method, my adjustment is now +5. Thanks so much for posting this.

 

[Bart_van_der_Wolf]

For the record, when I used the 45 degree method, I had my adjustment set to -2. Using this method, my adjustment is now +5. Thanks so much for posting this.

Hi Bob,

Thank you for the feedback. It kind of demonstrates that calibration based on a tilted surface can give sub-optimal results.

When my AF Microadjustment target is not parallel to the sensor plane, it shows as a difference in the moiré amplitude across the image. I try to align my LCD and camera with a mirror (center the lens reflection in the viewfinder). However, the moiré method is so sensitive that even then it is hard to get things perfectly squared, but at least the target will give a signal. At some distance it gradually becomes less critical for calibration, and it's easier to stay within the narrow DOF of a wide open lens.

Bart

 

[Peter Galbavy]

Bart,

Can you perhaps consider editing / supplementing the first posting with the various recommendations and tips that have come up in the forum since your excellent first port over a year ago ?

I have now got lost as to what is true best practise using your LCD chart...