Next chapter in our ongoing 50D review:
Auto Focus (AF) Microadjustment
(Also known as C.Fn.III-7 fun-o-rama!)
I’ll fess up, one of the first things I did after getting my Canon 50D was to pop my EF 24-105 f4L IS in place, put the camera on a tripod, point the lens at the concentric circle AF Microadjustment target that seems to be making the rounds, and start trying to tweak my auto focus. I thought it would be quick and easy – and that I’d see great new sharpness from all of my shots… <chuckle>
Okay – so I didn’t think it was going to be that easy, but I didn’t have much luck in my first attempt.
The pattern I used, which I found by searching with good old Google, is posted later in this post. I found the first versions of it in this thread on openphotographyforums.com. I also found a page w/ a similar target (same pattern, different file size) on Northlight Images’ website here; but the instructions seemed a touch incomplete to me. (As an FYI – I do think the Northlight page – the one linked to the word "here" – is worth reading – it is much shorter than the series of posts, responses, questions, etc. in the thread.)
Allow me to ‘splain – no – there is no time…
Allow me to sum up…
In short – you need to use a special target on an LCD based monitor (a printout will not work – it has to be shown on a LCD) – the test procedure counts on the interference between the LCD screen showing the pattern and the sensor to show as a pattern (called moiré – pronounced "mwah-ray") when the focus is sharp. Show the pattern at 100% on said monitor, set the camera for center AF point only, aim your camera square to the screen at the center of the pattern, turn on live view, get the most distinct interference pattern you can, turn off live view, and then watch the distance indicator window on your lens as you half press the shutter to have the AF focus on the screen. If the scale moves in the window, the AF Microadjustment is off… Make a tweak and then re-test. Oh – and be sure to do each lens individually, I’d be shocked (and would tell you to go out quick and buy a lottery ticket) if all your lenses came back the same correction…
Sounds pretty easy… but… I had just a few questions after reading the post I reference above:
- How far away should the camera be from monitor showing the target?
- What aperture should I shoot with?
- So, if the distance scale does move, which way do I need to "adjust" the lens? + or – ?
- Is the pattern test the best test?
The first attempts were admittedly rushed, and I had a hard time seeing any real difference, and I thought that zero adjustment was the best for the 3 or 4 lenses I tried… But I knew intuitively that couldn’t be the case, so I stopped that test, and got on with some other playing with my new camera that day.
After doing some shooting, the dead pixel testing with its discoveries/fix (doh! – been meaning to post that one – perhaps Thursday), and the shooting of an event behind me, it was time to get back to the AF Microadjustment.
So today I got things setup; and, being just a touch less rushed, I was able to refine my process just a bit (okay – quite a bit), and was soon seeing some results.
After reading through the long series of posts from the OPF, I was clearer on a few things. 1 – keep your lens zoomed to its maximum zoom level; and 2 – be sure to keep your aperture cranked as open as possible (smallest number your lens will shoot). This does two things – it ensures the shallowest depth of field, and it ensures the shallowest depth of field… (no – that wasn’t a typo, both actions minimize the DoF).
So having thought on this a bit since my first attempt, I figured that when the tiny pixels on my laptop’s high res but not huge screen combined with the finer resolution of the sensor on the 50D, and with the finer resolution on the camera’s screen, perhaps the moiré wasn’t the best test for the 50D AF Microadjustment. I quickly created my own test pattern with one pixel wide lines in a grid, with a few of the lines either wider or a different color so you could tell where you were in the grid when zoomed in on live view. My thought was that I should use laptop monitor at a 45 degree angle to the camera, and I could then use the tight DoF and the snazzy new screen to pixel peep to see which individual pixels on the laptop monitor were out of focus relative to where the center point focus pipper was placed. I guessed it would work better than the tutorials which had been created for the 1D Mark III and 1Ds Mark III with their lower resolution screens.
To save details – nope – didn’t work so great for the short lenses, but did come in handy later… I found that the moiré works great for wide to medium lenses, and that my pattern worked better for longer lenses.
The Answers: (well – what worked for me at least)
Keep in mind that you should have the zoom at its maximum zoom – I had live view set to "Quick Mode" and never hit the "AF-On" button – I manually turned the focus ring. Oh yeah, and one shot AF with the center AF point selected.
1. How far away should the camera be from the monitor showing the target?
When testing wide to medium zoom (say, up to about 100 mm), I used the circular pattern and found that it helped to have the pattern come close to filling up the frame top to bottom. For my 24-105, it turned out to be about 50 inches from screen to where the sensor is in the camera (you know that little circle w/ the horizontal line through it that you see when looking down on the top of your camera? That horizontal line shows where the lens is focusing – the front of the sensor); for my 17-55 – about 30 inches; and something like 20 inches for the 10-22. When in live view, just ensure you see the interference pattern clearly – play around with the 5x and 10x live view if need be.
2. What aperture should I shoot with?
Use the most wide open aperture (the smallest number) to minimize DoF.
3. So, if the distance scale does move, which way do I need to "adjust" the lens? + or – ?
I was a touch confused by the "Forward" and "Backward" labels on the adjustment screen. I then almost scratched a hole in my head trying to come up with a way to explain what they meant – I soon gave up and came up with how I got my head around it.
Seated behind the camera, I’d see the distinct pattern in live view; I’d turn off Live View then half press the shutter. If I saw (for example) the distance scale shift to the right as I looked down, that meant I had to turn the focus ring counter clockwise (or – ) to get the moire pattern back. If I saw the scale jump to the left, that meant I had to turn the focus ring clockwise (or + ) to get the pattern back. So I soon ignored the words on the adjustment menu, and just moved the pipper to the – or the + as needed.
4. Is the pattern test the best test?
Well – yes and no. I found that for my 10-22, 17-55, and 24-105 – the circular interference method worked the best; but when the longer lenses came out, my 70-200 (zoomed to 200 – remember – zoom to high end) and 100-400, I found that it took so little to shift the pattern, that I had difficulty getting a distinct pattern to be consistent. It might have been that I was testing in my kitchen and was running out of room to move the tripod away from the monitor, but I found using my target worked pretty darn well. Again – show the target at 100% view, but this time, turn the monitor on an angle so that it is about a 45 degree angle to the camera. Center the camera on the thick lines that cross in the center of the target. Now turn on live view, zoom to 10x and you should be able to manually focus so that the vertical line is sharp, with DoF fall-off to the left and right (assuming you pivoted the whole laptop on the counter/desk, if you tilted the monitor away – the horizontal line should be sharp w/ DoF fall-off above and below). Now turn off Live View, half press the shutter to let the AF operate, and then turn Live View back on – zoom back to 10x… if the live view display has changed, you easily be able to see if the focus has shifted to a portion of the screen that is closer to you (if so, you would need to turn the focus ring clockwise, so add to the AF Adjustment) or if it is focused further away (you need to turn the focus ring counter-clockwise to get the center point back in focus, so subtract to the AF Adjustment).
A few pics w/ my target being used (click to enlarge):
From left to right:
the pipper on the crosshairs of my target;
this 10x zoom shows what it would look like if you needed to subtract correction.
micro adjustment, micro adj, nikon, 1D mark III, 1D mk III, 1Ds mark III, 1Ds mk III, sigma, tamron, micro-adjustment
Okay – so that’s it then? You betcha!
But wait – there’s more!
NOW how much would you pay?
For those who are interested in a longer winded version of the test… here goes… I’ll just try to bang this out as it is getting late and I do need to get some shut eye.
1. Be patient – take your time – don’t rush this – or you will get *very* frustrated. The moiré pattern’s changes as you barely nudge the focus ring can be VERY SUBTLE (in caps for irony’s sake) – so don’t do this after three espressos and a red bull.
2. Setup your camera on a very steady tripod. When you get to your longer lenses, you will notice any subtle movement is magnified greatly by the distance to the screen. (See step 1…) Ensure your camera is leveled.
3. Display your choice of target at 100% on a LCD display – I put my laptop on the kitchen counter.
4. Set the height of the center of your lens to be the same height as the center of the target you are using.
5. If you are using the moiré target, square up the camera to the screen – if you are using my target, angle the screen at about a 45 degree angle to the camera.
6. Set your Live View to quick mode, your aperture to wide open (lowest number), set AF to use only the center point, and set the distance between the camera and the target to the closer end of the focus range of the lens. As an example If your lens can focus at 1/2 meter, then move the camera to about 3/4 of a meter away from the target. The target should fill most of the height of the frame (you don’t want it too big – unless you are using my target). Some lenses that have IS (image stabilization) will sense if the camera (or lens) is mounted in a tripod, but to be sure – I suggest you turn it off manually to be sure.
7. Focus on the target in Live View – If you are using the circular pattern target, get the most moiré pattern you can (see the example images on this page from Nortlight-images.co.uk), or if you are using my target – focus on the intersection of the thick black lines (3 px wide) at the center of the target.
8. Turn off Live View, watch the focus distance scale in the window on your lens to see if it moves while you 1/2 press the shutter to activate the camera’s auto focus. remember which way the focus scale moved
9. If the focus distance scale didn’t move – you are done! (If this happened the first try – you might try steps 7-9 a few times to ensure that your precision in step 7 was accurate) You can also verify by going back into Live View and seeing if you see the same amount of moiré you did before. Or if you are using my scale, see if the thick line is still in focus properly.
If the scale did move – proceed to step 10
10. hit the "menu" button, index wheel over to the Custom Functions menu (the second from the end) and choose "C.Fn III: Autofocus/Drive", then choose 7 – AF Microadjustment, and ensure it is set to 2 – Adjust by lens.
11. If this is the first time AF Microadjusting this lens on this camera, hitting the "info" button below the LCD will "register" the lens and take you to the AF-Microadjust screen; if you’ve had this lens on before and adjusted it, or if you are looping through, making changes and checking, hitting the "info" button will allow you to "change" the microadjustment in the AF-Microadjustment screen.
12. Now – make an adjustment to the correction value. If it moved a lot, try adjusting the value by 10, if it moved a little, try 5. + or – ? Well – if you are using the circular rings target: If the distance scale shifted to the right – adjust in the minus ( – ) direction… If the distance scale shifted to the left, adjust in the plus ( + ) direction. My Target: if the focus shifted to a portion of the pattern that is closer to you, adjust in the plus ( + ) direction… if it is focused on part of the pattern further away, adjust in the minus ( – ) direction.
13. Be sure to hit the "Set" button (the button in the center of the thumb-dial) to save your change!
13. Now go back up to step 7 and repeat the process until you do not see the focus ring shift.
nikon, sigma, tamron, micro adjustment, micro adj, 1D mark III, 1D mk III, 1Ds mark III, 1Ds mk III, micro-adjustment
Okay – that pretty much wraps it up…
Just as an illustration, here are the adjustments that my lenses needed:
|EF-S 10-22 f/3.5-4.5
|EF-S 17-55 f/2.8 IS
|EF 24-105 f/4L IS
|EF 70-200 f/2.8L IS
|EF 70-200 f/2.8L IS + 1.4x TC
|EF 70-200 f/2.8L IS + 2x TC
|EF 100-400 f/3.5-5.6 L IS
Be sure to note that the camera does save separate micro adjustments when you are using TCs with your longer zooms. So there are three adjustments saved with the 70-200 f/2.8L IS, one by itself, one when using the 1.4x TC, and one when using the 2x TC. Pretty cool.
The moiré target from this thread at OPF:
My target – feel free to share early and often:
Okay – I’m going to have to crater now – good night and thanks for reading!
micro adjustment, 1D mark III, 1D mk III, 1Ds mark III, 1Ds mk III, nikon, sigma, tamron, micro adj, micro-adjustment