Home · Register · Join Upload & Sell

Moderated by: Fred Miranda

New fredmiranda.com Mobile Site

New Feature: SMS Notification alert

New Feature: Buy & Sell Watchlist

FM Forums | Canon Forum | Join Upload & Sell

2 3 … 37 38 end

Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Fred Miranda
Offline
Admin
Upload & Sell: On

p.1 #1 · p.1 #1 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Fellow FM member snapsy has developed DotTune: a free and fast method for fine-tuning the autofocus on your lenses.

"Unlike existing methods, with DotTune you don't need to take test photographs at multiple AF tune values to find the best tune value. Instead you focus once in Live View to establish critical focus, then use the focus confirmation dot within the viewfinder to find the range of AF tune values which provide a reliable focus confirmation, then simply select the midpoint of that range as your final and optimal AF tune value."

Watch video tutorial below:

Segment Links:
00:00 : What is DotTune?
00:51 : DotTune Quick Reference
01:56 : Step 1 - Set up your AF Target and Camera
03:08 : Step 2 - Establish critical focus in Live View
03:35 : Step 3 - Switch to Manual Focus and set AF Tune to 0
04:18 : Step 4 - Evaluating Viewfinder Focus Confirmation
07:54 : Step 5 - Cycle AF Tune values to find confirmed range
11:42 : Step 6 - Set final AF tune value to midpoint of range
12:51 : Additional Notes
15:02 : Sample DotTune session

Download Bob Atkin's focus chart used in the video.

Here is a link to the DotTune Midpoint calculator used in the video.

Nikon version of the same AF tuning method started on the Nikon SLR board.

Feb 19, 2013 at 02:39 PM

snapsy
Online
• • • • • •
Upload & Sell: On

p.1 #2 · p.1 #2 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

It's free, it's fast, and it's much more accurate than any other method I've seen so far. What makes it faster and more accurate is that it doesn't rely on actual phase-detect focusing to derive the AF tune value. Because PDAF has shot-to-shot variability, all existing AF tune techniques require multiple images taken at each tune value to eliminate that variability from the equation. And it's the need for multiple shots that make those techniques more prone to AF tune measurement error. The concept is based on a recent discovery about the true nature of AF tuning - you can read about that here (I'm "horshack" on dpreview).

I developed and tested this method first on a Nikon D800, and it has been confirmed by several other owners so far (Nikon thread here). I've also tried it on my 5DM3 with a Sigma 35mm and it appears to work. I'm looking for feedback from other Canon owners to confirm it on their bodies+lenses.

Canon version of technique

1. With the camera mounted on a tripod, enter Live View and establish critical focus on a high-contrast subject over the center point. Exit Live View.

2. Set the lens to MF (Manual Focusing). Be careful not to jostle the focus ring when you move the switch - you don't want to alter the focus you established in step #1.

3. Set AF tune to 0.

4. Look through the viewfinder to see if the camera thinks the subject is in focus by checking if the green focus confirmation dot is lit. You'll need to half-press the shutter (or AF-ON) while doing this to keep the metering/rangefinder from going to sleep. Assuming you get confirmation, increase the AF tune value and recheck the viewfinder, leaving the lens set to the original focus you established in step #1 - do not refocus the lens. You want to keep increasing the AF tune value until you no longer get a green confirmation dot. As you reach the edge of the range the green dot will either fail to light up or it may flicker...you may need to cycle the half-press of the shutter several times to confirm that the confirmation you get is consistent. You're done when you find the highest positive AF tune value that gives you a consistent green-dot confirmation - that value represents the positive end of the AF tune range.

5. Return AF tune to 0 and repeat the procedure in step 4, but this time decreasing the value of AF tune (going negative). When done you've established the negative end of the AF tune range

6. Calculate your final AF tune value, which is the midpoint of the range you established in steps 4 and 5. For example if your range was -2 to +8 then the midpoint is +3. If your range was 0 to +12 then the midpoint is +6. Set your AF tune value to your calculated midpoint. You're done.

Notes
* If you don't get a green confirmation dot at an AF tune of 0 then increase or decrease the AF tune value until you do. Your goal is to find the range of AF tune values that show focus confirmation - that range may be all positive or all negative, instead of the -/+ range described in the steps above.
* If either end of the AF tune range (-20 or +20) still gives you focus confirmation, this means the range of AF tune values is wider than the camera's supported +/-20 AF tune range. Marianne Oelund came up with a clever workaround for this here.

Edit 2/10/13: Decided to call this technique "DotTune", to have a quick and searchable moniker when referring to it. It's a bit more identifiable than just "new AF tune technique"

Edit 2/17/13: I've created a simple midpoint calculator on my website. DotTune Midpoint Calculator

Edit 2/18/13: I've published a DotTune tutorial on YouTube.

" target="_blank">DotTune Youtube Video

Edited on Feb 18, 2013 at 05:31 PM · View previous versions

Feb 03, 2013 at 11:56 AM

Fred Miranda
Offline
Admin
Upload & Sell: On

p.1 #3 · p.1 #3 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Sounds interesting snapsy. I will try it today. For me 'distance' is the biggest variable. For example, I can AF tune the Sigma 35 to nail focus at 4 feet but at 15 feet, AF misses the mark...and vice-versa.
Fred

Feb 03, 2013 at 01:23 PM

snapsy
Online
• • • • • •
Upload & Sell: On

p.1 #4 · p.1 #4 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Fred Miranda wrote:
Sounds interesting snapsy. I will try it today. For me 'distance' is the biggest variable. For example, I can AF tune the Sigma 35 to nail focus at 4 feet but at 15 feet, AF misses the mark...and vice-versa.
Fred

Thanks Fred. Irrespective of the AF tuning method used, I've found that for large-aperture primes it's best to tune for far distances rather than shorter. There's an asymmetrical relationship here - if I tune for short distances then far distances are OOF with that tune value. However if I tune for far distances then short distances still work fine. I have a few theories as to why this is but I haven't spent the time to prove them out yet.

Feb 03, 2013 at 01:34 PM

jstntym
Offline
• •
Upload & Sell: Off

p.1 #5 · p.1 #5 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Interesting snapsy! Reminds me of when we had to calibrate voltage sensing relays on a voltage throwover panel. Dialing the lens AF tune sounds similar here, I'll give it a go!

Edited on Feb 03, 2013 at 02:41 PM · View previous versions

Feb 03, 2013 at 01:43 PM

skibum5
Offline
• • • • • •
Upload & Sell: Off

p.1 #6 · p.1 #6 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

sounds like it might work out well

i've gotten pretty good results simply aiming at a crack in the sidewalk (or some such) at an angle and adjusting until snaps tend to center best on the crack and it works out well, or for sports, fine tuning by aiming at some player standing around and getting the grass DOF at their feet to be placed as you want it; quick and dirty but it works well and can be quickly adjusted to exact task at hand in the field

I will give it a try, but NOT today. Today is the SUPERBOWL!!!!!!!!!

Edited on Feb 03, 2013 at 01:52 PM · View previous versions

Feb 03, 2013 at 01:49 PM

skibum5
Offline
• • • • • •
Upload & Sell: Off

p.1 #7 · p.1 #7 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
Thanks Fred. Irrespective of the AF tuning method used, I've found that for large-aperture primes it's best to tune for far distances rather than shorter. There's an asymmetrical relationship here - if I tune for short distances then far distances are OOF with that tune value. However if I tune for far distances then short distances still work fine. I have a few theories as to why this is but I haven't spent the time to prove them out yet.

Yeah that does seem to be. Usually tuning anything other than very close in works better overall.

Although for some lenses, like say 50 1.4 they tend to need tuning for at least two if not three different distances and tuning at a farther one can not be the best for close in work.

Feb 03, 2013 at 01:51 PM

EB-1
Offline
• • • • • • •
Upload & Sell: Off

p.1 #8 · p.1 #8 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

I tried a simplified centering method about a year ago with the dreaded D7000. Although it allowed more accurate assisted manual focus, the AF problems continued. Ultimately if the camera has poor AF reproducibility you'll never get consistently accurately focused shots. Of course it is useful to get the center point as close as possible to where it should be. Ironically if the AF has good reproducibility, it is a snap to obtain calibration with a lens Align or similar target.

Eventually the camera manufacturers should build in an automatic AF calibration function that can be done quickly in the field. I'd rather have that than the useless video.

EBH

Feb 03, 2013 at 02:02 PM

skibum5
Offline
• • • • • •
Upload & Sell: Off

p.1 #9 · p.1 #9 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Feb 03, 2013 at 02:05 PM

Fred Miranda
Offline
Admin
Upload & Sell: On

p.1 #10 · p.1 #10 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
Thanks Fred. Irrespective of the AF tuning method used, I've found that for large-aperture primes it's best to tune for far distances rather than shorter. There's an asymmetrical relationship here - if I tune for short distances then far distances are OOF with that tune value. However if I tune for far distances then short distances still work fine. I have a few theories as to why this is but I haven't spent the time to prove them out yet.

From your tests, what seems to work as best 'far distance' for fast primes like 24mm, 35mm, 85mm, etc?

Feb 03, 2013 at 02:38 PM

jj_glos
Offline
• • •
Upload & Sell: Off

p.1 #11 · p.1 #11 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

skibum5 wrote:
video also brought us liveview which even many strict stills shooters find far less than useless

Liveview was featured on dslrs way before video.

Feb 03, 2013 at 03:31 PM

snapsy
Online
• • • • • •
Upload & Sell: On

p.1 #12 · p.1 #12 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

EB-1 wrote:
I tried a simplified centering method about a year ago with the dreaded D7000. Although it allowed more accurate assisted manual focus, the AF problems continued. Ultimately if the camera has poor AF reproducibility you'll never get consistently accurately focused shots. Of course it is useful to get the center point as close as possible to where it should be. Ironically if the AF has good reproducibility, it is a snap to obtain calibration with a lens Align or similar target.

I agree completely. You have correctly discerned PDAF variability from AF tune, and it is indeed ironic that when using the typical AF tune techniques which rely on actual focusing it's harder to find the optimal tune value vs bodies with less shot-to-shot variation. This is why I think the method in the OP has so much value - it allows you to establish the AF tune value independent of mechanical shot-to-shot variability.

What complicates the matter is that AF mistune can manifest as variability, since it moves the relative midpoint of the PDAF mechanical variability. I wrote an article about it here.

EB-1 wrote:
Eventually the camera manufacturers should build in an automatic AF calibration function that can be done quickly in the field. I'd rather have that than the useless video.

Agreed. I wrote an post about that as well here.

Feb 03, 2013 at 03:49 PM

snapsy
Online
• • • • • •
Upload & Sell: On

p.1 #13 · p.1 #13 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Fred Miranda wrote:
From your tests, what seems to work as best 'far distance' for fast primes like 24mm, 35mm, 85mm, etc?

I tuned mine at pretty far distances, like 30-50 feet, which happen to reach the infinity position of the lens elements.

Feb 03, 2013 at 03:57 PM

stanj
Online
• • • • • •
Upload & Sell: On

p.1 #14 · p.1 #14 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
1. With the camera mounted on a tripod, enter Live View and establish critical focus on a high-contrast subject over the center point. Exit Live View.

2. Set the lens to MF (Manual Focusing). Be careful not to jostle the focus ring when you move the switch - you don't want to alter the focus you established in step #1.

Then just reverse the steps. You don't need to have the lens in AF mode to perform what you're doing in your step 1.

Feb 03, 2013 at 04:05 PM

cputeq
Offline
• • • • •
Upload & Sell: Off

p.1 #15 · p.1 #15 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

jj_glos wrote:
Liveview was featured on dslrs way before video.

+1

Liveview is older than DSLR video, not the other way around. In fact, I still remember reviewers, etc. asking "okay, if we can see a live stream of this on the rear LCD, how long before we can record it as video?" IIRC the Canon 1D3 was Canon's first DLSR to have it, the Nikon D90 followed about a year later, maybe a little more. I think Oly had something going on earlier than both of these also, but unsure.

As to the OP - I'll try it out, sounds like an interesting method at testing AF fine tune.

Feb 03, 2013 at 04:14 PM

Mike K
Offline
• • • •
Upload & Sell: Off

p.1 #16 · p.1 #16 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

An alternate PD AF calibration technique comparing to Live View 10X

I have a different variation based upon my confidence in Live View Manual Focus. This method requires a high resolution LCD; all recent models of Canon DSLRs since the 5DII have about 1M dots or more. The fundamental premise is the with Live View 10X, the focus will be exactly as you see it on the LCD. Focus wise; What you see is what you get.

1. Canon strongly suggests performing phase AF calibrations at 50X the focal distance of the lens. Thus for a 50mm lens this is 2.5 meters from the target to the camera. Don't do PDAF calibrations at really close distances, like MFD, as there is much greater chance of that calibration being way off. For telephoto lenses, choose a detailed subject some distance away.

2. I like to use a resolution chart with very fine interference lines such as

For a much larger higher resolution version of this image visit
http://forums.dpreview.com/forums/post/50791608

However your choice of charts or subjects is not critical. The target does not have to be a monitor or a large print and can be nearly anything with a uniform distance from the camera, but should be well lit and with good contrast and detail.

3. Use a solid tripod so you can see the camera LCD well.

4. PD AF on the chart above. Switch the Live View on and magnify the LCD image to 10X. Carefully adjust the lens to maximize the Live View focus. Turn LV off. Look at the lens as you press AF again. Did the focus band move? Alternatively you can use Live View AF "quick". Turn LV on and again manually turn the focus ring on the lens to maximize the LV 10X focus. Again turn LV off and watch the lens as you press PD AF. Based upon the amount and direction of miscalibration between the PD AF position and the LV 10X position, go into the micro adjust menu and change the MA calibration for that lens.

5. Do step 4 again, multiple times. Its very fast, so one can repeat the comparison of PD AF Vs Live View 10X several times in a few minutes. There is some imprecision in both the PDAF position and the LV10X focus position, so repeat the exercise until you are happy that the AF and LV results are very close to one another. If you have a 1dX, 5DIII or 6D body, you need to perform this calibration at both ends of the Zoom position. Repeat this exercise for all of your Canon AF lenses.

6. Variation with greater resolution: using the HDMI output, hook up your camera to an external monitor such as laptop or monitor such as

https://www.fredmiranda.com/smallhd/

Press the info button several times to maximize the screen resolution and adjust the monitor to yield 1:1 resolution (1 pixel on the external monitor = 1 pixel from the camera output). This will make it easier to visualize fine adjustments of in/out of focus. Perform steps 1-5 above.

Edited on Feb 04, 2013 at 01:05 PM · View previous versions

Feb 03, 2013 at 04:23 PM

RCicala
Offline
• • • •
Upload & Sell: Off

p.1 #17 · p.1 #17 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Well done!

Feb 03, 2013 at 04:37 PM

Gunzorro
Offline
• • • • • •
Upload & Sell: On

p.1 #18 · p.1 #18 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

jj_glos wrote:
Liveview was featured on dslrs way before video.

True, but it was an offshoot of hand-held video cameras live monitoring off the sensor. Then the video was added to DSLRs after importing the video viewing technology -- no extra charge!

Edited on Feb 03, 2013 at 05:19 PM · View previous versions

Feb 03, 2013 at 05:15 PM

snapsy
Online
• • • • • •
Upload & Sell: On

p.1 #19 · p.1 #19 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Mike K wrote:
An alternate PD AF calibration technique comparing to Live View 10X

Hey Mike, I responded to you over on dpreview but I'll repost here for the benefit of those who don't like that visit that jungle

The problem with all the existing AF tune techniques, including the one you described, is that they all require one to perform PDAF cycles. That incurs the typical shot-to-shot mechanical variability seen on all PDAF systems, and so arriving at a AF tune value requires multiple shots to separate that +/- variability from the AF tune bias, a bias which is constant but is made variable when one starts performing PDAF.

Feb 03, 2013 at 05:18 PM

scalesusa
Offline
• • • •
Upload & Sell: Off

p.1 #20 · p.1 #20 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

The issue with AFMA is that the variations in the lens and camera for PDAF and Contrast Detect are high.

Contrast detect is not very accurate, and that green dot means very little.

You need to take about 10 shots setting the focus to mfd or infinity each time and throw out the outlier focus errors and take the average of the rest.

The closer you look at AF accuracy, the uglier it gets.

Feb 03, 2013 at 06:30 PM

1 2 3 … 37 38 end
FM Forums \| Canon Forum \| Join Upload & Sell 1 2 3 … 37 38 end

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.