Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Home · Register · Search · Software · Hosting · Software · Join Upload & Sell

FM Forum Rules
Canon SLRs, primes, and zooms lenses reviews

FM Forums \| Canon-mount SLRs \| Join Upload & Sell
1 2 3 … 6 7 8 … 15 16 end

snapsy
Offline
• • • •
Upload & Sell: On

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

zacheryjensen wrote:
I gave this method a try with my Nikkor 85mm ƒ/1.8G. I already have it tuned so I thought it would be a good test. The resulting range where I got an unwavering digital rangefinder "dot" was nearly 75% of the range of adjustments: 11 to -19.

That hardly gives me a sense of confidence.

That said, the center of that range, -4, is only one off of what I had previously determined to be the correct adjustment of -5.

I had less luck applying this technique to my 70-200/4 zoom. The rangefinder inconsistency seemed greater at some focal lengths than others. In fact, at 200mm, the rangefinder thought every single adjustment, 20 to -20 was in focus… not helpful, especially since test photos showed otherwise.

Anyway, that's just my experience with it.

Thanks for the feedback. Don't be concerned about the width of the AF tune confirmation. My theory is both Canon and Nikon intentionally add some slop into the range of focus that produces a confirmation in the VF - if the range were too tight then manual focusing would demand a lot more time and focus ring dexterity on the part of the photographer. If my theory is true this might explain why the midpoint of the VF conformation range is so reliable for DotTune.

Regarding ranges where either -20 or +20 (or both) are confirmed, there is a workaround I mention in the video but I didn't provide enough detail for it to be useful yet. I'm planning a follow-up video to describe it once I get the procedure down. In the meantime try reducing the DOF as much as you can by moving closer and see if that helps.

Feb 21, 2013 at 03:28 PM

snapsy
Offline
• • • •
Upload & Sell: On

p.7 #2 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Ian.Dobinson wrote:
OK first let me say im liking the idea of this method and when i have a spare couple of hours in the day i will have to give it a shot.
but I also have an idea (with a question) that may add to this method using Canon's softwre (i have no idea about nikons software)

when your tetherered and using Canon's softwre you can bring up the live view picture on screen. and you can also adjust focus using the '<<< < > >>> ' . my question is how do the smaller step <> buttons have any relation to a step in the MA values.?

Thanks Ian. Someone on POTN provided a link to a similar method. The article claims that the smaller step buttons correspond to a single AF tune value. I haven't tried it myself yet. Here's the link.

Feb 21, 2013 at 03:59 PM

Alek Komarnits
Offline
• • •
Upload & Sell: Off

p.7 #3 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

I'm a bit late to the thread, but super-awesome job Snapsy ... as a technical person, I love the reverse engineering you did and WOW, what a simple concept/implementation.

mrfozibear talked about the new TurboCal feature of Reikan Focal Pro - certainly sounds like they are using a similar approach as yours - nice to hear the results are consistent ... which ideally, they should be of course.

I'm a bit busy with some other issues going on, but once Reikan releases the "final" version of their software, I'll try to test it and compare to your method.

Thanks again - great job!

Feb 21, 2013 at 05:26 PM

Ian.Dobinson
Offline
• • • • • •
Upload & Sell: Off

p.7 #4 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
Ian.Dobinson wrote:
OK first let me say im liking the idea of this method and when i have a spare couple of hours in the day i will have to give it a shot.
but I also have an idea (with a question) that may add to this method using Canon's softwre (i have no idea about nikons software)

when your tetherered and using Canon's softwre you can bring up the live view picture on screen. and you can also adjust focus using the '<<< < > >>> ' . my question is how do the smaller step <> buttons have any relation to a step in the MA values.?

Thanks Ian. Someone on POTN provided a link to a similar method. The article claims that the smaller step buttons correspond to a single AF tune value. I haven't tried it myself yet. Here's the link.

Thanks , way back in my distant memory I think someone on here mentioned the small step bein 1 point on the MA value. But that was before I had a MA body so I never tested that theory out.

I guess if it is then its another usfull way .

Feb 21, 2013 at 05:28 PM

clarence3
Offline
• • • •
Upload & Sell: Off

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

Ian.Dobinson wrote:
when your tetherered and using Canon's softwre you can bring up the live view picture on screen. and you can also adjust focus using the '<<< < > >>> '

I have DSLR Controller for Android...
http://play.google.com/store/apps/details?id=eu.chainfire.dslrcontroller&hl=en
http://dslrcontroller.com/

Seems like it'd be fairly straight-forward to write an app to use tethered Live View to step through each of the 40 AFMA steps, note whether focus confirmation was instantly achieved, then calculate the dot-tune midpoint.

0 clicks instead of ~50 clicks for FoCal, plus without the series of manual AFMA setting changes in the menu.

Feb 22, 2013 at 05:15 AM

Zenon Char
Offline
• • •
Upload & Sell: Off

p.7 #6 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
Ian.Dobinson wrote:
OK first let me say im liking the idea of this method and when i have a spare couple of hours in the day i will have to give it a shot.
but I also have an idea (with a question) that may add to this method using Canon's softwre (i have no idea about nikons software)

when your tetherered and using Canon's softwre you can bring up the live view picture on screen. and you can also adjust focus using the '<<< < > >>> ' . my question is how do the smaller step <> buttons have any relation to a step in the MA values.?

Thanks Ian. Someone on POTN provided a link to a similar method. The article claims that the smaller step buttons correspond to a single AF tune value. I haven't tried it myself yet. Here's the link.

It is really not the same I don't think. You have to connect a computer if you own a laptop. You are still judging what MFA is required by trusting your eyes rather than the DotTune method which tells you what is out.

I tried Arash's when it came first came out. I have tried most of them except Focal. I would prefer a system that tells me what is out and what needs to be done to correct it.

Feb 22, 2013 at 05:33 AM

snapsy
Offline
• • • •
Upload & Sell: On

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

The Magic Lantern developers are working on an automated version of DotTune. Someone on CR posted a link to the first source commit of the feature. Based on the source the user is prompted to establish critical focus in LV and exit, after which the ML does the rest. Source Code. I'll be posting to the ML DotTune feature request thread here as horshack.

Feb 22, 2013 at 08:56 AM

Beni
Offline
• • • • •
Upload & Sell: On

p.7 #8 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Question, would not the shot to shot variability in PDAF also apply to whether the green dot shows up? If I get a +/-3 of MA between shots on a camera for example then would I not also get the same +/-3 of Green Dot confirmation?

Feb 22, 2013 at 09:13 AM

snapsy
Offline
• • • •
Upload & Sell: On

p.7 #9 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Beni wrote:
Question, would not the shot to shot variability in PDAF also apply to whether the green dot shows up? If I get a +/-3 of MA between shots on a camera for example then would I not also get the same +/-3 of Green Dot confirmation?

In my testing the shot-to-shot variability appears to stem exclusively from the mechanical elements of the PDAF system (AF motor/servo), at least under good lighting with a good AF target and the camera stabilized.

Feb 22, 2013 at 03:24 PM

Jeff
Offline
• • • • •
Upload & Sell: On

p.7 #10 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Nicely done, Snapsy. This would have come in real handy prior to acquiring the 5DIII...

Feb 22, 2013 at 04:44 PM

Photon
Offline
• • • • •
Upload & Sell: On

p.7 #11 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
The Magic Lantern developers are working on an automated version of DotTune. Someone on CR posted a link to the first source commit of the feature. Based on the source the user is prompted to establish critical focus in LV and exit, after which the ML does the rest. Source Code. I'll be posting to the ML DotTune feature request thread here as horshack.
Wow!
Will you give us a heads up here when that becomes available? Thanks!

Feb 22, 2013 at 04:56 PM

JEFFERY Z71
Offline
• •
Upload & Sell: Off

p.7 #12 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Have to give it a try...thank you!

Feb 22, 2013 at 05:04 PM

Inga
Offline
• • •
Upload & Sell: Off

p.7 #13 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

First off, thanks for the new technique to check/adjust MFA Snapsy.

I have an 85L on my 5DIII and feel like it front focuses in real shooting conditions. I'm not ruling out that my technique is to blame. I have tried both focus-recompose, and using the full range of focus points, and sometimes is bang-on, but often for portraits and headshots etc (static subjects) it seems to front focus. I've tried Focal, and it tells me that lens is spot on, or sometimes +1. I just tried this technique of yours and am getting a -3.5 result averaged out over a couple of tests.

Thoughts?

Feb 23, 2013 at 02:52 AM

snapsy
Offline
• • • •
Upload & Sell: On

p.7 #14 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Inga wrote:
First off, thanks for the new technique to check/adjust MFA Snapsy.

I have an 85L on my 5DIII and feel like it front focuses in real shooting conditions. I'm not ruling out that my technique is to blame. I have tried both focus-recompose, and using the full range of focus points, and sometimes is bang-on, but often for portraits and headshots etc (static subjects) it seems to front focus. I've tried Focal, and it tells me that lens is spot on, or sometimes +1. I just tried this technique of yours and am getting a -3.5 result averaged out over a couple of tests.

Thoughts?

Seems pretty close. Which tune value is yielding the best results for you?

Feb 23, 2013 at 05:23 PM

snapsy
Offline
• • • •
Upload & Sell: On

p.7 #15 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

You need to see this: DotTune on Magic Lantern Demo (YouTube)

Feb 23, 2013 at 10:02 PM

dwweiche
Offline
• • •
Upload & Sell: Off

p.7 #16 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

snapsy wrote:
You need to see this: DotTune on Magic Lantern Demo (YouTube)

Excellent. Signifying the start of the demise of the pay sites and options?

Feb 23, 2013 at 10:12 PM

Matt Howell
Offline
•
Upload & Sell: Off

p.7 #17 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

That is awesome. I'm really hoping for a new Magic Lantern version for the 5D3 with this Dot-Tune feature that is not alpha and does not require a firmware update every time the camera power is turned on.

Feb 23, 2013 at 10:18 PM

wihakowi
Offline
•
Upload & Sell: Off

p.7 #18 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

I used Dot-Tune today for MFA and have a couple of questions ...

Why should I not use AI Servo mode while checking the "Dot"? I am in MF so I can't see any difference between AI Servo and One Shot Mode. Shouldn't I use the mode that I most use with that lens?

I took my initial readings under good light but when the lighting started to lessen I repeated the measurements
just for fun. Interestingly, I got basically the same "median" setting but the outside points were much less of a range. That stands to reason (I think) given that, since it has less light and therefore less contrast, the body can't read the focus as well and gives less confirmation dots on either side of the one that truly is the best focus. Does it make any sense to test under "less than optimum" light? Or is the above all false logic?

Lastly it would seem that, if Dot-Tune is truly showing what it purports to, it proves that, at least in MF confirmation, these cameras are pretty darn inaccurate. Does that explain why I have had less than stellar results when using MF and depending on the confirm Dot to assure proper focus? Of course Dot-Tune does nothing to improve MF accuracy since the range of confirm Dots remains the same!

Feb 24, 2013 at 05:25 AM

Beni
Offline
• • • • •
Upload & Sell: On

p.7 #19 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

Matt Howell wrote:
That is awesome. I'm really hoping for a new Magic Lantern version for the 5D3 with this Dot-Tune feature that is not alpha and does not require a firmware update every time the camera power is turned on.

Yup!

Feb 24, 2013 at 08:51 AM

snapsy
Offline
• • • •
Upload & Sell: On

p.7 #20 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes

wihakowi wrote:
Why should I not use AI Servo mode while checking the "Dot"? I am in MF so I can't see any difference between AI Servo and One Shot Mode. Shouldn't I use the mode that I most use with that lens?

You actually might be able to use AI Servo since the mode may not affect the dynamics of the VF confirmation. I just happened to use One-Shot for all my testing and so didn't want to introduce an unknown variable into the procedure.

wihakowi wrote:
I took my initial readings under good light but when the lighting started to lessen I repeated the measurements
just for fun. Interestingly, I got basically the same "median" setting but the outside points were much less of a range. That stands to reason (I think) given that, since it has less light and therefore less contrast, the body can't read the focus as well and gives less confirmation dots on either side of the one that truly is the best focus. Does it make any sense to test under "less than optimum" light? Or is the above all false logic?

The light levels, temperatures, and subject contrast definitely affect the width of the confirmed range. In my testing the range width increases as target quality/light decreases.

wihakowi wrote:
Lastly it would seem that, if Dot-Tune is truly showing what it purports to, it proves that, at least in MF confirmation, these cameras are pretty darn inaccurate. Does that explain why I have had less than stellar results when using MF and depending on the confirm Dot to assure proper focus? Of course Dot-Tune does nothing to improve MF accuracy since the range of confirm Dots remains the same!

The width of the confirmation range is actually not an issue for DotTune since it only cares about the midpoint, which correlates to the maximum phase correlation and thus best focus. I actually think Canon/Nikon might intentionally be making the range wider than it needs to be to aid in faster manual focusing - if the range were too tight then some photographers might have trouble in dynamic shooting situations. For your MF lenses I suggest DotTunning them, and then whenever you shoot just try to pick the center of the VF confirmation range when focusing.

Feb 24, 2013 at 03:16 PM

1 2 3 … 6 7 8 … 15 16 end
FM Forums \| Canon-mount SLRs \| Join Upload & Sell 1 2 3 … 6 7 8 … 15 16 end