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

Moderated by: Fred Miranda
Username   Password

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

1      
2
       3              23       24       end
  

Dot-Tune: Autofocus Fine Tuning in under 5 minutes
  
 
snapsy
Offline
• • • •
Upload & Sell: On
p.2 #1 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


scalesusa wrote:
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.


There are two separate forces at work here; the shot-to-shot variability in the mechanical aspects of the focus system (lens motor, servo, position sensors), and a fixed error bias in the optical aspects of that same system (mirror alignment, mount alignment, AF point alignment, lens abberations, etc..). The only reason existing AFMA techniques require multiple shots to find the fixed error bias is because those techniques rely on the variable output of PDAF focusing, which makes it harder to discern the fixed error bias from the mechanical variability. This is what the technique in the OP avoids.

An analogy is someone trying to find their true weight on a bathroom scale. The person knows his weight varies throughout the day, so he endeavors to weight himself multiple times and average the results. Lets say his weight varies by +/- 3 pounds, with a low of 160lbs and a high of 166lbs. So he concludes his weight is 163lbs. But what happens if the scale he used is out of calibration, with a fixed error of 10lbs under? That means his real range for the day was 170lbs to 176lbs, making his true weight 173lbs instead of 163lbs. Because of the scale's error, the person will not only fail to arrive at the correct average weight, he'll also fail in every single sample of his weighing, since the heaviest the scale will report is 166lbs, yet his actual range is 170lbs to 176lbs.

The variability of the man's weight represents the shot-to-shot mechanical variability std. dev of the PDAF system. The fixed error of the scale represents the fixed error the PD sensing mechanism resulting from optical/machined tolerances of the mount, mirror, AF sensor, and lens-specific abberations not accounted for in the PDAF design.

Now what happens if the man knows the scale is off but wants to find out by how much? And let's say he already knows his true (average) weight is 173lbs, but doesn't yet know that his range of weights throughout the day is 6lbs (+/- 3lbs). He can't just stand on the scale once and calculate what the error of the scale is relative to his true (average) weight, since his weight varies by an amount unknown to him. So instead he has to take measurements throughout the day to establish his range of weight. Only then can he compare his known true (average) weight to what the scale's biased average is, at which time he can calculate the scale's error at 10lbs. This is the same situation of trying to calculate an AF tune value by taking multiple photographs.

The purpose of AF tune is to center the range of shot-to-shot variation over the system's true center (the "0" on the bathroom scale). That way even with the unavoidable variation most of the shots will still be in focus. I've made a graphical depiction of this here.



Feb 04, 2013 at 01:40 AM
badlydrawnboy
Offline
• • •
Upload & Sell: On
p.2 #2 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


I'm excited to try this, in large part because I don't get consistent results with any other AF fine-tuning method (FoCal or FocusTune).

Two questions:

1. When you say "establish critical focus using LiveView" in step 1, does that mean manually or using CDAF? I've actually found it easier to nail focus in LiveView with AF than manual. Maybe my eyes are bad.

2. What focal distance do you recommend doing this at?



Feb 04, 2013 at 01:52 AM
sirimiri
Online
• • • •
Upload & Sell: On
p.2 #3 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


*following*


Feb 04, 2013 at 03:39 AM
gschlact
Offline
• • •
Upload & Sell: On
p.2 #4 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Such a simple yet apparently effective process. NOW, why can't Can't Canon just make it into a menu routine? ( would be too hard to disable the lens command to emulate the MF during that part of the process. Too much to ask?

Guy
Gschlact



Feb 04, 2013 at 04:01 AM
snapsy
Offline
• • • •
Upload & Sell: On
p.2 #5 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


badlydrawnboy wrote:
I'm excited to try this, in large part because I don't get consistent results with any other AF fine-tuning method (FoCal or FocusTune).

Two questions:

1. When you say "establish critical focus using LiveView" in step 1, does that mean manually or using CDAF? I've actually found it easier to nail focus in LiveView with AF than manual. Maybe my eyes are bad.

2. What focal distance do you recommend doing this at?


For step #1 you can use either CDAF or manual focus or a combination of the two. It's really up to your personal preference. This is why I put the MF step as #2 even though some may prefer to use MF in step #1.

As for focus distance, I have preferences that have worked well for me on the lenses I use, but I don't have a grasp of all the factors involved to give general advice. For my primes I like to tune at far distances, well beyond the normal 25x to 50x focal length multipliers, and find that those tune values hold up for short distances as well. I'd suggest first trying the tune technique at any arbitrary focus distance, just to get a feel for it, and to confirm that it produces an AF tune value that yields good focus results at that tuned distance. Once you establish that you can then start experimenting with different focus distances to find how those differences might affect the AF tune bias.

Lastly, the quality of the AF target you use is important. I like this one from Bob Atkin's site (focus on the intersection of the horiztonal and vertical lines). You also want a good amount of light. The better the target's contrast/detail and lighting, the easier it is for the PD sensing to discern in-phase vs out-of-phase, which will make the green-dot confirmation process at the margins more reliable and less finicky. Also, if you plan to use a lens in different lighting temperatures, I suggest trying the technique in each of those lighting scenarios to see how the results change - I'm interested in hearing feedback from anyone who does this.



Feb 04, 2013 at 04:29 AM
Mike K
Offline
• • • •
Upload & Sell: Off
p.2 #6 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


snapsy wrote:
An analogy is someone trying to find their true weight on a bathroom scale. The person knows his weight varies throughout the day, so he endeavors to weight himself multiple times and average the results. Lets say his weight varies by +/- 3 pounds, with a low of 160lbs and a high of 166lbs. So he concludes his weight is 163lbs.


Contrast detection AF is actually more accurate than PD AF as there is no issue with calibration (offset). Rather, the actual lens/camera body combination under consideration simply keeps adjusting AF until contrast is at a maximum. If you are having trouble with CDAF Pecision (reproducible final settings) one should consider a target with good contrast, in good light, and a solid tripod.

Your body weight analogy is fine, as the difference between the true value Vs measured value is a measure of accuracy. the +/- 3 lb issue defines the precision of measurement. As your example clearly delineates, the precision and accuracy can be distinctly different from one another.
Both of us use Live View (or magnified LV) assessment as a measure of accuracy. This works well on the more modern ~1 mdot sensors (since the 5DII and later). This is the key as comparisons can be quickly made without taking pics, downloading them, etc.

The only difference between our methods is that you are using focus confirm to indirectly assess PDAF while I press the back button auto focus to compare the AF setting against magnified LV. Since we are calibrating the PDAF system I would prefer to use that entire hardware chain in the calibration. For example there may be a different result when AF from MFD Vs infinity. A handful of replicate comparisons can be performed within a minute or so, thus I do not see the added imprecision, if any, to be onerous.
Mike K



Feb 04, 2013 at 05:13 AM
Andrew J
Offline
• • • •
Upload & Sell: On
p.2 #7 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Very interesting, thanks for posting. I am concerned that this method will not address focus shift. When my 600/f4 is stopped down 2 clicks the focus moves +5. My 500/f4 also moved 5 points. FoCal confirmed this and knowing the new number and using it is a big help.


Feb 04, 2013 at 05:36 AM
snapsy
Offline
• • • •
Upload & Sell: On
p.2 #8 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Mike K wrote:
The only difference between our methods is that you are using focus confirm to indirectly assess PDAF while I press the back button auto focus to compare the AF setting against magnified LV. Since we are calibrating the PDAF system I would prefer to use that entire hardware chain in the calibration. For example there may be a different result when AF from MFD Vs infinity. A handful of replicate comparisons can be performed within a minute or so, thus I do not see the added imprecision, if any, to be onerous.
Mike K


Mike, I think we're nearly on the same page. My only point of departure is that since the AF tune bias is solely the result of static tolerances which are completely divorced from the PDAF mechanical side, I don't see the utility in incorporating an entire hardware chain into the process used to derive the AF tune value. I'll still use PDAF to as a sanity check once I've completed the process and arrived at my final AF tune value, but using it in the AF tune process itself for me only serves to introduce variability into the portion of the system that otherwise has none.



Feb 04, 2013 at 05:44 AM
snapsy
Offline
• • • •
Upload & Sell: On
p.2 #9 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Andrew J wrote:
Very interesting, thanks for posting. I am concerned that this method will not address focus shift. When my 600/f4 is stopped down 2 clicks the focus moves +5. My 500/f4 also moved 5 points. FoCal confirmed this and knowing the new number and using it is a big help.


The only way a single AF tune value can account for focus shift is if the PDAF system has a database of lenses with a list of spherical abberation vectors based on various apertures, which could be used to dynamically alter the AF bias applied depending on the photographer's selected aperture. Absent that system we can only tune for either wide-open or stopped down (or a compromise in between). If you'd like to tune for stopped down using the new AF tune technique, for step #1 dial in your stopped-down aperture and hold the DOF preview button while establishing focus in Live View. The AF tune value you arrive at using the remainder of the technique will then have focus-shift correction built into it.



Feb 04, 2013 at 05:51 AM
JameelH
Offline
• • •
Upload & Sell: On
p.2 #10 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Sounds interesting. I'm going to try this with one of my lenses tomorrow.


Feb 04, 2013 at 06:35 AM
 

Search in Used Dept. 



paulfeng
Offline
• • •
Upload & Sell: Off
p.2 #11 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


snapsy, thank you for this contribution to the body of knowledge - I'm going to try it out, just need to make some time...


Feb 04, 2013 at 07:05 AM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #12 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


skibum5 wrote:
I will give it a try, but NOT today. Today is the SUPERBOWL!!!!!!!!!


I will give it a try, but not for a few months. Today was the superbowl.
Almost the greatest ever. So much shaky play and still the comeback of all comebacks and then no runs, four silly passes, game lost.

maybe in a few years I will be recovered and give this AF tuning technique a test

(but all the other teams better watch out, Kaepman is real)



Feb 04, 2013 at 07:34 AM
GammyKnee
Offline

Upload & Sell: Off
p.2 #13 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Snapsy - thanks for taking the time to post this up and explain it so clearly. I love the concept I'm certainly going to give it a try once I get the time.


Feb 04, 2013 at 09:10 AM
Beni
Offline
• • • • •
Upload & Sell: On
p.2 #14 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Following! I have a 5D3 and 50L coming and was about to buy Focal but perhaps I'll wait...


Feb 04, 2013 at 10:34 AM
Zenon Char
Offline
• • •
Upload & Sell: Off
p.2 #15 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


badlydrawnboy wrote:
I'm excited to try this, in large part because I don't get consistent results with any other AF fine-tuning method (FoCal or FocusTune).

Two questions:

1. When you say "establish critical focus using LiveView" in step 1, does that mean manually or using CDAF? I've actually found it easier to nail focus in LiveView with AF than manual. Maybe my eyes are bad.

2. What focal distance do you recommend doing this at?


I thought I was the only one.



Feb 04, 2013 at 03:12 PM
Zenon Char
Offline
• • •
Upload & Sell: Off
p.2 #16 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Well now that I have had a chance to read this a few times it sounds very interesting. Only if I could dig up a few old threads at DPreview. I said long time ago of LV is accurate why can't it be used as a base for MFA. Very cool idea
snapsy.

I've never been a real supporter of MFA and I'm blacklisted at Dpreview for telling it like it is. I've never had consistent results. I could MFA one day, repeat it the next day and get different results. I have tried at 8 10 and 12' and get different results. Even the users manual states for best results to MFA at location so distance must be important. So basically you set it at the distance you shoot at most often.

I basically gave up on MFA and if I think something is off my lens goes to Canon to get calibrated by a trained technician that has all the tools to do this. A pain but it is worth it. I realize there is slight variation in the Canon specs so MFA will fine tune for example -2 to +2.

The other thing that I found odd was yes Canon does provide a procedure but you have to look for it. They give you the tool but no procedure in the manual to use it correctly. We all know there must be 2 dozen procedures out there created by owners. Is it only me that finds again a little off as well. If Canon had a tried and true method it would be in the manual. I know at people are happy with the results of other methods so what do I know.

Actually at one time I suggested why can't the firmware use LV to MFA. You are on location, you press the external MFA button and it does a LV to phase detect comparison and sets it automatically. How about at every shot?

Anyway enough of my little rant. This method sounds pretty good. I'm always open to something new if it makes sense. I'll give it a go. Again good work.



Feb 04, 2013 at 03:35 PM
Zenon Char
Offline
• • •
Upload & Sell: Off
p.2 #17 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


One more thing. Chuck Westfall suggests a procedure but a the end states "if you can do better go for it" ?? We're working with expensive precision equipment. You tell me how I get there consistently - not have me trying to figure it out.


Feb 04, 2013 at 03:39 PM
Zenon Char
Offline
• • •
Upload & Sell: Off
p.2 #18 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Mike K wrote:
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 20-40X the focal distance of the lens. Thus for a 50mm lens this is
...Show more

I thought Canon suggested to test 50X the focal length of the lens?



Feb 04, 2013 at 04:30 PM
Zenon Char
Offline
• • •
Upload & Sell: Off
p.2 #19 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Well I just did a quick test on my 5D3 and 24-70II wide open at 70mm and got a +2. Seemed pretty consistent between +2 and and +3, more so at +2. The hard part os determine where to cut off is. Sometimes at +18 you can press the shutter 3 times and will light up. Sometimes it lights up as you are taking your finger off the shutter. Takes a bit of work to find that sweet spot where it does not light up a both ends.

I must admit it is a very quick method. Beats taking the card out to view on the monitor. I guess making sure that the sensor is parallel to the target still applies?




Feb 04, 2013 at 04:38 PM
Mike K
Offline
• • • •
Upload & Sell: Off
p.2 #20 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes


Zenon Char wrote:
I thought Canon suggested to test 50X the focal length of the lens?


I stand corrected, 50X focal distance of the lens under test.
That would yield 2.5 m for a 50mm lens.

Here is another interference pattern one can use for testing AF
http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html
Mike K



Feb 04, 2013 at 07:03 PM
1      
2
       3              23       24       end




FM Forums | Canon Forum | Join Upload & Sell

1      
2
       3              23       24       end
    
 

You are not logged in. Login or Register

Username   Password    Retrive password