December 20, 2014, 09:12:49 PM

Author Topic: New MFA method  (Read 23389 times)

AlanF

  • 1D X
  • *******
  • Posts: 1176
    • View Profile
Re: New MFA method
« Reply #75 on: February 17, 2013, 05:52:23 PM »
Larry
Comsense is not saying he doesn't want to have precise focus. He is complaining that Horshack hasn't explained clearly what he  meant in his reply to my earlier message, Horshack also confused me, and he hasn't replied to my last questions to him, four or five posts back.
5D III, EOS-M, Powershot SX50, 300/2.8 II, 1.4xTC III, 2xTC III, 70-200/4 IS, 24-105, 15-85, 100-400 II, Sigma 10-20, EOS-M, 18-55, f/2 22.

canon rumors FORUM

Re: New MFA method
« Reply #75 on: February 17, 2013, 05:52:23 PM »

horshack

  • PowerShot G1 X II
  • ***
  • Posts: 60
    • View Profile
Re: New MFA method
« Reply #76 on: February 17, 2013, 06:38:29 PM »
If you don't want ability to focus to a precise point within the DOF, why would you bother to AFMA? I am amateur here with fewer shutter clicks than most but I  have about 10 Canon lenses, mostly primes. And none of them would need AFMA if I did not want that precise a focus. Maybe you are missing something in terms of explanation!!!
DOF is defined as the range of distances between the camera and subject that will be rendered with sufficient sharpness for a given viewing or print size. Since DOF is not a precise point but instead a range, when you tune to a 2-D object you can't be sure what the balance of front/rear DOF will be relative to the 2-D point you're tuning to. This is fine if you'll be shooting with enough DOF for everything you need to be sharp, or if you're shooting 2-D real-life subjects. However if you're shooting in DOF-limited situations (small DOF, like f/1.4), many times you'll want to know or specify the precise balance of DOF relative to where the AF is tuned to, for example when shooting a portrait where you want both the eyes and nose to be in focus vs eyes and ears or eyes and hiar.

horshack

  • PowerShot G1 X II
  • ***
  • Posts: 60
    • View Profile
Re: New MFA method
« Reply #77 on: February 17, 2013, 06:40:16 PM »
The "0" of my home grown system is placed in the same plane as the target. The ruler rests at about 30deg at the side of a cardboard box and the target is pasted on the front face.  The line on the ruler that touches the target is the 0 point.  The target is perpendicular to the line of sight from the camera.  What is not working is the dot assist on manual focussing relative to focussing in live view.  Does anyone use manual focussing and the dot assist regularly - the spread of 15-20 units is not impressive? Maybe my mistake has been to use the automatic focussing in live view and not tether to a computer and blow up for manual focussing in live view?
Again, precise use of a 3-D target to establish the DOF balance requires tight tolerances between the 2-D target and the 3-D portion (ruler), along with a mechanism to calibrate alignment between the camera and target.

comsense

  • Rebel T5i
  • ****
  • Posts: 106
    • View Profile
Re: New MFA method
« Reply #78 on: February 18, 2013, 03:30:06 AM »
If you don't want ability to focus to a precise point within the DOF, why would you bother to AFMA? I am amateur here with fewer shutter clicks than most but I  have about 10 Canon lenses, mostly primes. And none of them would need AFMA if I did not want that precise a focus. Maybe you are missing something in terms of explanation!!!
DOF is defined as the range of distances between the camera and subject that will be rendered with sufficient sharpness for a given viewing or print size. Since DOF is not a precise point but instead a range, when you tune to a 2-D object you can't be sure what the balance of front/rear DOF will be relative to the 2-D point you're tuning to. This is fine if you'll be shooting with enough DOF for everything you need to be sharp, or if you're shooting 2-D real-life subjects. However if you're shooting in DOF-limited situations (small DOF, like f/1.4), many times you'll want to know or specify the precise balance of DOF relative to where the AF is tuned to, for example when shooting a portrait where you want both the eyes and nose to be in focus vs eyes and ears or eyes and hiar.
I understand DOF. For good or worse, I have physics/spectroscopy background. What you have still not explained is why LensAlign target is so critical compared to 2-D target with precisely placed ruler? On practical note, you need AFMA only when you are using larger apertures/ small DOF except for rare manufacturing/ quality control failures. And front/back focusing you refer to as 'balance' is actually part of performing AFMA. So, essentially you are saying that one needs to find front/back focus with LensAlign target to use your precise dot tune? What I also fail to understand is that if you get that far, it does not take a lot more to do AFMA. So why bother with dot tune? If i have to spend money/time why not on something that works for sure? Again, neither against nor for; just trying to understand.

sama

  • EOS M2
  • ****
  • Posts: 181
    • View Profile
Re: New MFA method
« Reply #79 on: February 18, 2013, 11:06:55 PM »
see another post in this forum with video tutorial : http://www.canonrumors.com/forum/index.php?topic=13056.msg234394;topicseen#new

East Wind Photography

  • 1D Mark IV
  • ******
  • Posts: 879
  • EWP
    • View Profile
Re: New MFA method
« Reply #80 on: February 18, 2013, 11:31:33 PM »
Still doesnt work even with a video!  ;D

see another post in this forum with video tutorial : http://www.canonrumors.com/forum/index.php?topic=13056.msg234394;topicseen#new

AlanF

  • 1D X
  • *******
  • Posts: 1176
    • View Profile
Re: New MFA method
« Reply #81 on: February 19, 2013, 11:41:46 PM »
I have suggested why DotTune fails in the DotTune Video thread.  Maybe failure means the camera should be sent back to Canon for calibration. 
5D III, EOS-M, Powershot SX50, 300/2.8 II, 1.4xTC III, 2xTC III, 70-200/4 IS, 24-105, 15-85, 100-400 II, Sigma 10-20, EOS-M, 18-55, f/2 22.

canon rumors FORUM

Re: New MFA method
« Reply #81 on: February 19, 2013, 11:41:46 PM »

kaihp

  • Canon 70D
  • ****
  • Posts: 302
    • View Profile
Re: New MFA method
« Reply #82 on: February 20, 2013, 12:02:47 AM »
I wonder if Focal Method "TurboCal" is using similar technique as "DotTune" to perform MFA prediction in 30 Seconds  with no shutter activation in an automated way.

http://www.reikan.co.uk/focalweb/index.php/2013/02/focal-1-8-release-candidate-available-for-focal-pro-users/

Thanks for the link! I agree, they might be doing something similar to DotTune, provided the Canon SDK+USB interface provides a way to both change the AF tune value programmatically and also feed the PDAF evaluation information over USB to the computer as well.
FWIW, I forwarded a link to this discussion to Reikan some days ago. I have no idea if their method is the same nor if they have acted independently or not.

East Wind Photography

  • 1D Mark IV
  • ******
  • Posts: 879
  • EWP
    • View Profile
Re: New MFA method
« Reply #83 on: February 20, 2013, 12:35:34 AM »
Yeah I replied to you on that.  I think there are more factors that can throw off the Dot tune method including non-symmetric AF detect.  Different lenses have different in-focus characteristics and while not a defect causes the AF detect to respond in a non-symmetric way.  Some lenses have more DOF behind that subject than in front, other lenses just focus differently in front or behind the subject.  ex: front focus yields good blur...back focus starts producing halo's before it begins to blur.  AF detect may get confused in that scenario.

I don't believe it's a good way to tell if your camera is out of spec and I still believe it's a bad way to calculate AFMA due to the many factors that can throw it off.  YOu still need to verify with a 3D calibration target or something similar and at that point you might as well just do it manually.

I have suggested why DotTune fails in the DotTune Video thread.  Maybe failure means the camera should be sent back to Canon for calibration.

comsense

  • Rebel T5i
  • ****
  • Posts: 106
    • View Profile
Re: New MFA method
« Reply #84 on: February 20, 2013, 11:29:57 AM »
Yeah I replied to you on that.  I think there are more factors that can throw off the Dot tune method including non-symmetric AF detect.  Different lenses have different in-focus characteristics and while not a defect causes the AF detect to respond in a non-symmetric way.  Some lenses have more DOF behind that subject than in front, other lenses just focus differently in front or behind the subject.  ex: front focus yields good blur...back focus starts producing halo's before it begins to blur.  AF detect may get confused in that scenario.

I don't believe it's a good way to tell if your camera is out of spec and I still believe it's a bad way to calculate AFMA due to the many factors that can throw it off.  YOu still need to verify with a 3D calibration target or something similar and at that point you might as well just do it manually.

I have suggested why DotTune fails in the DotTune Video thread.  Maybe failure means the camera should be sent back to Canon for calibration.
+1 exactly the point I made in my earlier thread. Front/back focussing is one of the common reason to perform AFMA and with 3D target, your job is mostly done anyways....

Knut Skywalker

  • PowerShot G1 X II
  • ***
  • Posts: 65
    • View Profile
Re: New MFA method
« Reply #85 on: February 20, 2013, 12:30:57 PM »
Just found out my 50mm 1.4 is -11 on my 5D Mark II.  :o The difference is mind-blowing. Thank you so much for this. :)
5D Mark II + Grip
50mm 1.4 and 100mm 2.0
430EX II and YN-565EX + 4 YN622c

digital paradise

  • Canon 70D
  • ****
  • Posts: 290
    • View Profile
    • Zenon Char Photography
Re: New MFA method
« Reply #86 on: February 20, 2013, 12:53:46 PM »
Just found out my 50mm 1.4 is -11 on my 5D Mark II.  :o The difference is mind-blowing. Thank you so much for this. :)

So there is a noticeable difference in your shots now or you just providing the values after testing with DotTune?   

Knut Skywalker

  • PowerShot G1 X II
  • ***
  • Posts: 65
    • View Profile
Re: New MFA method
« Reply #87 on: February 20, 2013, 01:02:16 PM »
The photos with the 50mm 1.4 are significantly sharper now that i found its "sweet spot". I knew that the 50mm 1.4 wasn't the sharpest lens wide open and i thought this would be normal, but now the photos have a good sharpness even wide open. :)
5D Mark II + Grip
50mm 1.4 and 100mm 2.0
430EX II and YN-565EX + 4 YN622c

canon rumors FORUM

Re: New MFA method
« Reply #87 on: February 20, 2013, 01:02:16 PM »

digital paradise

  • Canon 70D
  • ****
  • Posts: 290
    • View Profile
    • Zenon Char Photography
Re: New MFA method
« Reply #88 on: February 20, 2013, 01:12:28 PM »
The photos with the 50mm 1.4 are significantly sharper now that i found its "sweet spot". I knew that the 50mm 1.4 wasn't the sharpest lens wide open and i thought this would be normal, but now the photos have a good sharpness even wide open. :)

That is good to hear. The author is getting some heat on this method here and I don't really know (and I have absolutely no interest) the technical aspects as to why or why not it would not work. All I need to know is that it works for me. I've been busy but I want to try it on one my suspect lenses with the new focus acquisition method.   

It is getting good reviews on other sites. Fred Miranda is a pretty respected photographer and he liked it.           

East Wind Photography

  • 1D Mark IV
  • ******
  • Posts: 879
  • EWP
    • View Profile
Re: New MFA method
« Reply #89 on: February 20, 2013, 01:52:27 PM »
So lets start out by saying that a 50mm 1.4 has DOF at 12ft of about 17" assuming a full frame body.  you could be within +-3 or 4 AFMA and still get pretty sharp images.  Dot method might have gotten you there or it might have just got you close within the acceptable DOF.  Either way for you that would be good because your images are sharper.

Where the method breaks down is when we start using long telelphotos where exact AFMA is more critical such as when your DOF is 6" or less wide open.  Off by 1 AFMA will give you soft images with an occasional in focus shot.

As I described earlier every lens has it's own unique characteristics too and if you are looking for exactness and not just getting you close enough, then you need to use a 3D target such as a lensalign or some other targeting and measuring system to, if not make the proper AFMA selection, but to at least verify that you have selected the best AFMA for your lens/camera combination.

The photos with the 50mm 1.4 are significantly sharper now that i found its "sweet spot". I knew that the 50mm 1.4 wasn't the sharpest lens wide open and i thought this would be normal, but now the photos have a good sharpness even wide open. :)

canon rumors FORUM

Re: New MFA method
« Reply #89 on: February 20, 2013, 01:52:27 PM »