DotTune MFA question...

[ErikH]

Hi - I just MFA my 70D using the "DotTune" method outside on a bright day with 2 of my lenses. Here are my results and my questions below:

55-250 STM

(W)ide -8 +5 = -1 (or -2)

(T)ele -5 +3 = -1

18-135 STM

W -6 +5 = 0

T -7 + 5 = -1

1) Because my adjustments are so minimal, is it even worth making the change? The reason I'm asking, is (if I'm understanding correctly) this method balances the focal plane evenly whereas typically the focal plane is thought to be 1/3 in front & 2/3 behind (speaking generally).

Since my camera has a (slight) tendency toward the minus side, could that be part of the correction DotTune is trying to make to even "things" out? (hopefully that makes sense).

Any comments appreciated!

 

[Zeee]

My 300 F4 – minimal focusing distance is 4.9 ft

50 X 300 = 0.01 feet

???

0.01 ft or 0.12 inches total DOF

0.12 divided by 8 = 0.015 inches

0.12 Split front and back = 0.06 inches

0.015 X 4 = 0.06 . Still a shift of +- 4

I get that, but, what does that have to do with 300X50?

Sorry. That was a typo on my part. It should have been 4.9 x 300. It is still correct as I used the value of 0.01 for the rest of the calculations.

 

[Zeee]

I redid the 300.

300 F4 – minimal focusing distance is 4.9 ft

4.9 X 300 = 4.82 feet to target

4.82 = 0.01 ft or 0.12 inches total DOF

0.12 divided by 8 = 0.015 inches

0.12 Split front and back = 0.06 inches

0.015 X 4 = 0.06 . Still a shift of +- 4

f1.2 @ 25 ft

25 X 50mm = 4.1 feet to target

4.1 = 0.09 ft or 1.08 inches total DOF

1.08 divided by 8 = .135 inches

Split front and back = .54 inches or just over ½ inch

.135 X 4 = .54 inches. Still a shift of +- 4.

 

[WilbaW]

I redid the 300.

300 F4 – minimal focusing distance is 4.9 ft

4.9 X 300 = 4.82 feet to target

4.82 = 0.01 ft or 0.12 inches total DOF

0.12 divided by 8 = 0.015 inches

0.12 Split front and back = 0.06 inches

0.015 X 4 = 0.06 . Still a shift of +- 4

f1.2 @ 25 ft

25 X 50mm = 4.1 feet to target

4.1 = 0.09 ft or 1.08 inches total DOF

1.08 divided by 8 = .135 inches

Split front and back = .54 inches or just over ½ inch

.135 X 4 = .54 inches. Still a shift of +- 4.

I'm baffled by what you're trying to do here.

You're dividing the DOF by 8, then dividing 1/2 the DOF by 4, and saying you get the same answer... of course you always will. What are you actually trying to calculate?

 

[Zeee]

I redid the 300.

300 F4 – minimal focusing distance is 4.9 ft

4.9 X 300 = 4.82 feet to target

4.82 = 0.01 ft or 0.12 inches total DOF

0.12 divided by 8 = 0.015 inches

0.12 Split front and back = 0.06 inches

0.015 X 4 = 0.06 . Still a shift of +- 4

f1.2 @ 25 ft

25 X 50mm = 4.1 feet to target

4.1 = 0.09 ft or 1.08 inches total DOF

1.08 divided by 8 = .135 inches

Split front and back = .54 inches or just over ½ inch

.135 X 4 = .54 inches. Still a shift of +- 4.

I'm baffled by what you're trying to do here.

You're dividing the DOF by 8, then dividing 1/2 the DOF by 4, and saying you get the same answer... of course you always will. What are you actually trying to calculate?

First dividing by 8. Then in half to show what is on the + side of zero and then what is on the - side of zero.

But that is fine. Lets not divide at all.

f1.2 @ 25 ft

25 X 50mm = 4.1 feet to target

4.1 = 0.09 ft or 1.08 inches total DOF

1.08 divided by 8 = .135 inches

There is a shift of 8 MFA points one way or the other of 1.35 inches within the DOF zone that is in focus based on the online DOF chart. So is one move of 1.35 going to show that much of a difference if both the lens and camera are very close to factory calibration. f1.2 probably. f8 at 50 feet probably not. I realize it starts to blur at the edges but it does not blur all the way through.

The ⅛ of DOF was presented and I'm just trying to find out what the value of DOF is. It never made any sense to me before so I just brushed it off. Without a value it is like asking how long is a piece of string. Today I'm curious about it.

Also I was surprised as you said you always get the 8 values by dividing the DOF total but based on distance, f stop, and mm they are always different. So how does this equate to the 50X canon recommends and others who MFA at MFD successfully?

As I said it appears the further you are from the target the less forgiving this process is. I'm just exploring this because I don't really like to MFA because of the distances and I'm not nailing targets to trees and shooting out my window.

No one has told me of my assumption that DOF in that equation is the total DOF based on the DOF chart or is it another value yet.

So in a nut shell I'm trying to find a more practical way check 400mm with with a 1.4 TC and I'm trying to figure out if both the lens and camera are very close to factory calibration will you see a difference with a 1 MFA shift at f8 for example shooting at 50 away.

 

[WilbaW]

f1.2 @ 25 ft

25 X 50mm = 4.1 feet to target

You're testing a 50mm f/1.2 lens at 25f. Okay.

4.1 = 0.09 ft or 1.08 inches total DOF

1.08 divided by 8 = .135 inches

So one click of AF MA should shift the auto focus 0.135". Bonza.

There is a shift of 8 MFA points one way or the other of 1.35 0.135 inches within the DOF zone that is in focus based on the online DOF chart.

Let's not forget, the inch of total DOF is not "in focus". It is, in theory, not noticeably out of focus to someone with ordinary eyesight viewing an image under the conditions assumed in the calculation. Those assumptions don't apply for instance if you're peering at the image at 100% on a nice big monitor.

So is one move of 1.35 0.135 going to show that much of a difference if both the lens and camera are very close to factory calibration.

You certainly won't see it under the conditions assumed in the DOF calculation, e.g. a 10x8 viewed from 12". It is still hard to tell a click difference using images or the EOS Utility zoomed in, which is why DotTune approaches the task in a very different way.

I think DotTune is the quickest, easiest, and most portable way of AF tuning, but I don't believe it's the most accurate way. This is how I tune for best accuracy, comparing images deliberately defocussing from the position achieved by the AF system - http://www.dpreview.com/forums/thread/3840249

Also I was surprised as you said you always get the 8 values by dividing the DOF total but based on distance, f stop, and mm they are always different.

I don't think it was me said that.

So how does this equate to the 50X canon recommends and others who MFA at MFD successfully?

I think there has to be a standard scenario for the 1/8 DOF idea to make sense (e.g. wide-open at 50f), but I'm not convinced that would be consistent for all lenses, especially those made by other manufacturers. But in the end does it matter whether it's 1/8 or 1/5 or 1/nth of the DOF? If you have a way of getting the plane of focus on the subject, nothing else matters.

So in a nut shell I'm trying to find a more practical way check 400mm with with a 1.4 TC and I'm trying to figure out if both the lens and camera are very close to factory calibration will you see a difference with a 1 MFA shift at f8 for example shooting at 50 away.

As with most of these questions, the easiest way to answer it is to try it! Don't get hung up on trying find out what the answer should be, just find out what the answer is.

 

[guy schlacter]

Each +1 or -1 is 1/8th of the DOF.

@55mm f/5.6 10ft: the DOF is 2.15ft ==> -1 == 3 inches off

@250mm f/5.6 10ft: the DOF is 1.2 inch ==> -1 == 0.15 inches off

It does make the difference between a sharp photo and an almost sharp photo (to me it does)

 

[guy schlacter]

Unless there is another answer to the DOF value the calculations based on a perfectly calibrated lens and body, no deviation from perfect 0, which is impossible, you would need to shift + or - 4 to see any difference.

The entire depth of field is not in equally sharp focus. There is still a focus point within the DOF.

That is interesting. You can see that on a ruler test but you can't change the focus point independently of the DOF area.

I bet a lot of the DOF is equally in focus here.

There is actually very complex math for dof calculation with differing formula for what is considered near field and far field. I don't pretend to be able to explain it. However, by definition and related to the circle of confusion, the entire dof is in focus, and the single point is rendered as a single point, not oblong. Bokeh has similarly complex math.

 

[guy schlacter]

So, wrt why are small values of mfa important boils down to more than just the focus plan being in Focus. Often photographers are choosing apertures to capture a desired amount of dof, so centering it is important. Another big reason though probably has to do with consistency of focusing the left side exactly putting the focal plane where intended. These phase detect Af mechanisms Do have shot to shot variance, so having he plane centered is important so that when you have shot to shot variance, your subject has best chance of being in the dof.

 

[WilbaW]

There is actually very complex math for dof calculation with differing formula for what is considered near field and far field.

The formulae are complicated, but the physics is just similar triangles.

However, by definition and related to the circle of confusion, the entire dof is in focus, and the single point is rendered as a single point, not oblong.

No, by definition a point in the scene appears as a circle of confusion on the sensor (assuming a perfect lens), and only points in the plane of focus appear as points on the sensor (COC = 0). We arbitrarily define an "acceptable" limit to the COC in the hope that viewers of an image don't notice that elements in front of or behind the plane of focus are actually out of focus.

 

[guy schlacter]

There is actually very complex math for dof calculation with differing formula for what is considered near field and far field.

The formulae are complicated, but the physics is just similar triangles.

However, by definition and related to the circle of confusion, the entire dof is in focus, and the single point is rendered as a single point, not oblong.

No, by definition a point in the scene appears as a circle of confusion on the sensor (assuming a perfect lens), and only points in the plane of focus appear as points on the sensor (COC = 0). We arbitrarily define an "acceptable" limit to the COC in the hope that viewers of an image don't notice that elements in front of or behind the plane of focus are actually out of focus.

Wilba,

I agree with your better explanation about circle of confusion, and you are really saying the same thing, that within DOF, we deem things look sharp.

However, in the first point, I believe you are confusing Field of View which is simple triangle math, with DOF which is much more complex.

a source: http://en.wikipedia.org/wiki/Depth_of_field



DOF ~~ 2N c f^2 s^2

f^4 - (N^2 c^2 s^2)

used for moderate to large distances defined as s>f and s<H. We then substitute s for H.

where we Let f be the lens focal length, N be the lens f-number, and c be the circle of confusion for a given image format. The hyperfocal distance H and S is the distance to subject (focal plane distance).

 

[Zeee]

f1.2 @ 25 ft

25 X 50mm = 4.1 feet to target

You're testing a 50mm f/1.2 lens at 25f. Okay.

4.1 = 0.09 ft or 1.08 inches total DOF

1.08 divided by 8 = .135 inches

So one click of AF MA should shift the auto focus 0.135". Bonza.

There is a shift of 8 MFA points one way or the other of 1.35 0.135 inches within the DOF zone that is in focus based on the online DOF chart.

Let's not forget, the inch of total DOF is not "in focus". It is, in theory, not noticeably out of focus to someone with ordinary eyesight viewing an image under the conditions assumed in the calculation. Those assumptions don't apply for instance if you're peering at the image at 100% on a nice big monitor.

I did say there is a little room for error for that.

So is one move of 1.35 0.135 going to show that much of a difference if both the lens and camera are very close to factory calibration.

You certainly won't see it under the conditions assumed in the DOF calculation, e.g. a 10x8 viewed from 12". It is still hard to tell a click difference using images or the EOS Utility zoomed in, which is why DotTune approaches the task in a very different way.

I think DotTune is the quickest, easiest, and most portable way of AF tuning, but I don't believe it's the most accurate way. This is how I tune for best accuracy, comparing images deliberately defocussing from the position achieved by the AF system - http://www.dpreview.com/forums/thread/3840249

Also I was surprised as you said you always get the 8 values by dividing the DOF total but based on distance, f stop, and mm they are always different.

I don't think it was me said that.

You said that if you divide by 8 then you always get the same number which is true. I was saying that no matter what lens, distance or f stop you would get a total DOF that is within focus. If you divide that by 8 you would get 8 equal values but these values would always be different based on the input.

50mm, f1.2 @ 4 ft = 8 units of .135 inches

300mm, f4 @ 50 ft = 8 units of 0.015 inches

So how does this equate to the 50X canon recommends and others who MFA at MFD successfully?

I think there has to be a standard scenario for the 1/8 DOF idea to make sense (e.g. wide-open at 50f), but I'm not convinced that would be consistent for all lenses, especially those made by other manufacturers. But in the end does it matter whether it's 1/8 or 1/5 or 1/nth of the DOF? If you have a way of getting the plane of focus on the subject, nothing else matters.

So in a nut shell I'm trying to find a more practical way check 400mm with with a 1.4 TC and I'm trying to figure out if both the lens and camera are very close to factory calibration will you see a difference with a 1 MFA shift at f8 for example shooting at 50 away.

As with most of these questions, the easiest way to answer it is to try it! Don't get hung up on trying find out what the answer should be, just find out what the answer is.

Why not? LOL. I'm interested in this stuff. Not trying to give MFA a hard time. I have not learned or discovered anything new in a while.

I have tried them all from the first days of downloading paper scales, to the defocusing one, the one created by the birders where you use Canons lens utility, I have the Lens Align and the program that is similar to Focal. I have Focal and have used Dot Tune. I have at one time spent many hours on this. I don't know what is going on but when the moves are small I get different results on different days and I definitely get different results at different distances to the subject so I started second guessing the results. Mayne that is my mistake. I decided I own $5000 of precision equipment (one lens and camera lol), I'm not a technician and don't have the facilities so I'll let Canon deal with it.

I changed my approach. I shoot real world and do my thing. If it is off (which means it is usually really off) I use Dot Tune or Focal to tell me what is going on and it usually goes to Canon. I have only sent in 2 cameras and 1 lens to Canon for this so far and I have owned all the crops from the 20D, all the 5D's and I'm not sure how many lenses I have bought and sold in the last 10 years.

I did get on answer on another thread. It is Depth of Focus, not Depth of Field so that changes everything. So now I'm trying to figure out what that means.

As usual thanks for your help. Always good info.

 

[Zeee]

I just checked your link and it is the utility one. I'll review everything.

 

[guy schlacter]

MFA values from Canon:

this is clarifying what Chuck said about Field of Focus either in front or behind the focal plane.

http://www.learn.usa.canon.com/resources/articles/2011/af_microadjustment_article.shtml

When setting the Microadjustment, you'll see a scale on the camera's LCD monitor with up to + or - twenty steps. Each step is a very fine increment, equal to 1/8th of the depth-of-field you'd have with the current lens wide-open. And that 1/8th of the depth of field is only moving forward (toward the camera) or back (toward the background) from the sharpest plane of focus. The main thing to remember here is that these are very fine increments.

 

[Zeee]

MFA values from Canon:

this is clarifying what Chuck said about Field of Focus either in front or behind the focal plane.

http://www.learn.usa.canon.com/resources/articles/2011/af_microadjustment_article.shtml

When setting the Microadjustment, you'll see a scale on the camera's LCD monitor with up to + or - twenty steps. Each step is a very fine increment, equal to 1/8th of the depth-of-field you'd have with the current lens wide-open. And that 1/8th of the depth of field is only moving forward (toward the camera) or back (toward the background) from the sharpest plane of focus. The main thing to remember here is that these are very fine increments.

So based in this what is the answer? I know there will be an impact if I make one MFA move on an f1.2 lens at close distance. I will see a significant impact. If whatever MFA device I use tells me I need one move will I see it on an f5.6 lens at 100 feet? I'm trying to prove mathematically you won't see a difference.

If I were Canon I would have never gave a value of ⅛ and just said fine increments. There is still no real value for what ⅛ is.

 

[Zeee]

I had to step out. The reason I'm exploring this because although MFA is a good tool I think it can overused. Just because it is there people feel they have to apply it. It reminds me of the case numbers. Just because they are there people were tweaking the heck out of everything for the simplest shots. Canon even says case 1 is all you may ever need. Sure go out and learn and apply all of that but first keeps things simple and learn how to use the camera.

If you are wondering why I'm pursuing this I have tried to follow complex conversations about how cameras work with formulas, etc that lost me after a while and really never effected me anyway. Canon presents some values so I'm jus trying to understand what they mean and how they effect wide/narrow apertures and short/long distances.

It does make it tougher because I'm not on the MFA bandwagon. My motives are not dark. I have no issues people who use it for everything. Good for them. If we are going to say 1 MFA makes a big difference across the board a person may accept that. If it doesn't and a person who MFA's a lens realizes 1 tick won't make that much of a difference but chooses to leave it or go back to zero that is fine. As long an individual understands that and they can make a choice.

I don't think 1 stop makes a difference across the board for all situations but I can't prove it.

It is not often someone asks why a member is missing. Karl who was respected and disliked was knowledgeable and against the overuse of MFA as well.

 

[Zeee]

that is pretty comprehensive. Lets see what the conclusions are.

Canon EOS Autofocus - BFCV and micro-adjust (MA)

In this note, I will talk about two related matters in the context of the Canon EOS autofocus system, the best focus correction value (BFCV) and AF micro-adjust (MA). Phase Comparison The Canon EOS AF system revolves around the quantitative determination of the "error" in the current focus...

www.openphotographyforums.com

 

[guy schlacter]

MFA values from Canon:

this is clarifying what Chuck said about Field of Focus either in front or behind the focal plane.

http://www.learn.usa.canon.com/resources/articles/2011/af_microadjustment_article.shtml

When setting the Microadjustment, you'll see a scale on the camera's LCD monitor with up to + or - twenty steps. Each step is a very fine increment, equal to 1/8th of the depth-of-field you'd have with the current lens wide-open. And that 1/8th of the depth of field is only moving forward (toward the camera) or back (toward the background) from the sharpest plane of focus. The main thing to remember here is that these are very fine increments.

So based in this what is the answer? I know there will be an impact if I make one MFA move on an f1.2 lens at close distance. I will see a significant impact. If whatever MFA device I use tells me I need one move will I see it on an f5.6 lens at 100 feet? I'm trying to prove mathematically you won't see a difference.

If I were Canon I would have never gave a value of ⅛ and just said fine increments. There is still no real value for what ⅛ is.

Zee,

Each "click" of MFA is equal to 1/16 of the whole DOF from DOF calculator (calculated for the specific lens focal length, wide open aperture, the subject distance) for mfa. So, the 1/16 is a number that changes based on subject distance for a fixed focal length lens at its widest aperture.

Each "click" can be represented as (1/16) x Total DOF*

*BUT NOTE: this is for the lens calculated DOF at wide open aperture. Any stopping down will incur the value from the wide open aperture calculated.

Again, small numbers are small changes, but MFA does help AF place/center the focal plane during auto focus. Said another way, it shifts the actual focal plan and DOF relative to the desired focus plane. This is useful when your DOF just "fits" your subject or other reasons necessitating alignment of the whole DOF around the desired af plane.

Example: Shooting a face wanting eyes, nose and ears to be in focus when focusing on the eyes. IF no MFA done, the AF might create sharp eyes at the front of the DOF so the nose is OOF. Thus negative MFA would be needed.

 

[c h u n k]

There were worthy adjustment on all of my lenses, except 18-135. I had a -1 and a +1 (short n ling end). I shot with those adjustments and then with out - even at the long end and close for narrowest dof, and i couldnt tell any difference for the life of me, so I just left it

 

[elfroggio]

Each "click" of MFA is equal to 1/16 of the whole DOF from DOF calculator (calculated for the specific lens focal length, wide open aperture, the subject distance) for mfa. So, the 1/16 is a number that changes based on subject distance for a fixed focal length lens at its widest aperture.

Each "click" can be represented as (1/16) x Total DOF*

Quote from Canon:

When setting the Microadjustment, you'll see a scale on the camera's LCD monitor with up to + or - twenty steps. Each step is a very fine increment, equal to 1/8th of the depth-of-field you'd have with the current lens wide-open.

Each step = 1 click. How does an 1/8th becomes a 1/16th?

 

[guy schlacter]

Each "click" of MFA is equal to 1/16 of the whole DOF from DOF calculator (calculated for the specific lens focal length, wide open aperture, the subject distance) for mfa. So, the 1/16 is a number that changes based on subject distance for a fixed focal length lens at its widest aperture.

Each "click" can be represented as (1/16) x Total DOF*

Quote from Canon:

When setting the Microadjustment, you'll see a scale on the camera's LCD monitor with up to + or - twenty steps. Each step is a very fine increment, equal to 1/8th of the depth-of-field you'd have with the current lens wide-open.

Each step = 1 click. How does an 1/8th becomes a 1/16th?