[BruceBanner]

My sincere apologies for bringing this up again, I just wanted to quickly ask if my assumptions are correct or misplaced on a specific point relating to Focus Chart/Sheets and fine adjustment/tuning for lenses.
Some of you may know I have visited this subject before with mixed results/feelings. I have actually felt I have had better results fine tuning 'in the field'. For example I have adjusted the FA 50mm and DFA 100mm to both being +5 whilst out and about, and when retesting with Focus Charts that does seem to be a pretty sweet balanced spot.

My thing is this;

- If you have the tripod set up, all three legs anchored out correctly, height of the tripod perfect for the camera mounted to it, use the Electronic Level to get both levels perfectly in the middle, use Live View and a grid mode to also confirm that the middle grid line is perfectly pushing though the focus sheet text and smack bang in the middle of the '0's of each left and right side (ie its not tilted at all), BUT... (and here's the but..), when you go to check on the result and zoom in on the pic, if you zoom in say x2 or x4, and the screens aspect ratio of 3:2 shows that perhaps the top blurred number digit is say '6' and the bottom blurred digit is say a different number entirely (like 7), or that it's still the number 6 but you can't see as much of it... then does that mean the focus chart/sheet is NOT perfectly at a 45 degree angle, and therefore ALL results are not to be trusted when fine tuning? Because I have tried my best using folded pieces of paper and protractors to get as near perfect 45 degree angle as I can with the focus sheet but it just occured today that perhaps if when I am zooming/cropping in on camera to check the results, if I pan across to the left or right to check the focus of the digits, if the positive (above 0) and the negative digits (below 0) are not perfectly in 'approximately' the same place/plane on the Live view, then I've messed up?

Does what I say make sense?


Pretend this is the Live view;

------------------------------------------
7
6
5
4
3
2
1
0
1
2
3
4
5
6
------------------------------------------

Notice how at the bottom the last digit seen is 6, whereas at the top of the digit the last digit seen is 7 (both would be blurryish.)

Cheers,

Bruce

---

Last edited by pjv; 10-25-2017 at 01:11 AM. Reason: Remove masked swearing

[pschlute]

The 45 degree angle is not relevant. Anything approximating 45 degrees will do. The important thing is that the focus chart is level; the camera is level; and that the camera is aligned straight. The fact that you can see more of the top of the chart than the bottom when zooming in is to do with perspective, not misalignment.

You do the focus test at widest aperture and should see clearly if FF or BF exists.

But if you find that adjustments "in the field" are giving you the results you want then stick to that method.

[BruceBanner]

The 45 degree angle is not relevant. Anything approximating 45 degrees will do. The important thing is that the focus chart is level; the camera is level; and that the camera is aligned straight. The fact that you can see more of the top of the chart than the bottom when zooming in is to do with perspective, not misalignment.

You do the focus test at widest aperture and should see clearly if FF or BF exists.

But if you find that adjustments "in the field" are giving you the results you want then stick to that method.

Ah pschlute, you come to my rescue again

Yes, as I said before I had kinda mixed results with Focus Charting before, and I kinda gave up on it. It wasn't until this particular day that i had to hang outside that I got bored and just grabbed a letter from the mail box, threw it down onto the garden path, grabbed my camera, sat down and braced the camera pretty well (but it was Australia and broad daylight so fast shutter speeds with aperture wide open anyway) and did some practice shots (center focus on the text/mailing address etc).
I then (another day) set up the Focus Chart/Sheets to see if what I changed for outside corresponded with a more stricter set up and yeh it kinda did, no need to readjust anything, +5 seemed to be decent for both lenses.

But then I got thinking about the angle of the chart and what if all this time I had the chart more at 43 degrees or 48 etc, would that then skew my results up when trying to balance out the focus! Would that explain my initial error and 'giving up' on Focus Sheeting entirely?

---------------

Ok, but here now, what about this point. All Focus Charts/Sheets that I have seen are about configuring the AF for the center point only. I read that on my K-1 the 3 middle spots (in the vertical line, not horizontal 3) are optimised for accuracy AF, and that other 'spots' are less reliable.

I tend to avoid using center spot when shooting, if I want to snap a portrait of someone wide open, and want their eye in focus, the eye is often not in the center of the shot when framing, so I might spot focus upper left or right of middle more. Could you perform the Focus Chart with a spot that was more in this placement? Would the results and fine adjustment tuning differ? Or is it just that the AF system finds it harder to 'lock' onto that spot as opposed to the centre or 3 vertical center points? I do notice when using AF.C for example and choosing a point of focus further away from the middle mark that the AF seems to struggle to lock on, you could be perfectly still and it goes back and forth trying to lock on and fail. I'm now just pondering the actual accuracy of the focus on these non center points now...
If this was true, in theory you could set up the different User Modes on the camera to being different spot focus points and have subsequently different Fine Adjustments for each point/User mode to increase the accuracy of getting point in focus.

[Alex645]

A few thoughts to share: I believe the main reason the center point is used for focus adjustments is that the center of modern lenses is almost always the sharpest part of the lens with the greatest resolution. If you get the center fine tuned, the rest of the field will benefit. Bottom line: Using a non-centered focusing point should not make any difference unless your lens is de-centered.

The angle of the chart is not that significant. Yes, at exactly 45 degrees with everything level will give your pluses and minuses an equal correction, but you could also just have a 90 degree flat subject with detail perpendicular to the focal plane with a wide open aperture and test results at various fine tuned + or - until you found the sharpest setting. Angled charts just give you quicker info if it's front or back focusing.

[pschlute]

Could you perform the Focus Chart with a spot that was more in this placement?

I can't see that it would make any difference. When you perform the AF Fine Adjustment the camera physically changes the position of the sensor (or is it the AF sensor, i cannot remember). This adjustment will affect all AF points together.

[GUB]

Maybe this is relevant. Here is a link to a dof calculator. as you can see the dof beyond the focus point is greater than in front. The degree varies in different setups.
Online Depth of Field Calculator

[BruceBanner]

I've put my 'thinking hat' on, and I now understand that if the chart is at a different angle other than 45 degrees it matters not, the extent of the DoF will become more or less extreme when viewing the digits away from the center focus point etc, but I am now still pondering about the concept of when tilting the focus chart, if that the upper and lower portions of the Live View screen show different numbers, or perhaps a full '9' is seen top but at the bottom the '9' is almost fully not there, perhaps only a 1/5th seen, does this not suggest that the camera is not bang on spot in the middle of the chart? It can look like it is, the grid lines pass exactly through the center and the '0's perfectly, but if the angle and tilt of the camera is off then this would present unevenness? Basically I'm curious if a guideline to getting things really spot on for the test could be to ensure that when zooming in that the numbers seen both positive and negative to the 0 are showing approximately the same? You want a set up so that this is seen;

----------------------------------------------
| 3 3|
| 2 2|
| 1 1|
| 0 0|
| 1 1|
| 2 2|
| 3 3|
----------------------------------------------

and not this (exaggerated example, pretend the '0' looks more centered than this);

----------------------------------------------
| 4 4|
| 3 3|
| 2 2|
| 1 1|
| 0 0|
| 1 1|
| 2 2|
| 3 3|
----------------------------------------------

[Alex645]

Bruce, if I understand your question, the answer is that it is not significant. It does not matter as long as your AF point is on the zero plane.

Also, keep in mind that the microfine AF tuning is for use with the optical viewfinder which uses phase detection which is a faster AF but less accurate. If AF in Live View, the camera is using the slower but more accurate contrast detection right off the sensor, and thus not intended for AF fine tuning.

Kudos to Pentax that both your K-1 and K-50 have 100% coverage in your OVF.

[pschlute]

Maybe this is relevant. Here is a link to a dof calculator. as you can see the dof beyond the focus point is greater than in front. The degree varies in different setups.
Online Depth of Field Calculator

At the distance one performs AF tests, and at maximum aperture, the DOF will be as close to 50/50 front/back as makes no difference. The point of AF testing is to ensure that the point of sharpest focus is in the centre of the chart

---------- Post added 10-25-17 at 11:28 PM ----------

Bruce, if I understand your question, the answer is that it is not significant. It does not matter as long as your AF point is on the zero plane.

This is correct. Bruce you are overcomplicating matters. Unless your camera/lens combo has been dropped from a great height you will only be looking at the -2/-1/ 0/+1/+2 numbers

[BruceBanner]

Bruce, if I understand your question, the answer is that it is not significant. It does not matter as long as your AF point is on the zero plane.

Also, keep in mind that the microfine AF tuning is for use with the optical viewfinder which uses phase detection which is a faster AF but less accurate. If AF in Live View, the camera is using the slower but more accurate contrast detection right off the sensor, and thus not intended for AF fine tuning.

Kudos to Pentax that both your K-1 and K-50 have 100% coverage in your OVF.

Yeh, I would just use the Live View mode for assistance in grid lines and ensuring that what I had setup looked like everything was bang on centered, then toggle to OVF for firing the shots (and following guidelines in other focus sheet/charts such as making the AF work each time u fire a shot such as placing yer hand in front of the camera briefly etc).

At the distance one performs AF tests, and at maximum aperture, the DOF will be as close to 50/50 front/back as makes no difference. The point of AF testing is to ensure that the point of sharpest focus is in the centre of the chart

---------- Post added 10-25-17 at 11:28 PM ----------



This is correct. Bruce you are overcomplicating matters. Unless your camera/lens combo has been dropped from a great height you will only be looking at the -2/-1/ 0/+1/+2 numbers

Over-complicating things is my Superpower!

See... the focus chart should look like this right; (EDIT See below attachment)

---------------
/ \
/ \
/ \
/ \
/-----------------------\
/ \
/ \
/ \
/ \
/--------------------------------\

Even tho it's tilted/perspective changed, the distance/height from the middle line (where 0 should be) to the top should be the same distance/height from the middle line to the bottom in order for calibration to be correct. The viewfinder is 3:2 aspect ratio, if using a grid on the view finder (just to help set things up) and that grid line is a middle line passing through the 3:2 View Finder, and is presenting as being bang on the middle line, and you can zoom in and verify that the grid line is nicely 'slicing' through the middle of the '0's left and right, but then when you pan and look around and have a look at the top of what the view finder can see and it shows a digit that the bottom doesn't get to, doesn't that suggest the camera (despite using the Electronic level (horizontal and vertical) is actually not being completely truthful (because I wonder how well calibrated that system is...).

Wouldn't it suggest this is occurring (exaggerated); (EDIT, see below attachment)

Focus Sheet > /
/
/
--------------------------/
-------------------------- /
Camera>------------------ /
/


i.e the camera is tilted up slightly and not looking out completely horizontal.

It should ideally be like this;

Focus Sheet > /
/
/
Camera>-----------------------------------------------------------------------/
/
/
/

Now if calibrating and using fine adjustments for the former scenario above, doesn't that mean yer correcting AF incorrectly? lol

PS: I'm having way too much fun drawing these stupid charts.

Also... This was something I never saw/noticed/pondered about before when calibrating the 50mm 1.4 and 100mm 2.8, it wasn't until I was testing the 15mm f4 that I had to start looking further away from the 0 mark to judge whether the blur or sharpness was the same for both positive and negative values or not that I realised that the numbers corresponding on the Playback View Finder at the bottom and top edges were not completely the same, that either the top had an entire extra digit compared to the bottom, or that perhaps half a digit was missing or more etc, and then that made me think about the whole test.

You hate me now, i know... I kinda hate myself

EDIT: Damn the thread isn't displaying correctly what I am seeing in develop mode, I shall attach screen shots

[Alex645]

Over complicating things is not a superpower....itʻs kryptonite

The key here is to simplify and reach your objective (no pun intended). Having everything level or aligned x/y/z is nice, but not critical unless youʻre a troll

And when you mentioned problems fine tuning your 15mm...the DOF is so huge I wouldnʻt bother other than maybe an infinity check. Also the 15mm is going to seriously distort your expectations of the image at 50mm and is certainly not going to be the same with an UWA 15mm. Note the amount of wide angle MF lenses for Pentax? Itʻs simply because the DOF is so great, AF is effectively as helpful as the no-smoking light on planes.

[BruceBanner]

Over complicating things is not a superpower....itʻs kryptonite

The key here is to simplify and reach your objective (no pun intended). Having everything level or aligned x/y/z is nice, but not critical unless youʻre a troll

And when you mentioned problems fine tuning your 15mm...the DOF is so huge I wouldnʻt bother other than maybe an infinity check. Also the 15mm is going to seriously distort your expectations of the image at 50mm and is certainly not going to be the same with an UWA 15mm. Note the amount of wide angle MF lenses for Pentax? Itʻs simply because the DOF is so great, AF is effectively as helpful as the no-smoking light on planes.

How do u do an Infinity check?

[Alex645]

How do u do an Infinity check?

With a wide angle prime outdoors with objects in the distance (preferably mountains, buildings, things with contrast and hard edges) AF to infinity. Then bracket a series of +1, +2, +3, -1, -2, -3, etc. and pixel peep comparing it to your base shot at 0.

For telephoto lenses, a full moon is often the preferred target, but obviously that's not going to work on any normal to wide angle focal length.

[pschlute]

Bruce, what you are seeing is perspective. Because your focus chart is at an angle sloping away from your camera, the number 9 at the top of the chart is a lot further away from the lens than the number9 at the bottom of the chart. The further away an object is , the smaller it is recorded by the camera. Depending on what lens and distance you are photographing from this will be more noticeable in some instances than others, but is quite likely that the number 9 is visable at the top but not the bottom.

[BruceBanner]

With a wide angle prime outdoors with objects in the distance (preferably mountains, buildings, things with contrast and hard edges) AF to infinity. Then bracket a series of +1, +2, +3, -1, -2, -3, etc. and pixel peep comparing it to your base shot at 0.

For telephoto lenses, a full moon is often the preferred target, but obviously that's not going to work on any normal to wide angle focal length.

Ok. So mountains are fine here, I live in the Blue Mountains, have plenty of them to test with

When you say 'AF to infinity' yer then obviously meaning the furthest mountain/skyline thingy away. Like also... with my DA 15mm lens, is infinity considered something like anything past as certain distance?

Wait... there's no way to bracket with different fine adjustments, so do you mean snap the shot, go into the camera, change the fine adjustment, shoot again, and so on so forth? Manually making the changes each shot? I'm not sure EXIF would show any Fine Adjustment change, so perhaps it's wise to start at -10 and work my way up to +10, then when reviewing the shots at home on pc I can better gauge which shot looks the best for the mountain range in the distance? So for example, of the 21 shots taken (-10 through to +10), if I deem the file that is 15th out of 21 is the best then +4 is the adjustment used for the shot.

Bruce, what you are seeing is perspective. Because your focus chart is at an angle sloping away from your camera, the number 9 at the top of the chart is a lot further away from the lens than the number9 at the bottom of the chart. The further away an object is , the smaller it is recorded by the camera. Depending on what lens and distance you are photographing from this will be more noticeable in some instances than others, but is quite likely that the number 9 is visable at the top but not the bottom.

Yep, now that I have had some time away at work and come back I realise what a numpty I have been.

This pic illustrates. As you say perspectives;