Index  •   • New posts  •   • RTAT  •   • 'Best of'  •   • Gallery  •   • Gear  •   • Reviews
New posts  •   • RTAT  •   • 'Best of'  •   • Gallery  •   • Gear  •   • Reviews

Thread started 02 Mar 2018 (Friday) 18:52

How to tell which AFMA is sharpest?

Mar 03, 2018 08:03 |  #16

texshooter wrote in post #18576256
I'm sure it's easy to see the difference between 0 and +10 AFMA on the ISO12233 chart. But have you tried comparing +9 vs +10 points? Not so easy I bet.

As Teamspeed points out, +/-1 changes are so small that even at f/1.2 you'll be had pressed to notice a difference. Also, PDAF is inherently inaccurate. Rarely with the lens focus to exactly the same. It gets really close, close enough that you will never see the difference in normal shooting, but there are tolerances within the mechanics of the lens.

So as you have already determined that optimal focus is somewhere from -1 to -5, split the difference.

Not sure why, but call me JJ.
I used to hate math but then I realised decimals have a point.
.
::Flickr::
::Gear::

 LIKES 0

Mar 05, 2018 08:55 |  #17

You can take a look at the profiles of edges or line pairs or whatever on your target image and decide if there is some meaningful threshold that you are willing to accept as the best value, or you can look at all of the images relative teach other and choose the "best" profile.

Attached is a contrived example that I made in PS - I drew a black line and paired it with a copy with differing arbitrary separation distances - when this is done precisely, these are "line pairs" and the ability of the optical system to discriminate between them is a measure of the spatial resolving power of the system.

I applied a gaussian blur to the lines, progressively increasing the blur, at four levels. I plotted the gray value of lines across the line pairs and looked at the results, compared to the reference. Using the plot of the grayscale profile, you can see how the edges of the lines get softened and their signal decreases as a function of blur. Maybe an evaluation of your target images with particular areas picked out for evaluation can be analyzed this way.

Also, you need to shoot images in the field and evaluate those to really test the "bench top" settings you choose with a target. This is the proof of your setting, not some planar target.

kirk

HOSTED PHOTO

Kirk
---
images: http://kirkt.smugmug.c​om

 LIKES 0

Mar 05, 2018 10:45 |  #18

The point of AFMA is to get the plane of focus to best coincide with the location of your DESIRED plane of focus.

So you should be looking at scene in which you can assess that the camera AF plane coincides with the location of your desired plane of focus -- NOT 'how sharp' a single plane happens to be.

Look at photos 2 and 3 of this old post to see how you can assess point of best focus with a flat plane target and an inclined plane ruler for location assessment of the plane of focus.

You need to give me OK to edit your image and repost! Keep POTN alive and well with member support https://photography-on-the.net/forum/donate.p​hp
Canon dSLR system, Olympus OM 35mm system, Bronica ETRSi 645 system, Horseman LS 4x5 system, Metz flashes, Dynalite studio lighting, and too many accessories to mention

 LIKES 0

Mar 05, 2018 10:53 as a reply to  @ Wilt's post |  #19

But Wilt, to really assess what is in focus, you have to zoom in quite a bit. So something like this so you can determine a +1 or -1 adjustment...

I agree though, with just a 100% pixel view (or less actually), the layout you describe works well. There are other versions of the same type of method too.

HOSTED PHOTO

 LIKES 0

Mar 05, 2018 11:10 as a reply to  @ TeamSpeed's post |  #20

In that older thread that I linked, a few posts later windpig made a testimonial about my methology being a good one that he found to be 'best' and 'easiest'.

One must keep in mind that 'in focus' is within Canon's standard of 'within one DOF'.
At 200mm, 50* FL (Canon's suggested test distance) or 394mm yields a 20/20 visual acuity DOF zone which extends only 2.6" in front/behind the plane of focus, and we can clearly determine in photo 3 of my post that plane of focus is best within about 0.5" back...even without 100% view assessment.

You need to give me OK to edit your image and repost! Keep POTN alive and well with member support https://photography-on-the.net/forum/donate.p​hp
Canon dSLR system, Olympus OM 35mm system, Bronica ETRSi 645 system, Horseman LS 4x5 system, Metz flashes, Dynalite studio lighting, and too many accessories to mention

 LIKES 0

Mar 05, 2018 12:32 |  #21

digital paradise wrote in post #18576400
I think the one on the right looks sharper. If you want to take all the guesswork away you may consider this.

http://www.reikan.co.u​k/focalweb/

I was about to post the same thing. The scientific approach to MFA.

Sam
5D4 | 6D | 7D2 (2 bodies) | Reasonably good glass
Gear List

If I don't get the shots I want with the gear I have, the only optics I need to examine is the mirror on the bathroom wall. The root cause will be there.

 LIKES 0

Mar 05, 2018 12:35 |  #22

The one on the right looks sharper

 LIKES 0

Mar 05, 2018 14:56 |  #23

at 3200x you're at paralysis by analysis...it's not that critical. You'll never know the difference at print w/ normal viewing distance.

Sony a6300/ 16-70/4 / 70-200/4 G / Sony HVL-60M
https://designcostello​.com/

 LIKES 0

Mar 06, 2018 19:03 as a reply to  @ DagoImaging's post |  #24

I ran the test again and this time the software recommended -4 AFMA, which is within my visually expected range of -5 to -1, so I will go with that. There are many variables that can skew the software's results, such as focus inconsistency, invisible target glare, widest aperture defects, target-to-camera distance, etc. Best to trust but verify.

But I still stand by my original claim. If the software can tell which nearest neighbor AFMA is sharpest, then there should be a way to replicate that precision in PS using some sort of advanced focusing mask. But, I'll trust the software if, and only if, my eyeballs agree.

 LIKES 0

2,343 views & 3 likes for this thread
How to tell which AFMA is sharpest?
AAA
 x 1600 y 1600

Not a member yet?
Register to forums
Registered members may log in to forums and access all the features: full search, image upload, follow forums, own gear list and ratings, likes, more forums, private messaging, thread follow, notifications, own gallery, all settings, view hosted photos, own reviews, see more and do more... and all is free. Don't be a stranger - register now and start posting!