Focus/Sharpness Expectations

This gentleman wants to show us how knowledgeable and smart he is. So the perturbations in the design become his norm and everything else is worthless.

There is no reason to discuss this kind of sophistry here. You don't like shooting at charts to determine whether a lens front/or back focuses, it's your problem. My lenses all are calibrated and work very well.

I am out of here...

cdifoto wrote in post #16863995
Softness is usually one of three things:

1. Missed focus
2. Defect in workmanship
3. Optical design

That's exactly what I said.

Where I have a problem with test charts is that they are a two dimensional subject.

Something to consider. At what distance do you place the test chart? Is there are formula to what distance the chart is placed in relation to focal length?

Example, testing an ultra wide angle like EF-S 17-55mm mentioned in the OP, at 17mm. Wouldn't it be fair to say the chart to sensor distance would be somewhere near minimum focus distance if you want to fill the frame with the chart (depending on the size of the chart)? What are the dimensions of a typical lens test chart? A4, A3 or whatever? Is the 17-55mm at 17mm at its best near minimum focus distance? What about when testing at 55mm. Many zooms perform differently at different focal lengths. Do you discount a zoom because it performs better at 17mm than 55mm?

And finally, at what aperture would you test? Wide open, a couple of stops down or at minimum aperture. Whenever I buy a new lens, I always test at full aperture in low light because that is where and how I do the majority of my shooting. If the lens performs well at f/2 or f/2.8 with a highish shutter speed at 1600 ISO in low high, I'm happy. I shoot mainly with fast primes.

For me, test chart have too many variables which could cause seeing things that aren't really there. I have never used a test chart in my life and see no need to.

MakisM1 wrote in post #16864040
This gentleman wants to show us how knowledgeable and smart he is. So the perturbations in the design become his norm and everything else is worthless.

There is no reason to discuss this kind of sophistry here. You don't like shooting at charts to determine whether a lens front/or back focuses, it's your problem. My lenses all are calibrated and work very well.

I am out of here...

What type of test charts are you talking about? There a two I can think of the top of my head. One is to test focus accuracy. These have a scale of measurement and placed at 45 (or thereabouts) degrees to the sensor. The other is a flat chart usually stuck to a wall. These are used for measuring MTF (lens resolution in simple terms). The two aren't interchangeable.

Having a lens that constantly misfocuses can (and often is ) still optically sharp. A lens that is optically poor, can still focus accurately.

Now,, which test procedure are you talking about?

Also,, you still haven't answered my question. Why do testers, reviewers and manufactures give a number of MTF values, namely for centre, edge and sometime extreme edge resolutions? By the way, MTF has nothing to do with auto focus accuracy.

Ideally you use both. You check for focus with one and sharpness with the other. No one here is even talking about MTF numbers. Heck I couldn't care less about those personally. But if I'm having trouble in the real world, I want to know why. Shooting more in the real world won't help me figure it out. Throwing together a controlled situation sure as heck will.

Heck, even some batteries lined in a row and then staggered side by side is better than shooting out in the real world with all kinds of variables at play.

Without getting too carried away with technicalities, what I'd be doing is testing focus outside, in good natural light. No need for fancy test charts, just position yourself beside something like an outside wall of your house, and try focusing on a distinctive feature such as a door/window frame. At about a 30 degree angle to the wall. Brace yourself off the wall with your elbow. Shoot level. Shoot wide open aperture, as this is where any focus inaccuracy will show up. Like this. I focused on the little black door latch. Something with good contrast.

IMAGE LINK: https://flic.kr/p/npsp​iD

I'm not sure why all the bickering, I doubt it helps the OP when those who are trying to help end up arguing with each other...!

As has been said, there are various considerations, but it doesn't have to be all that difficult and complicated to step through a few things to address those considerations, such as how well the camera/lens focuses under good" use/conditions, and given good focus what can one expect regarding sharpness, of course in the center, but, hey, you can check out corners and edges too!

And, if during the course of things we do suspect front or back focusing (that is not caused by us not "properly" using the AF), well, sure, doing the angled chart test, then using MFA if you can address things that way...

speedync wrote in post #16864100
Without getting too carried away with technicalities, what I'd be doing is testing focus outside, in good natural light. No need for fancy test charts, just position yourself beside something like an outside wall of your house, and try focusing on a distinctive feature such as a door/window frame. At about a 30 degree angle to the wall. Brace yourself off the wall with your elbow. Shoot level. Shoot wide open aperture, as this is where any focus inaccuracy will show up. Like this. I focused on the little black door latch. Something with good contrast.

If this looks okay, you're set. If not, check things such as shutter speed (make sure you have plenty) IS on & off and ISO (not unnecessarily high)
There is no home test to check if sharpness is at an acceptable level though. It's up to your eye

The problem with shooting with a wide open aperture is that it can mislead you into believing that your lens is "soft", since most consumer lenses are, in fact, softer at wide open apertures. Typically a stop "down" from wide open will give decent results.

However, this is not to determine whether a camera/lens combo actually successfully focuses on your "target"...your suggested approach can be fine, although for some folks taking steps to protect from camera shake can be a real consideration!

Yeah, I only suggest shooting wide open, to maximise, or make more apparent, any front/back focusing issues. Nothing to do with sharpness. I'd personally do that as the very first step, as you'll NEVER get a sharp image if the focus point/distance is way out of whack. Or it will be sharp where you don't want it to be

Yep narrow down focus first then go for optical quality.

MakisM1 wrote in post #16864040
This gentleman wants to show us how knowledgeable and smart he is. So the perturbations in the design become his norm and everything else is worthless.

There is no reason to discuss this kind of sophistry here. You don't like shooting at charts to determine whether a lens front/or back focuses, it's your problem. My lenses all are calibrated and work very well.

I am out of here...

You do realise making aspersions against a poster and not the post possibly contravenes forum rules?

If you have a comment to make, comment on the post and not the poster.

cdifoto wrote in post #16864130
Yep narrow down focus first then go for optical quality.

Yeah, things can be simplified to a few practical steps!

Back a while ago I put together a basic "starting test" for focus:

https://photography-on-the.net …php?p=9995436&p​ostcount=2

yogestee wrote in post #16864072
What type of test charts are you talking about? There a two I can think of the top of my head. One is to test focus accuracy. These have a scale of measurement and placed at 45 (or thereabouts) degrees to the sensor. The other is a flat chart usually stuck to a wall. These are used for measuring MTF (lens resolution in simple terms). The two aren't interchangeable.

Hmmmm, my flat chart (stuck to a bit of board and mounted on a tripod - a bit more manoevrable than a wall) is used to test focus accuracy. Multiple shots at differing MFA values tell me which is the best.

MakisM1 wrote in post #16857712

Front/back focusing is not the only reason a lens produces blurry photos.

I always recommend using a target designed for the task AND a proper tesdting process.

You can get a target from this site for free. Print it and glue it on heavy posterboard to avoid distortions/inaccuracy​.

Read their testing procedure and understand/apply it.

Use the procedure only in good natural light.

http://www.peleng8.com …k-focus.html#.U1kSTo_ANa​g

Yogestee,

This is my original post which you quoted but obviously did not bother to read.

In it, in the first line, I clearly state:

and proceed to advice the OP on how to determine whether front/back AF is the source of softness, or eliminate it.

The advice contains a URL where he can find a printable testing target (which operates on the principle of a graduated scale at 45 degrees to the sensor plane, while the high contrast AF target is parallel to the sensor, to ensure repeatability, accuracy and no ambiguity in the results).

In your subsequent posts, apart from trying to put tits on a chicken, you present a number of facts that while they may be correct in themselves, they are irrelevant.

We are talking about a specific lens, the 17-55, therefore all the discussion about soft lenses by design have no bearing. However, while as a fact (the existence of the EF 135 2.8 soft focus for instance) is correct, all it does is to obfuscate the specific issues while inspiring awe at the breadth and depth of the knowledge of the poster.

Misleading the discussion with this technique is called sophistry.

Back to our subject now, the OP is not complaining about soft edges on the 17-55. The 17-55 is well known as being a near L quality lens with good sharpness credentials. It should not be visibly worse than the nifty fifty.

It appeared that the OP was applying MFA at random (or at least in sequence from + to - and vice versa) and shooting different subjects at different instances trying to determine whether the MFA choice of the moment was right.

For this reason it is preferable to shoot targets appropriate for the task. They provide results that are consistent, repeatable, accurate and not ambiguous.

From your post #23 you started derailing the discussion towards softness around the edges, flat field lenses while offering your barbs about

As I explained above, 'real life' photos may indicate you have a problem but they are not controlled or repeatable conditions adequate for testing.

All the discussion on edge sharpness, MTF lines etc etc would have some meaning if they addressed a specific problem (soft edges, possibly mild lens decentering). In this thread, they only serve to obfuscate the issue.

Past closing this post, I am leaving shortly for a trip, this is the reason I wanted to check out the conversation last night.

You strung together a number of correct, but nearly irrelevant facts, while offering no guidance to the OP other than not to shoot charts and take real life photos.

There was nothing aggressive in my first answers (as a matter of fact I was rather concilliatory, asking whether we are saying something different). You are the one that brought irony and condescension in the debate with your comments about tits on the chicken and my photographic life.

I think it's a bit rich to complain about breaking forum rules when it suits you.

bumpintheroad wrote in post #16864035
The raw files would be better for review. Not only would they be bigger and more detailed, but DPP could show what focus points were selected.

Could you show me how that works without conversion to a viewable format?

MakisM1 wrote in post #16858408
You should have seen the thread where I posted the results of my (faulty) first EF 24-70 MkII!... People were about to lynch me for not going out to the real world to shoot (I did, that's how I got clued that something was wrong with it), accusing me of trying to calibrate a delicate instrument with a sledgehammer (yeah right, my EF 70-200 MkII AND my nifty were spot on...), not knowing how to experiment (spent a career in R&D, including model simulations and model testing).

Of course, nobody bothered when the replacement EF 24-70L MkII arrived and the results showed it spot on...

I feel your pain. I had exactly the same experience when my camera turned out having a faulty AF circuitry failure. I got accused of everything that you describe, and more. It was particularly funny because I have 40 years of experience as a test engineer!