Autofocus in Live View mode - Bad idea?

I hear one should not use autofocus in Live View mode when precise and reliable focusing is required. Rather, one should zoom in 10x and focus manually instead. This is because lenses are not designed for Live View focusing and, as such, cannot be calibrated like viewfinder autofocus mechanisms can. Is this true?

I've never had an issue when auto focusing in live view.

I only use LV to see how accurate PDAF is. Phase detect is never better than LV. It can only be as good as LV.

texshooter wrote in post #17473284
I hear one should not use autofocus in Live View mode when precise and reliable focusing is required. Rather, one should zoom in 10x and focus manually instead. This is because lenses are not designed for Live View focusing and, as such, cannot be calibrated like viewfinder autofocus mechanisms can. Is this true?

Contrast detection AF is more accurate than PD simply because it does not need to be calibrated - it requires more time and confidence that the specific feature in the AF zone is the one the user wants to use. It is usually simpler/quicker to zoom in and manually focus because WYSIWYG - you can see that the target is in focus kin the image that will be captured.

"Modern lenses are not optimized (in gearing) for MF" is perhaps what you have heard - that means that you only need a small amount of rotation of the focus ring to go from MFD to infinity (it's designed this way so AF can work quickly) but for manual focus you need to be very precise in your rotation. Movie lenses and older MF lenses you need to turn the focus ring a long way to get from MFD to infinity, so it's easier to make fine adjustments to focus position.

The Contrast Detect AF system that is used in LV, unless you use the quick focus mode that flips the mirror up does a normal Phase Detect AF and drops the mirror, uses the actual sensor to determine the state of the focus. Because it is measuring the focus at the sensor plane it needs no calibration. PDAF uses a sensor mounted in the bottom of the mirror box, which has to be mounted at exactly the same optical distance from the lens as the sensor is. This is where the possibility of miscalibration comes from.

The real problem is that CDAF is perfectly accurate, but really really slow. Also you still have the problem of making the area selected from the image that is used to Af actually look exactly at the spot you want in focus. By using manual focus, and 10× magnification you can ensure that the exact point you want to be in focus is, and usually without worrying about how much contrast you have present, or it's exact orientation.

Alan

Au contraire. According to Michael Tapes, the designer of LensAlign, auto focusing using Live View is unreliable:
https://mtd.zendesk.co​m …an-I-calibrate-Live-View-

texshooter wrote in post #17473479
Au contraire. According to Michael Tapes, the designer of LensAlign, auto focusing using Live View is unreliable:
https://mtd.zendesk.co​m …an-I-calibrate-Live-View-

I'm a tad curious. If you already have the answer, why ask the question?

GeoKras1989 wrote in post #17473704
I'm a tad curious. If you already have the answer, why ask the question?

Second opinion. Do you believe everything you're told?

What he's referring to is that the AF motor drives in the lenses are designed for quick point to point operation. From current position to position of correct focus, as calculated by the standard AF system.
He seems to neglect that today's USM drives are designed with both rapid execution of standard AF commands as well as the slow motion more suitable for video production. Look at the power focus function in lenses like the EF 300 mm f/2.8L IS II USM or the new EF 400 mm f/4 DO IS II USM, for example.
He also totally forgets that the STM focus drive motors are fully designed to support slow, creeping focusing operation as well as rapid focusing.

So I dare say the statement quoted above is either several years old, or he doesn't know what he's talking about.

texshooter wrote in post #17473749
Second opinion. Do you believe everything you're told?

I understand wanting a second opinion. It is telling those who provide what you ask for that they are wrong that I find quite odd.

apersson850 wrote in post #17473752
What he's referring to is that the AF motor drives in the lenses are designed for quick point to point operation. From current position to position of correct focus, as calculated by the standard AF system.
He seems to neglect that today's USM drives are designed with both rapid execution of standard AF commands as well as the slow motion more suitable for video production. Look at the power focus function in lenses like the EF 300 mm f/2.8L IS II USM or the new EF 400 mm f/4 DO IS II USM, for example.
He also totally forgets that the STM focus drive motors are fully designed to support slow, creeping focusing operation as well as rapid focusing.

So I dare say the statement quoted above is either several years old, or he doesn't know what he's talking about.

I lean heavily in this direction. I've had 500D, 60D and 6D that all have LV focus. It is dead-on with any lens I mount. CDAF is so accurate because it is an iterative process. PDAF is so useful because it is so fast. They are two entirely different animals.

GeoKras1989 wrote in post #17473799
It is telling those who provide what you ask for that they are wrong that I find quite odd.

I reserve the right to tell anybody they are wrong whenever I decide they are wrong. I also have the prerogative to change my mind about who is wrong after I've elicited their opinion--even while I wait for a third one. Now then... after reading apersson850's reply, I'm on the fence again.

texshooter wrote in post #17473816
I reserve the right to tell anybody they are wrong whenever I decide they are wrong. I also have the prerogative to change my mind about who is wrong after I've elicited their opinion--even while I wait for a third one. Now then... after reading apersson850's reply, I'm on the fence again.

Exactly! And that is why we ask for others' opinions. I did the same thing when considering a 1D3/7D a little while ago. I decided to stick with my 60D. Now I am thinking about upgrading my crop body to the 1D4! The more you ask, the more you know. Sometimes it helps you decide. Sometimes it just adds to your confusion. Sometimes it just makes your head hurt.

Well, the problem with opinions is that they are often wrong.

Or they may be right for the opinion-holder and wrong for you.

But its nice to share opinions - even if we may decide not to adopt them.

Whats really nice is to have this conversation and then go out and test it for yourself and tell us what worked for you.

This is based on, wait a minute, it's actual data, not opinions or guesses (There's a follow-up here.) Roger Cicala is one of the few people who is in a position to test these issues on multiple lenses and is extremely methodical/systematic about his testing.