[tblachowicz]

How often do you use AF in selective (SEL) mode?

The reason why I'm asking is that I've noticed that AF is far more accurate when in AUTO mode than in SEL mode. The images focused in AUTO mode are razor sharp and the same shots in SEL mode tend to be a bit blurry.

I've also noticed that in AUTO mode often two AF points are often used together (the red dots in viewfinder). This never happens in SEL mode. Can you please help me understand this?

Please note this question is a cross-post with StackExchange question: autofocus - Why AF performance is different in automatic and selective mode? - Photography - Stack Exchange.

Cheers,
Tom

[starbase218]

I've also noticed that in AUTO mode often two AF points are often used together (the red dots in viewfinder). This never happens in SEL mode. Can you please help me understand this?

I would be very surprised if it used more than one point in SEL mode. The whole point of SEL is that you can SELect a single focus point among all 11 points, for the camera to focus on. In Auto mode, you never know what area the camera will focus on until you actually half-press the shutter. It then decides which AF point or points it will use to focus.

As for your problem with blurry images, remember that the actual focus area for a single AF point is far larger than the little red square in the viewfinder. So you may be thinking that the camera focusses on one detail, while it actually focusses on something else.

I use SEL almost exclusively. Sometimes I leave the center point active, but at times I chose a different point. This can reduce focus error when recomposing after having confirmed focus. The reason I don't normally use Auto is that I like to decide what the camera should focus on myself .

[philbaum]

Tom,
an intriguing question. It could be that the camera seeks out the best phase focusing targets in auto mode, rather than being forced to focus on less desirable targets (from a systems standpoint) in sel mode.

Can you post an image where this has occurred? Otherwise your question may be met with disbelief.

Creatively, a camera is not going to know what you had in mind from a composition point of view. i suspect most cameras just auto focus on the nearest targets - kind of a mindless approach if one is serious about the image. Works find for casual shots or street shots, perhaps.

[bxf]

OK, this may be a record for the resurrection of an old thread, but here goes anyway:

As I always use Center AF point, I'm not in a position to argue the OP's claim that Auto yields more precise focus than SEL. However, I am nevertheless curious what it means when two AF points light up when using AUTO (and they do, in case you're wondering). Is it possible that two points light up when the target is determined to be between these two AF points? And if so, would it not make sense to think that the software is able to adjust the focus distance accordingly, rather than using just one AF point or the other? If this is done, then it would certainly result in more accurate AF.

Does anybody have any practical experience with this?

EDIT: I was lying down, relaxing, and it suddenly occurred to me to wonder what I was thinking when I wrote the above. It implies that AF points are aimed at some fixed locations, which makes no sense. So before somebody else says "stupid", there I've said it myself. Still, one can't help but wonder what it means for two AF point lights to come on.

---

Last edited by bxf; 11-18-2013 at 12:20 PM.

[Joojoo2010]

This is intresting...i will try auto mode af next time

[adwb]

Yes I have also wondered for years about the two red lights and what it is actually focusing on or if it is in between.
Remember the focus is looking for a contrast the more strong the better between light and dark.

What if the 45 degree focus tables that are used for back focus were used as a auto focus target? As they have a large expanse of blank with numbers on the sides and a bar in the centre it would be intesting to see what happens in auto and what is in focus in the resulting image compaired with using select and one focus point and selection on one of the previous selected auto points. Camera on tripod obviously and evry thing in manual, wb, iso , aperture and shutter etc.

[bxf]

An experiment for a rainy day, I guess.

[adwb]

ok with out posting images lets see if i can explain what I found.

k 5 on tripod using a sigma 24-60 hsm lens at 2.8 shooting a test chart set up 4 foot away upright but angled so the left side is further away form the camera , hope that make sense.

on this chart is a row of number on both sides and a very small bar in the centre with very small white text surrounded by white

first shot using select and the centre square point on the bar and writing it is sharp and the left side [further away ]numbers are sharper than the right side [which is closer]. that shows I have a back focus issue I know and am happy with it. this is image 1

second shot using auto 11 the camera selects and puts a red square in the centre again but also on the RIGHT side row of numbers. the result is now the opposite of shot one. Where the squares were located is now sharp and the left is now blurred. image 2

third shot select in manual select the same square only on the right side number .This time the right side number are still sharp but the centre bar and the fine writting on it is not as sharp as the center only selection image 1. this is image 3 .

i have tried this a couple of times at different apertures and seem to get the same result.

not exhaustive test and i don't really know what it proves perhaps someone can take it further.

[bxf]

ok with out posting images lets see if i can explain what I found.

k 5 on tripod using a sigma 24-60 hsm lens at 2.8 shooting a test chart set up 4 foot away upright but angled so the left side is further away form the camera , hope that make sense.

Yup.

on this chart is a row of number on both sides and a very small bar in the centre with very small white text surrounded by white

first shot using select and the centre square point on the bar and writing it is sharp and the left side [further away ]numbers are sharper than the right side [which is closer]. that shows I have a back focus issue I know and am happy with it. this is image 1

second shot using auto 11 the camera selects and puts a red square in the centre again but also on the RIGHT side row of numbers. the result is now the opposite of shot one. Where the squares were located is now sharp and the left is now blurred. image 2

third shot select in manual select the same square only on the right side number .This time the right side number are still sharp but the centre bar and the fine writting on it is not as sharp as the center only selection image 1. this is image 3 .

i have tried this a couple of times at different apertures and seem to get the same result.

not exhaustive test and i don't really know what it proves perhaps someone can take it further.

As you say, one can't draw definitive conclusions from your test, but it does indicate that there are possible differences in behaviour, of which one may want to take advantage sometimes. Specifically, if we determine that the Shot 2 behaviour is consistent, one may choose to use AUTO when a front bias is desired. I wonder if it would make sense to try this when one wants more background blur.

Thanks for your efforts.