Nikon AF help/info

[PhotomanJohn]

I have always assumed that when Nikon refers to 3-D focusing, the third dimension is color which is used to help separate the subject from other objects as it moves.

John

[AB26]

Dan,

Sorry for barging in.

You need no vast technical knowledge of the AF system to get the shots in sharp focus. All you need to know is the basics of what the system does with different settings.

The system works with Focus AREAS. AREA is the important word to keep in mind. With the Focus Area Selector you choose different AF AREA modes to be used, not just a focus mode.

1. Single AREA AF: The user chooses a SINGLE focus AREA for the camera to focus in. You have a single focus Area you can move around within the focus ZONE, using the focus selector.
i. The preferred choice when you want your subject not to be in the centre of the image and you don’t want to re-compose.
ii. Use the focus selector to focus on the subject.

Re-composing is not an option when using AF-C. It will defy the object of AF-C Mode. When using back button AF you will need two thumbs to “lock” focus in AF-C if you are going to re-compose.

2. Dynamic AREA AF: You select the Focus AREA but the focusing system uses information from different focus AREAS to determine focus. When the subject leaves the specific AF AREA momentarily the system will focus on the subject based on information from other focus AREAS.
i. This mode will be used when shooting BIF in conjunction with AF-C mode.
ii. In this mode it is easier to “track” a moving subject.

Viewing images in View NX2 and realizing the focus AREA is in blue sky means you have probably used this setup and the image was not on the specific AF AREA when the shutter was released.

3. Group Dynamic AF AREA mode: You select the GROUP of AF AREAS to be used but the CENTRE AF AREA in the group is the one that has priority in determining focus.
i. This is the mode to be used for moving subjects when you know where you want the subject to be in the image.

4. Dynamic AREA AF with closest subject priority: Pretty much self explanatory.
i. The AF AREAS at the edges are less sensitive to obtain focus.

It is important to know where the cross tipe AF AREAS are in your camera.

It is better choosing focus lock before the shutter can be released. That way you learn better camera technique and an almost 100% hit rate.

If you are using a Sigma lens on a Nikon D800 you should not expect to see very sharp BIF shots. We can get back into the discussion of how photons are directed to specific pixels on the sensor. Lets just say that the Nikon D800 sensor far out-resolves the Sigma lens.

But why do BIF with a D800 if you have a D4 to do it with?

It is important to know how and when to use different Focus AREA modes. The best way of finding out how to is to just do it and see what the camera does in different modes.

It only takes 10 000 shots in each mode to master that mode.

[pnodrog]

Andre, I think you will find Dan uses something a bit better than a Sigma lens and he possibly uses the Nikon D810 so that he has plenty of cropping ability if needed.

[Mark von Kanel]

well this thread is about to get interesting....

[JR1]

Own D4s, D810, D7100, D7000, I NEVER use anything other than centre focus point

Also never used the BF button

[Manfred M]

Andre - I'm not sure if you caught what the real issue is; REPEATABILITY.

When one takes two shots under similar shooting circumstances and they end up with different results, this is what the fundamental problem is. This apparent "randomness" in a modern camera's autofocus system when using multipoint and 3D focus modes is what causing all of the issues. Using a single focus point has some inherent issues too, but at least those; for example having to reframe your shot, are predictable and to some extend, managable.

Your suggestion of 10,000 shots only works if the system is consistent and predicatable. To use an analogy; it would be like driving a car, where the brakes acted differently (and seemingly randomly so) for similar braking situations.

We can all read our user's manuals, but unfortunately, these do not provide a deep enough level of understanding as to how the camera makes its decisions to be at all useful. The material on the internet tends to be either best-guesses by different "experts" or a cookbook approach the tells the reader what worked for the author under specific instances without any knowledge regarding why this worked.

Just as an aside; I find I have the same issue with TTL flash work using multiple flashes. I find that I cannot get easily repeatable results, so I tend to go back to basics and shoot them on manual.

[chauncey]

REPEATABILITY

You're essentially talking about "keeper rate"...I don't know of any photog that claims 100% keeper rate.
The sheer number of variables is mind boggling, not only gear associated but the shooter himself.

My GBH and Egret series, there was once, of each, that I captured a series of 11 images in a series.
All the other attempts, under similar conditions, led to one or more images missing focus, thru no fault
of the shooter. Sometimes things just come together and sometimes they don't. There is nothing that
functions perfectly every time. Standard deviation exists in the gear as well as the shooter.

[PhotomanJohn]

Re-composing is not an option when using AF-C. It will defy the object of AF-C Mode. When using back button AF you will need two thumbs to “lock” focus in AF-C if you are going to re-compose.

No, you simply take your thumb off of the AF-ON button.

John

[NorthernFocus]

Thanks for all the input, everyone.

John, using color in 3D mode is what Nikon claims to be the difference from other area modes. I've never had much luck with it, perhaps due to not enough difference in colors typically between wildlife subjects and BG.

Andre, thanks for the input. A bit more time reading the thread may have saved you a lot of time typing. Manfred's last post sums up/restates the issue.

Mark, it seemed so

Manfred, thanks for the well written summary of the issue.

Jeremy, I'm finding more and more shooters of high end gear who have reached the same conclusions and are shooting single point AF. I guess that says something.

Chauncey, yours is probably the best advice. We should just quit worrying about it and move on. I just don't do random very well. And it is the seemingly random nature of the variability that frustrates At least when the variability is photographer induced I know what's going on

John, nicely put.

[Peeshan]

"Jeremy, I'm finding more and more shooters of high end gear who have reached the same conclusions and are shooting single point AF. I guess that says something."

According to a bird photography book (by professional photographers) I just read, it seems every time it was safe to use single point they used it, then if there was a chance the bird would move they would use the next step (9 points), then if they where expecting the bird to fly they would add even more focus points. When following the flight they would use only 9 focus points.

I think to get the most out of this feature one must precisely know what s/he wants to shoot and the predictability of the subject. So maybe we should see it like ISO, the lowest the better, but without being afraid to use it if there is a risk.

---P-chan

[Manfred M]

You're essentially talking about "keeper rate"...I don't know of any photog that claims 100% keeper rate.
The sheer number of variables is mind boggling, not only gear associated but the shooter himself.

My GBH and Egret series, there was once, of each, that I captured a series of 11 images in a series.
All the other attempts, under similar conditions, led to one or more images missing focus, thru no fault
of the shooter. Sometimes things just come together and sometimes they don't. There is nothing that
functions perfectly every time. Standard deviation exists in the gear as well as the shooter.

Yes and no; while it is about the "keeper" rate; I can handle shots being non-keepers because I have made either a technical mistake (camera shake, inappropriate ISO, exposure error due to inappropriate metering, shutter speed or aperture selection) or if I have made a compositional error.

What I cannot tolerate is when the "miss" is due to camera error. Would you be happy with your camera if the shutter speed you selected did not work, or the selected aperture or ISO was not done properly by the camera? Of course not, so why would I accept a shot where the camera did not focus properly? I think this is the crux of the issue.

[chauncey]

C"mon Manfred, with your background in engineering you should understand the problems
with complexity. Expecting that AF system to "lock-on" every time is an exercise in futility...
too many variables in the equation.

[Manfred M]

C"mon Manfred, with your background in engineering you should understand the problems
with complexity. Expecting that AF system to "lock-on" every time is an exercise in futility...
too many variables in the equation.

I don't have an issue with not locking; but I do have an issue with not undertanding why I don't get a good lock and without a good workaround.

I think there is a parallel with the built-in light meter; the light meter will "miss" under specific conditions; but I understand that and can dial in exposure compensation to do something about it. There is no such "feauture" on a modern camera; the focus indicator (which is not something all cameras have) is not great for shooting a moving target and the focus "throw" on a modern lens is so short that fast, manual focus is all but impossible. The old pre-AF focusing screens at least gave you a fighting chance to do a fast and accurate manual focus.

I certainly do understand the technical complexities of an automated focusing system. What bothers me is the half-hearted work-around that a modern DSLR has. There is an almost grudging view that manual focus should not be needed; and while this may be correct for consumer level cameras and lenses; it bothers me that this view has also impacted the design of the pro glass and camera bodies.

[ajohnw]

Focus by wire cures the focus throw problem Manfred so manual focusing is easy on m 4/3 - if there was enough pixels in the electronic view finder. Maybe too fine manually really. Perhaps in some cases mounts and compatibility issues prevent it's use elsewhere.

The focusing has it's interesting aspects. Phase Difference Focusing that I noticed some started using recently is a complete misnomer. It's just a conventional split focus range finder arrangement with a sensor on the end.. These have known F ratio of the beams used accuracy limitations. That's why I suggested holes in a card earlier.

What I sometimes feel is that continuous is not as precise as single AF. One is designed to be agile and may include predictive aspects and the other isn't. Given the nature of control loops it's very likely that one will be more precise than the other. Prediction is fine but is anticipating ahead so if it doesn't happen there will be a lag before it corrects.

I remember a comment on mirrorless from some one - I want a precise focus not the fastest one. Well in real terms they all race to have the fastest AF what ever type of camera it is and manufacturers are well known for concentrating on numbers and costs rather than practicalities. Cameras probably come into it as well. I'm reminded of Colin's rock solid focus on 1D* comment. Nikon's equivalent is the D4s not the D80*. I suppose people might think that camera manufacturers wouldn't do things that don't work out - I wonder how many were disappointed when compacts switched to contrast focusing to save a few pence. Actually on the ones I picked up in a shop and quickly put down were by Nikon but the others were not much better so I put them all down.

Perish the thought but they might employ control engineers - as I am from the automotive industry I would wonder about that. Hope there aren't any around on here. If so I am out.

There is also a manufacturing philosophy that loosely goes sell early and sort out problems in the field. Doesn't seem to be that popular in the camera world. Another - don't do it all or that well so that a new model can be introduced later. Otherwise why would some one buy the new model? Often parts /facilities from superior cameras drift down a model range. Easy to see why they do that but then the superior model is likely improve further at some point. The real pro cameras though remain at very high prices so I wont be buying one to see if the AF really is rock solid.

John
-

[NorthernFocus]

...Expecting that AF system to "lock-on" every time is an exercise in futility...

Surely not. I am confident that the camera "locks on" very nearly every time. If it's not "hunting" then by definition it is locked on to something. The issue at hand is an understanding of what it decides to lock on to and why when used in any of the multipoint modes. That is where the variability comes in that is perplexing.

[NorthernFocus]

...The real pro cameras though remain at very high prices so I wont be buying one to see if the AF really is rock solid.

Per my last post, the issue that this thread is aimed at isn't the accuracy nor how "rock solid" focus lock is on the system. The issue is understanding the behavior of the system when used in multipoint focus mode. I'm perfectly happy with every body that I own when shooting in single point mode. They are all fast, accurate, and repeatable in single point mode (assuming they've been tuned to a given lens).

[ajohnw]

Playing with my D7000 the multipoint AF seems to be pretty predictable in SF mode. I'll use the word contrast.

If the central AF point is over something contrasty the 1st press of the shutter button will focus on that one. Press again and it will choose something else if it can generally still towards the centre. Press again and it will move further out - if it can.

Behind that is a set up that looks at the lot and chooses the strongest focusing signal but it's still definitely biased to use the central one if there is something their.

I just pointed the camera at a scene that definitely had a lot of possible strong focus points and kept pressing the button. Sort of camera sanity check. I must have a reason for keeping on pressing the shutter button. Scenes are often not that simple but it should try to focus as above and there will probably be some sort of time out where it goes back to the beginning again.

In CF mode I would suspect that the camera is perfectly happy about suddenly changing AF points to help with panning and there may also be a mode that allows focus to be lost for some period of time in case something is briefly obscuring the view. That is a mode on an E-M1 but could be permanent on other. The other aspect is that CF is very likely to be predictive. If for some reason that goes wrong there will be a delay before it gets things right again. In this case it might choose another AF point because as far as it's concerned the one it was using didn't work out and lost focus.

Contrasty - in principle if there is no features for it to range find on it can't focus.

Tell you something - it can be very difficult to decide what to do in software based on the information that is available especially if something doesn't totally work out or fails. Take the no shot taken till focus is locked on my EM's - keep the button pressed long enough in SF and it will take the shot in or out of focus. I wouldn't be at all surprised if it did that one hell of a lot quicker in CF as it would be the logical thing to do if focus had been acquired recently but at some point it would also be sensible to focus again.

John
-

[NorthernFocus]

...Tell you something - it can be very difficult to decide what to do in software based on the information that is available..

Ergo, the only way to get predictable AF performance seems to be using a single focus point with tracking turned off.