Modern Digital Cameras and Aperture's Function ... a question...

[Abitconfused]

It is my understanding that most modern DSLR lenses have an aperture that remains at its largest possible diameter (the smallest f-number of the lens) until the shutter is fully depressed. So the camera is able to use all focus points available at the largest aperture of the lens. When the shutter is fully depressed, the aperture closes to that determined by the user (Aperture priority) or as otherwise configured. Is my understanding correct?

[chauncey]

All of the focus points are utilized regardless of the f/stop used...is that what you're asking?

[ajohnw]

It is my understanding that most modern DSLR lenses have an aperture that remains at its largest possible diameter (the smallest f-number of the lens) until the shutter is fully depressed. So the camera is able to use all focus points available at the largest aperture of the lens. When the shutter is fully depressed, the aperture closes to that determined by the user (Aperture priority) or as otherwise configured. Is my understanding correct?

Yes but it may choose to prioritise which AF point to use. Basically using the fastest aperture provides most light for the AF sensors.

John
-

[royphot]

Hi.

The iris staying fully open until the shutter is released has been an SLR function since about the mid 1960s. In those old days it was done mechanically, and its purpose was to make sharp focussing easier.

Roy

[DanK]

So the camera is able to use all focus points available at the largest aperture of the lens.

Closing the aperture does not obscure any of the AF points.

[Abitconfused]

If the choice of aperture does not obscure AF points, than why is it that telephoto lenses with fixed f-numbers in the f/8 range are supposed to be more difficult to focus?

[ajohnw]

If the choice of aperture does not obscure AF points, than why is it that telephoto lenses with fixed f-numbers in the f/8 range are supposed to be more difficult to focus?

The sensors require a certain amount of light to function. Many AF's have been aimed at F5.6 and a certain amount of ambient light. Some cameras now have been aimed at F8 at the same light levels. If you try to use an F8 lens on a camera intended to be used with at at least an F5.6 maximum aperture it just means that more ambient light will be needed for the AF to function correctly. F8 to F5.6 is I stop / exposure value or EV so that much extra light is needed.

John
-

[Manfred M]

If the choice of aperture does not obscure AF points, than why is it that telephoto lenses with fixed f-numbers in the f/8 range are supposed to be more difficult to focus?

Your camera's autofocus mechanism needs sufficient light to perform the autofocus funtion, so lenses with smaller maximum apertures will not necessarily work reliably. Frankly, in dim light, even a fast lens does not autofocus reliably.

My D90 user manual states that it needs a maximum aperture of f/5.6, but I have successfully shot images with a lens that at 500mm had a maximum aperture of f/6.3.

[Glenn NK]

Ed:

Your understanding is correct and both John and Manfred clarified why.

There are some long lenses that won't focus particularly well with a 2X extender (which doubles the focal length and halves the f/stop, thereby letting in less light for the focusing system).

Glenn

[Inkanyezi]

It is easier to understand why there is a minimum aperture that will work for AF if we consider what "phase detection" stands for in this respect, because it has nothing whatsoever to do with the phase of electromagnetic waves. It is exactly the same function that we had in the split image focus detection on the screen and also the microprisms worked the same way.

The AF sensors have two or more sets of receptors that capture light coming from the lens. One set captures light coming from one side, the other captures light from the other side of the lens. Hence the images formed from these receptors will coincide if the lens is focused, but if out of focus, the image from one will be at the side of the image of the other. When the system can identify objects in those images, it can calculate how much to adjust the lens to find focus.

There are a few ways that this system will not work well. One of them is too soft image without edges, another is repetitive patterns, and one is when there is no image from one set of receptors in the AF sensor.

The cross-eyed sensor is aimed at two points at the lens, supposed to be at its exit pupil and about f/11 from the centre, one to the right, the other to the left, or one up and the other down. F/11 from the centre to both sides is f/5,6. If the diaphragm is closed down to a smaller aperture, the cross-eyed AF sensors will not see an image, but only the back-side of the diaphragm. Therefore, they cannot focus.

Actually, there is some latitude, and the peripheral sensors lose one image before the other when stopping down, and the centre AF point can often focus even at f/8, but closing the diaphragm more disables it, and it is not because of too little light. It is because there is no light at all reaching the AF sensors from the subject that you're aiming at. The sensors will only see the diaphragm blades when the aperture is too small.

[jcuknz]

These limitations do not appear to apply to contrast detection AF systems which seem to work as well and as fast whatever the aperture or light level.

Made an interesting discovery recently, or a penny dropped .... working in Aperture Priority a calibrated iris is not required as the camera accepts whatever 'hole' it is working through and sets a shutter speed accordingly .... result of purchasing a 'cheap' "TV lens" with no calibrations of the iris.

[Manfred M]

These limitations do not appear to apply to contrast detection AF systems which seem to work as well and as fast whatever the aperture or light level.

While I can't comment about the minimum maxiumum aperture requirements for contrast detect systems, with my cameras, they fail when light levels are low, just like the phase detect ones. There are shots that I've illuminated the subject I was trying to focus on with an external light source.

I look at my video camera that uses contrast detect and it uses the working aperture (i.e. it is stopped down") to adjust focus, so I expect that you are correct that this system requires less light than phase detect.

As for the functioning of the iris, I had noticed that as well when I shoot with non-CPU lenses. Light metering seems to work just fine, in fact I suspect this means the trasmitted light (T-value) is measured rather than the calculated f-stop is actually used to calculate exposures.

[ajohnw]

I've just been trying out focus peaking with a 500mm F8 mirror lens in P mode. Haven't tried adjusting the P selection for obvious reasons but exposures are coming out fine. That's on an E-M5. On one of the Pen's I did have to select another mode or maybe just I thought I had to.

There have been comments about needing cpu type lenses for accurate exposure but I haven't seen any signs of that on any camera I have used.

John
-

[Manfred M]

There have been comments about needing cpu type lenses for accurate exposure but I haven't seen any signs of that on any camera I have used.

Depends on the camera; my D90 will only meter with a CPU lens and my D800 handles any lens without a problem.

As with anything else, the camera manufacturers add features based on the price point, even though these are generally just handled in firmware

[William W]

If the choice of aperture does not obscure AF points, than why is it that telephoto lenses with fixed f-numbers in the f/8 range are supposed to be more difficult to [AUTO] focus?

Exactly what Urban wrote.

WW

***

Specifically (in precise) -

1. “The AF sensors have two or more sets of receptors that capture light coming from the lens.


2. “One set captures light coming from one side, the other captures light from the other side of the lens.


3. “If the diaphragm is closed down to a smaller aperture, the cross-eyed AF sensors will not see an image, but only the back-side of the diaphragm. Therefore, they cannot focus.”

[ajohnw]

Exactly what Urban wrote.

WW

***

Specifically (in precise) -

1. “The AF sensors have two or more sets of receptors that capture light coming from the lens.


2. “One set captures light coming from one side, the other captures light from the other side of the lens.


3. “If the diaphragm is closed down to a smaller aperture, the cross-eyed AF sensors will not see an image, but only the back-side of the diaphragm. Therefore, they cannot focus.”

I think this shows signs of a fundamental miss understanding of how irises work. Also that if a camera had this problem as outlined it would have a rather serious miss focusing problem.

John
-

[William W]

I think this shows signs of a fundamental miss understanding of how irises work. Also that if a camera had this problem as outlined it would have a rather serious miss focusing problem.

Well, it was a simplified answer and (assumed understood as per the original question as asked) referring to the maximum aperture of the lens and not the iris blades.

You might care to expand on how Phase Detection Auto Focus actually works . . . below is a bit more detailed, but the answer, to satisfy the OP's question - is still basically the same.

***

If the choice of aperture does not obscure AF points, than why is it that telephoto lenses with fixed f-numbers in the f/8 range are supposed to be more difficult to focus?

It's really nothing to do with the amount of light as one can get AF at f5.6 in poor light but not get AF at f8 in bright light.

It is actually complicated geometry - BUT - one can think of it as being similar to the case of split screen focusing of manual cameras, (which is the analogy that Urban referred to) and where it is a common observation that the split screen section ‘blacks out’, when one stopped a lens down.

Many (most?) DSLRs user SIR (Secondary Image Registration) TTL Passive Phase Detection.

This is basically a beam splitter which uses the light from opposite sides of the aperture.

It's looking for the overlap of the images from both sides of the aperture - and if the angle between the paths of those images is too small (f stop too small), it's hard to tell when the images are in or out of registration (i.e. focus shift doesn't make much difference).

The diagram below is indicative and not geometrically accurate.

It is not detailed in regard to the workings of shift type AF system – but hopefully it is an aid to the explanation above, showing how a smaller max aperture lens, provides less differentiation, for the function of the AF system.





Typical Cross-section View: DSLR with Passive Phase Detection AF system


LEGEND:

1. Lenses of two different max apertures – open
2. Main Mirror position during viewing
3. Main Mirror position during exposure
4. Focusing Screen
5. Pentaprism
6. Viewfinder
7. Digital Sensor
8. Sensor Filter
9. Shutter
10 AF Mirror
11 AUTOFOCUS SYSTEM


WW

[ajohnw]

I thought about posting when obscuring was mentioned but time etc. My understanding of how phase af works is like this taken from the wiki



It doesn't matter how the split image merges and gets to the af sensors but the point is that it must co inside with the focus on both the focusing screen and the sensor when the mirror is out of the way and the shot is taken. The iris doesn't change the angle of view or the size of the image and the focal length - in principle - is always the same. What I am saying in effect is that if the iris obscured it would also obscure the sensor and the focusing screen.

What the typical cross section misses is the field of view is the same for both apertures so it needs the pairs of parallel rays drawing at angles that indicate the field of view. That never changes as the iris is closed. Only the light levels change. Actually many lenses loose 1 stop or more at the corners when they are wide open because the front element isn't large enough. They may skimp on internal elements as well. Close the iris 1 to 1 1/2 stops and the effect usually disappears. Vignetting.

Tried to find a decent drawing of an iris in action but couldn't find anything at all worth posting. It's easier to see on a symmetrical lens. The iris prevent rays from the out regions of the front element from contributing to the image. The size of it's aperture sets how much is allowed through.



This might help if similar plane waves are considered coming in at an angle. They would produce another point offset from the one shown at a distance proportional to the angle. Change the diameter of a lens and all that happens is that less light is gathered.

http://upload.wikimedia.org/wikipedi...wavefronts.gif

This also shows why vignetting occurs. When the rays enter at an angle they see a smaller aperture as it's eliptical so has less area and consequentially collects less light. Actually this gif is a much better way of showing how lenses work. Ray diagrams are for design purposes and the typical section would need an object drawing in and rays drawn through the system form each end of it. Often only rays from the centre point and one end would be shown. The one from the centre would go straight through without bending. The other one would go to the point where it is imaged.

John
-

[Inkanyezi]

I thought about posting when obscuring was mentioned but time etc. My understanding of how phase af works is like this taken from the wiki


-

As a matter of fact, it does NOT work that way. As I wrote before, it has nothing whatsoever to do with PHASE of wavefronts.

There are a minimum of two sensor groups in the same position in the PDAF, and these receive light coming from different parts of the exit pupil of the lens. If focused, both images from these sensors will coincide, but if out of focus, they will not.

The points where these sensor groups are aiming at are at a distance of about f/11 from the centre of the exit pupil. This means that when the lens is stopped down further than f/5.6 no light from the subject will hit the AF sensor, which will only "see" the diaphragm blades from the backside.

The AF sensor is squint-eyed. One part of it looks at the left side of the lens, the other looks at the right side. To work at all, there must be a left and right side, but at f/22 there isn't; only a point midbetween where the AF sensors are aimed, the very centre of the exit pupil.

The Wiki image is wrong, as it does not show the path of rays. I can do a better sketch, but I'm rather tired right now, but if you want one, I can do it tomorrow.

[ajohnw]

http://www.the-digital-picture.com/P...Explained.aspx

The problem is that the DOF varies with aperture.

John
-