TCs and Auto-focus

I don't have any TC's (or even an SLR) but after reading a lot of the chatter on here about losing autofocus with a TC, I have a few questions.

Why is autofocus limited by the maximum aperture? I understand how auto-focus can be limited by light levels, but this would seem to vary shot by shot regardless of the aperture used.

Some possible follow-ons :

Even if the camera meters and focuses wide open, can you lose auto-focus just because of low light levels?

If you have an f/5.6 lens and a 2x TC, could you still possibly auto-focus in very bright light?

If you bought a lens that had a maximum aperture of f/8, would it have to be manual focus?

Even if the camera meters and focuses wide open, can you lose auto-focus just because of low light levels?

in low light, the AF becomes slower; in extremely low light it may not be able to find the subject.

If you have an f/5.6 lens and a 2x TC, could you still possibly auto-focus in very bright light?

No - at least on Canon cameras.

If you bought a lens that had a maximum aperture of f/8, would it have to be manual focus?

yes, unless you have a 1 series camera or the lens uses some "trick" (if I remember well, some years ago Sigma produced a 1000mm f/8 lens that could be used with AF because it reported a maximum aperture of f5.6 to the camera)

Thanks Juza, but that still doesn't quite answer the main question - what I'm really curious about. WHY doesn't the camera auto-focus in these situations? As you said, auto-focus slows down with light loss. Wouldn't a small aperture and bright light still be able to auto-focus? Maybe I'm looking for a more technical answer...

Hi David, basically autofocus relies on edge-contrast to establish when the image is sharp - this means that the brighter the light hitting the AF sensor(s) the more chance there will be some contrast for the AF to use to focus. In theory the camera would still be able to autofocus in bright conditions with a small max aperture but any other conditions would lead to the camera failing to get focus - hence the manufacturers actually limiting the camera from even trying. It is possible to tape the electronic pins on the lens or teleconvertor to stop it from sending the aperture information to the camera and therefore enabling autofocus...but of course, autofocus is still limited by the light getting to the autofocus sensors. HTH

Thanks Mark...that seems like what I was looking for. I couldn't figure out why the AF was enabled by the aperture and not the meter. I guess it's a matter of preventing "hunting". (Still seems like it could go off the meter somehow though.) However, wouldn't even something like f/2.8 still be a problem in low light situations? Do you guys with SLRs manually focus all those long exposure shots?

Yes, even an ultra-fast autofocus lens like Canon's 70-200/2.8 can 'hunt' in low light or low contrast, foggy conditions. So bright, soft light can challenge the AF because of the lack of contrast.

I use manual focus for pretty much all scenic and static scenes, whatever the lighting conditions.

I like to shoot in low light situations with a 2.8 lens. An example is a lighthouse - I have to put the ssensor on the lanternhouse and wait for it to blink in order to attain focus. Somtimes, if there isn't a light to focus on or nothing silhoueted by a brighter sky, the camera simply won't focus until it gets brighter as dawn approaches.

For a camera to autofocus, the image formed by the right side of the lens is compared to that of the left side (and/or the top compared to the bottom) This is analogous to the baseline of a rangefinder. If the relative aperture is too small, the AF has little to work with, and these images are minimal or non-existant. This is similar to the loss of the center split image of rangefinder screens when attempting to manually focus long, high f/# lenses.

Alex, your statement, above, is incorrect.

Baseline rangefinders compare the alignment between 2 images formed by 2 lenses separated by a fixed distance between the 2 lenses. A system of mirrors and a swiveling prism is used to bring the 2 different images into view in a single plane. An adjustment cam between the camera lens and the prism is turned such that when the 2 images coincide within the viewfinder, the camera lens is focused. Baseline rangefinders do not even use the image formed by the camera lens. The greater the distance between the 2 lenses in the rangefinder, the greater the accuracy of focus.

Autofocus sensors, on the other hand, adjust the focus of the lens so that the contrast of the subject/scene within the autofocus sensor's field of view has maximum contrast. That is why it is important that the subject/scene has adequate illumination and contrast for an autofocus sensor to work properly. When light levels fall too low or the subject/scene has too little contrast an autofocus sensor will be unable to detect enough contrast to properly adjust camera focus.

The scenario you describe, above, would only allow cameras to have just 2 autofocus sensors: one to determine the sharpness of the right side of the image and one to determine the sharpness of the left side of the image. As we all know, autofocus cameras may contain up to 45 individual autofocus points, each one selectable to focus on just the image within each autofocus sensor's limited field of view.

Please, do not mix up the functioning of a baseline rangefinder with autofocus points within an autofocus camera.

There are two different techniques for autofocussing.The contrast technique is usually reserved for digital P&S cameras and rapidly move the focus of the lens in searching for max contrast. Its advantage is that it needs no additional optics, but only microprocessing to determine the proper focus. The mid and high level of AF (digital or film)cameras need the more sophisticated phase comparison method that compares images formed by two different zones of the camera lens. This gives much speedier AF, focus prediction and tracking.
A good article explaining modern AF cameras is http://www.nikon.co.jp/main/eng/portfolio/about/technology/nikon_technology.

The complete url is http://www.nilon.co.jp/main/eng/portfolio/about/technology/nikon_technology/caf/index.htm.

Interesting article Alex. The article isn't completely clear, but I guess the multiple AF points in the camera are just different parts of the line sensor? Or maybe multiple small line sensors?

By the way, if anyone didn't notice, here's the correct link.

David--there are multiple line sensors. Some cameras have some (but not all) of their AF points as "Cross" sensors, which are two perpendicular line sensors. IIRC, these need f/2.8 to work properly.

Anyone remember the old MF cameras with a split prism focussing aid? Ever put a 2x on an f/5.6 lens with one of those? Half of the split prism would go black, regardless of how much light there was. The principle is the same.

-Paul

Thanks for all the responses everyone - this is very interesting. But it still leaves a question as to why you can't use AF in very brightly lit situations. I'm assuming that the cross sensors are arranged in some pattern. I would also think that, regardless of aperture, the sensor in the center would always receive light.

So with bright light, you could probably always use the auto-focus if you only used the center cross sensor? But they disable that intentionally regardless of light at a certain aperture since some of the cross sensors get cut off and won't be receiving any light? Or do the prisms get cut off in some way that AF is actually impossible?

It's not that light is getting cut off. I think it's more that with smaller apertures, the sensors aren't getting a wide enough range of phase difference to make a good detection. (The 1-series Canons have higher quality sensors that can work with smaller apertures.)

But note that it *is* possible to AF in bright light and good contrast scenes. In the Canon system, anyway, you can tape a couple of pins on the TC so the camera doesn't know it's there. AF will then attempt to work, but won't always be successful. (You can do a search for how to accomplish this, but note that the pins you have to tape are on the side of the TC that face the lens, not the camera.) I don't know if there's a similar trick for Nikons.

I suspect that Canon decided to make the AF not function at all with small apertures because it won't work well in many (most?) situations, and they didn't want to get complaints about the AF not working properly.

-Paul

This is very interesting. I was not aware that some manufactures actually disable their cameras from even trying to autofocus at apertures smaller than f 5.6. This explains a lot because both of my cameras will autofocus quite nicely at an effective aperture of f 8 ( f 5.6 with 1.4TC = f 8). I had heard from more than one source that Canon's won't do that and I couldn't believe that they were incapable. Now I find out that they are merely disabled on purpose.

Thanks for the info.

Bob

Robert: Canon "1" series bodies will autofocus the f8 (f5.6 + 1.4TC) combo you mentioned.