Minium AF aperture question

sirch

Lu-Tze
Admin
Messages
105,879
Name
The other Chris
Edit My Images
Yes
Quick question and confirmation of my understanding; Canon say that the minimum aperture on my 6D for auto focus is f5.6 and I assume that this is because of depth of field but what happens if I use a zoom with f6 or a tele-converter that reduces the aperture? I have tried it and it seems to at least try to AF but is it just that it won't be accurate?
 
Quick question and confirmation of my understanding; Canon say that the minimum aperture on my 6D for auto focus is f5.6 and I assume that this is because of depth of field but what happens if I use a zoom with f6 or a tele-converter that reduces the aperture? I have tried it and it seems to at least try to AF but is it just that it won't be accurate?
What it means is that with slow lenses your camera will struggle to focus because of the lack of light. Some slow lenses report a faster aperture and your camera will try to focus. Other lenses will report correctly and the camera won't attempt it.
 
Thanks Phil but the thing I don't get is that the camera is happy to AF in very low light conditions, for example indoors with just a tungsten bulb so why wouldn't it focus with an f6 lens outside in bright sunlight? Obviously there is a limit at which there is too little light but if it was only light related then would the manufacturer not specify a light level rather than an aperture?
 
Presumably they work on an almost worst case scenario so expect their bodies to be reasonably reliable at AF with an aperture of f/5.6 in light that users can see by or that the AF assist light can illuminate. You might find that a body that claims to need f/5.6 to AF can actually manage it with smaller apertures as long as it doesn't know. The examples that spring to mind are the long Sigma and Tamron lenses which are f/6.3 at the long end and also some 3rd party teleconverters. I have a feeling that these fib to the bodies and report as being f/5.6 so the bodies try their hardest rather than saying "Computer says 'no'."

Hopefully Hoppy or similar will be along and tell us better than I can!
 
I believe it is to do with the cone of light received by the AF sensor. If this is designed for f5.6, but the aperture is f6, the cone will be narrower and AF will struggle. Most cameras switch AF off at this point rather than let it keep trying. It doesn't mean the lens won't focus just that it isn't allowed to by the camera. Some lenses at f6.3 fool the camera by reporting f5.6, because mostly in good light focus will be achieved anyway.
 
There is a difference between specifications and reality, unless the system intentionally cripples the AF above a certain aperture, which I don't think they do, then the amount of light does make a difference.

However, I recently learned something from @HoppyUK (hope I got it right) which is that phase detect AF works, not from 'amount of light but from comparing images from either side of the frame and trying to make them align. If the distance between these isn't enough then it can't detect the differences in the images well enough and THIS is really why aperture is a limiting factor (as it reduces the difference in the images). Of course, through experience we know that light does make a difference but even in the sahara there is a limitation driven by the aperture and the way phase detect works.
 
Thanks Phil but the thing I don't get is that the camera is happy to AF in very low light conditions, for example indoors with just a tungsten bulb so why wouldn't it focus with an f6 lens outside in bright sunlight? Obviously there is a limit at which there is too little light but if it was only light related then would the manufacturer not specify a light level rather than an aperture?
What they said ^

Though I did mention it. The camera has to have a measurable performance, so if it allowed focus of f8 lenses, then you tried to use it in the dark and it wouldn't lock, you slag it off as incapable. To stop that happening the camera switches off the AF.

it can be fooled by lenses that mis report (or more often teleconverters that don't report their presence) and that allows the camera to try and often succeed in focussing, but when they fail, there's generally a Hoo-ha.
 
It is all about how the focusing process works.
If you put some black tape across the centre of a lens on a plate camera , so that the only light getting through is coming from the extreme sides of the lens, then check the ground glass, instead of the usual outof focus image you will see two images some small distant apart. As your rack the focus you will notice that the images come together and at that point the image is in focus.
Some cameras took advantage of this phenomenon by putting small prisims in their focus screens that took the light rays coming from the two sides of the lens.These became common in all sorts of reflex cameras.

However such an arrangement needs the lens to be significantly wide open to have any accuracy. Small apertures do not work.

Much the same idea but at a pixel level is used on some mirrorless cameras.
most Dslrs do it by reflecting part of the light on to a separate focusing unit below the mirror. But work in much the same way, but by using the more accurate phase difference in the two parts of the beam.

The critical factor is still the minimum aperture width that can make the system work. It is more a problem of geometry than light levels. Though poor illumination also worsens the accuracy.
 
Last edited:
Thanks guys, so one way or another, even if the camera attempts to focus at smaller apertures it may not be accurate which is, I suppose what I wanted know. As Nod said above I was thinking about the Tamron at 600/f6.3.
 
Thanks guys, so one way or another, even if the camera attempts to focus at smaller apertures it may not be accurate which is, I suppose what I wanted know. As Nod said above I was thinking about the Tamron at 600/f6.3.


f/6.3 is easily within usability of pretty much all modern cameras.
 
My Fuji 55-230mm has no problems focusing at F6.7 at its tele end :)

The Hybrid systems used in Fuji csc's have both contrast and phase detection. They do not have the same problems as Dslr 's
 
As mentioned by some posters above, conventional DSLR phase-detect AF needs a certain size of physical aperture diameter (as seen by AF sensors, through the lens) to work reliably. It's basically comparing images from either side of the lens and the more separated and different they are when out of focus, the better it is able to confirm an in focus image. It is actually very sensitive and can work well in low light, especially if you give it a nice contrasty area of the subject to lock on to. On the other hand, contrast-detect AF (eg in live view with most DSLRs) is immune to aperture size, but doesn't work so well in low light.

With Canon, phase-detect AF switches out if the aperture rises above f/5.6, but some third-party lenses that are f/6.3 at the long end fool the system to keep it working, and as a result the camera will try to AF. Results vary - some cameras cope well, others less so. The same happens with teleconverters, with mixed results. Note than Canon does not make any lenses with a maximum aperture higher than f/5.6, although a few higher end models are AF enabled up to f/8 when using just the centre AF point (7D2, 5D3, 1DX). Nikon doesn't set a hard aperture limit though the same performance limitations apply above f/5.6 with all but the top end models. Conversely, there are minor AF performance advantages when using f/2.8 lenses or with a lower f/number - faster, more accurate.
 
Last edited:
May as well lock the thread now. @HoppyUK has provided a completely definitive explanation.
 
Sorry to drag this on, I am obviously being a bit thick but I'm still struggling a bit with exactly what is going on

It's basically comparing images from either side of the lens and the more separated and different they are when out of focus, the better it is able to confirm an in focus image.

So it is kind of depth of field related? Presumably the reason why a small aperture gives a large DoF is that there is less difference between light from either side of the lens? If I had a plate camera I would try @Terrywoodenpic's suggestion above, I am struggling to understand the concept of "images from either side of the lens" and I think seeing it in practice might help. But ho-hum, I don't have a plate camera.

Anyway, thanks all, you have answered the basic question which is that focusing may not be accurate above f5.6 and I had always wondered why people carried around great big f2.8 telephotos when most of the time they are probably shooting at higher f-stops
 
So it is kind of depth of field related? Presumably the reason why a small aperture gives a large DoF is that there is less difference between light from either side of the lens?
Yes, sort of, although it's probably more accurate to say that both phenomena are affected by aperture in similar ways: there's no direct relationship between them.

There is a really excellent first-principles description of depth of field here - https://physicssoup.wordpress.com/2012/05/18/why-does-a-small-aperture-increase-depth-of-field/ - and it clearly shows how and why a small aperture produces a larger depth of field. There is a good description of how phase-detect autofocus works here - https://photographylife.com/how-phase-detection-autofocus-works - and whilst it doesn't say explicitly why a small aperture would be less effective, it's quite apparent if you've read the depth of field article first.
I had always wondered why people carried around great big f2.8 telephotos when most of the time they are probably shooting at higher f-stops
And the reason, of course, is that the autofocus is performed when the lens is wide open. So shooting at f/8 on an f/2.8 lens will give you better AF than shooting at f/8 on an f/5.6 lens.
 
Last edited:
Sorry to drag this on, I am obviously being a bit thick but I'm still struggling a bit with exactly what is going on


So it is kind of depth of field related? Presumably the reason why a small aperture gives a large DoF is that there is less difference between light from either side of the lens? If I had a plate camera I would try @Terrywoodenpic's suggestion above, I am struggling to understand the concept of "images from either side of the lens" and I think seeing it in practice might help. But ho-hum, I don't have a plate camera.

Anyway, thanks all, you have answered the basic question which is that focusing may not be accurate above f5.6 and I had always wondered why people carried around great big f2.8 telephotos when most of the time they are probably shooting at higher f-stops

Kind of, though it's depth-of-focus related, not depth-of-field - it has to be, because the AF sensors are behind the lens. Depth-of-focus is the zone of acceptable sharpness in front and behind the focal plane (sensor).

As an aside, we don't really worry about depth-of-focus as you can't change it. It's a very important consideration for large format film cameras where you can move both the lens and the film planes and it also comes into play when using tilt-shift lenses on digital. It's affected by the same parameters as depth-of-field, but differently. Eg, greater focal length = more depth-of-focus, closer distance = more, lower f/number = less. Compared to depth-of-field, depth-of-focus is miniscule, measured in fractions of a mm with digital cameras.
 
@HoppyUK many thanks :notworthy: there is so much to learn and the more I learn the more questions I have, thanks for taking the time.
 
@HoppyUK many thanks :notworthy: there is so much to learn and the more I learn the more questions I have, thanks for taking the time.

Over time, the more you find out and learn, the more connections you can make, and the more you understand.
 
Back
Top