PDA

View Full Version : TC's, maxi. aperture and AF



Jerry van Dijk
08-18-2012, 05:38 AM
I'm looking into TC's, because I like the versatility and lower pricetag combining a 300mm with a TC over longer glass.
I am however puzzled by the loss of AF with certain combinations of lenses and TC's. The Nikon compatibility chart shows that lenses with max. apertures below f/4 will lose AF with TC's over 1.4x (i.e. 1.7 and 2x). Lenses with max. aperture of f/2.8 maintain AF with all TC's.
Is there any technical or optical reason for that? If so, does that also mean that the f/2.8 lenses with a TC will lose AF ability when stopped down to f/4 or smaller?
I couldn't think of a logical reason, but I do not sufficiently understand the workings of AF.
Thanks in advance for any insights.

Graeme Sheppard
08-18-2012, 06:41 AM
Firstly, your camera uses the lens fully open, e.g. F2.8, to focus.
It then shuts down the aperture to whatever value is set for the shot.
So, the limitation is irrespective of the chosen aperture size, only of the maximum possible size.

The other part I'm not so strong on.
However, your camera needs a certain amount of light to be able to focus properly. The optics of a TC necessarily limit the total amount of light hitting the focussing screen. Think of it like this: instead of a postage stamp area of light hitting the screen, the 2X enlarges that so the same amount of light hits a bigger area - the laws of physics mean that the light must be more spread out and weaker, so focussing becomes more difficult.
So, cameras need a minimum aperture size to be able to focus because they need a minimum amount of light.

In the future, with better circuitry, it is likely to improve.

I'm sure someone will correct me if I'm wrong.

Jerry van Dijk
08-18-2012, 02:34 PM
Hi Graeme,

Thanks for your response! First point is clear and the second also makes a lot of sense. However, the TC's only take away either one or two stops of light (depending on the magnification), which to me doesn't seem like an awful lot, especially in bright conditions. Stopping down one stop is not unusual as exposure compensation and generally, you don't lose AF then. Lets hear what the others chime in.

Graeme Sheppard
08-18-2012, 09:03 PM
It may not sound a lot, but reducing the aperture by one f stop reduces the area of the aperture by half.
So two f stops reduction gives you only a quarter of the light on the focussing screen.

And remember that stopping down a lens manually has no effect on AF, since AF works fully open whatever the settings are. The TC reduces the 'fully open' state.

Jerry van Dijk
08-19-2012, 06:49 AM
Thanks Graeme!


And remember that stopping down a lens manually has no effect on AF, since AF works fully open whatever the settings are. The TC reduces the 'fully open' state.

The point I was trying to make is that the amount of light taken away by a TC at maximum aperture is equivalent to stopping down a regular lens from it's maximum aperture by two f-stops (e.g. from f/4 to f/8), the latter only causing AF problems under very low light conditions.

Graeme Sheppard
08-19-2012, 07:06 AM
Thanks Graeme!



The point I was trying to make is that the amount of light taken away by a TC at maximum aperture is equivalent to stopping down a regular lens from it's maximum aperture by two f-stops (e.g. from f/4 to f/8), the latter only causing AF problems under very low light conditions.

I wonder if we're talking at cross purposes. Stopping a regular lens down by 2 f stops to f8 has no effect on AF, since the AF action will still take place at f4.

One way to see this is to set a very small aperture, look through the camera and press the 'aperture preview' button, if you have one. It will cause the lens to switch from its default fully open state to the set value whilst you hold it down.

Another illuminating (haha) demo is to set your camera to a 10s exposure. Set it to various f values and look at the lens as you focus and click the shutter release. This will show you how during AF the aperture stays open, then switches to the set size for 10s.

Hope that helps. A lengthy answer as it helps remind me too.

Jerry van Dijk
08-19-2012, 02:33 PM
I wonder if we're talking at cross purposes
No, I just completely misunderstood, sorry :e3. Thanks for taking the time to explain again .
My only remaining questions would be why the AF system needs a lot of light, and what it is that determines the minimum amount needed.

Charles Glatzer
08-19-2012, 11:41 PM
If you start with a 300 f/2.8, adding a 1.4x makes the widest aperture f/4, a 2x f/8.
The Nikon D3/4 series and D800 cameras will AF at f/8, and all models will AF at f/5.6 being max aperture.

What camera are you using?

Chas

Jerry van Dijk
08-20-2012, 03:28 AM
Hi Chas,

I'm using the Nikon D7000. I was planning on switching my 80-400 VR for a 300 f/4 with TC once the VR version comes out (Nikon already applied for the patents), but if I lose AF, that would no longer be an option and I need to go save for the 300 f/2.8. I know that the 1.4x TC will give me AF and and equivalent focal length of the 400, but I was hoping on adding the 2x to get to a 600mm equivalent. The max. aperture would then be f/8. The D7000 is in another league than the D3 and D800 (which are both FF), so I fear the worst.

Other than weighing my options on the gear, I'm just wondering about the why and how of AF and light availability.

Alan Lillich
08-30-2012, 10:17 PM
Jerry,

Phase detect AF changes from TCs have nothing to do with the reduction in the amount of light. Think about performance with a given lens in full sun versus moderate shade. A lot more than 1 or 2 stops of difference there, yet the focus isn't affected. The amount of light only plays a role when it is so dim that the electronics can't get a decent signal.

The max aperture issue, whether from a TC or simply a really slow lens, is about the physical width of the lens aperture. Phase detect AF works like binocular vision. It compares slices of the light coming from opposite sides of the lens. The farther apart those slices are, the better it can judge distance. Too close and it can't make up its mind.

Roger Cicala has a nice article here: http://www.lensrentals.com/blog/2010/07/how-autofocus-often-works

Alan

Jerry van Dijk
08-31-2012, 03:44 AM
Thanks Alan, that is very insightful!

John Chardine
08-31-2012, 05:46 AM
If you start with a 300/2.8 and add a 2x tc you get an f5.6 lens (f2.8>f4>f5.6).

Jerry van Dijk
09-02-2012, 03:17 PM
If you start with a 300/2.8 and add a 2x tc you get an f5.6 lens (f2.8>f4>f5.6).
Hi John, that's correct and the f/5.6 won't cause any trouble for the AF, but the f/4 which I was planning to get will be f/8 iwith the 2x TC, which apparently is a problem for AF.

Roger Clark
09-08-2012, 09:36 AM
Jerry,

Phase detect AF changes from TCs have nothing to do with the reduction in the amount of light. Think about performance with a given lens in full sun versus moderate shade. A lot more than 1 or 2 stops of difference there, yet the focus isn't affected. The amount of light only plays a role when it is so dim that the electronics can't get a decent signal.

The max aperture issue, whether from a TC or simply a really slow lens, is about the physical width of the lens aperture. Phase detect AF works like binocular vision. It compares slices of the light coming from opposite sides of the lens. The farther apart those slices are, the better it can judge distance. Too close and it can't make up its mind.

Roger Cicala has a nice article here: http://www.lensrentals.com/blog/2010/07/how-autofocus-often-works

Alan

Sorry I missed this thread--I was on travel.

Alan is correct and the cited article is pretty good (except the Figure 1 diagram is a little off: the beam splitter is not at the focus, no aperture masks are shown, and the micro lenses do not produce a perfect focus all the time as shown, but it does give an idea; Figure 2 is incorrect too, but you get the idea). Basically, for phase detect AF, there are two small apertures placed slightly out of focus (this is all in a module in the bottom of a DSLR), and the two apertures mask out most of the light from the lens, leaving light from two sides. The spread of these two apertures can vary: the further they are apart, the more accurate focus will be. But the light cone from the lens must be broad enough to get the light. So aperture masks designed for f/2.8 lenses will likely not get much, if any light from an f/8 system (e.g. f/4 lens 2x TC). The boundary is a fuzzy one, so aperture masks for f/5.6 lenses will likely work at f/8, but with reduced accuracy, and will fail as light levels fall. Manufacturers are putting software in their cameras to not do AF at f/8, probably so they don't get hammered in reviews about unreliable AF at f/8, and as a marketing tool (eg. onlt Canon 1D-series could AF at f/8). In 1D mark Iv and earlier Canon 1D-series cameras, Canon had one sensor designed for an f/8 light cone. The trick to get it to work on cameras that normally do not have f/8 focusing is to buy a 3-rd party TC that doesn't report the f/ratio change, or tape some pins to defeat reporting.

Roger

Jim Neiger
09-08-2012, 09:54 AM
Thanks Graeme!



The point I was trying to make is that the amount of light taken away by a TC at maximum aperture is equivalent to stopping down a regular lens from it's maximum aperture by two f-stops (e.g. from f/4 to f/8), the latter only causing AF problems under very low light conditions.

Jerry,

I think you missed Graeme's point. The lens focuses wide open no matter what aperture you are using to make an image. If you have an F4 lens and shoot it straight, it focuses wide open at F4. If you add a 1.4xTC, the lens still focuses wide open, but now wide open is F5.6 and it has one stop less light available to accomplish AF. If you use a 2.0x TC, the camera focuses wide open at F8 (assuming the camera can focus at F8 - some can and same can't) and has two stops less light available than the F$ lens by itself. That is way you can focus much faster with your aperture set to F8 on an F4 lens vs F8 with the lens plus 2xTC. The straight lens usies F4 for focusing and F8 for shooting. With the 2x it focuses and shoots at F8.

Jerry van Dijk
09-10-2012, 02:24 AM
Thanks all for taking the time to explain the workings of the AF. I've learned a lot from this thread!
All that's left for me is to find out whether the D7000 has AF with f/8, or use one of the shortcuts Roger mentioned. And of course waiting for that 300 f/4 VR to come out!

Gray Fox
10-08-2012, 11:28 AM
. . . The spread of these two apertures can vary: the further they are apart, the more accurate focus will be. But the light cone from the lens must be broad enough to get the light. So aperture masks designed for f/2.8 lenses will likely not get much, if any light from an f/8 system (e.g. f/4 lens 2x TC). The boundary is a fuzzy one, so aperture masks for f/5.6 lenses will likely work at f/8, but with reduced accuracy, and will fail as light levels fall.

Following up on the above relative to the 1DX AF system and Canon’s lockout of AF with f8 lenses (due presumably to less reliable performance), might there be other factors in addition to the size of the aperture masks in that particular AF sensor? I’m referring specifically to the fact that the 1DX (and 5DIII) AF sensor system has cross-type sensor elements rather far out toward the edges of the sensor array, as well as a wider array than previous 1D series cameras. One explanation given for Canon’s locking out AF at f8, seen on a B&H promotional video, was that Canon “spread out the light” coming to the AF array in order to enable more cross-type sensor elements out toward the edge of the array, leaving less for reliable focusing at f8. This was stated by a non-technical representative and might have been marketing gibberish. But it does lead one to ask if and how Canon might have manipulated the light impinging on the AF sensor.

Something is certainly different technically. The 1DX and D4 arrays are roughly of equal width and height, although the 1DX has a slightly higher density of elements in the central column. But, the Nikon D4 has several sensor elements that AF at f8, including a central cross-type element. However, only the sensor elements in the central section of the D4 operate as cross-type sensors at f5.6 and wider apertures. This concentration of cross-type sensing elements in the center of the D4 array contrasts markedly from the Canon implementation, and it does make one want to know more about differences in the way light is handled by the two sensor systems.

Roger Clark
10-08-2012, 01:15 PM
Following up on the above relative to the 1DX AF system and Canon’s lockout of AF with f8 lenses (due presumably to less reliable performance), might there be other factors in addition to the size of the aperture masks in that particular AF sensor? I’m referring specifically to the fact that the 1DX (and 5DIII) AF sensor system has cross-type sensor elements rather far out toward the edges of the sensor array, as well as a wider array than previous 1D series cameras. One explanation given for Canon’s locking out AF at f8, seen on a B&H promotional video, was that Canon “spread out the light” coming to the AF array in order to enable more cross-type sensor elements out toward the edge of the array, leaving less for reliable focusing at f8. This was stated by a non-technical representative and might have been marketing gibberish. But it does lead one to ask if and how Canon might have manipulated the light impinging on the AF sensor.

Something is certainly different technically. The 1DX and D4 arrays are roughly of equal width and height, although the 1DX has a slightly higher density of elements in the central column. But, the Nikon D4 has several sensor elements that AF at f8, including a central cross-type element. However, only the sensor elements in the central section of the D4 operate as cross-type sensors at f5.6 and wider apertures. This concentration of cross-type sensing elements in the center of the D4 array contrasts markedly from the Canon implementation, and it does make one want to know more about differences in the way light is handled by the two sensor systems.

Gray,
It is hard to critique the details without a good diagram of the design, which they have not supplied. I'll just say that the lock out of f/8 seems to be a marketing ploy as we've seen 3rd party TCs added to a lens and the 5D3 focuses fine at f/8. The 1D4 focuses fine at f/11 (stacked 1.4 and 2x TCs on an f/4 lens) (so did the 1DII). This is not hard technology, so it really seems odd for Canon to do such a lockout saying it is a technical limitation. We can only hope the competition from Nikon will make canon change (or we switch to Nikon).

Roger

Ulli Hoeger
10-14-2012, 10:27 AM
I looked into the whole TC and AF issue a few weeks back.
So far it was clear that AF would stop at f5.6 with my present camera bodies with my TC involved (D300, Sigma 500 4.5, and Sigma 1.4x).
With the recently released D600 and D800 and their ability to AF up to f9 I was wondering if one of those bodies would AF above lens and TC combo, as the resulting lens aperture would be f 6.3.
It didn't work!
Piecing the big picture together, that with some other brand TCs my current body/lens combo AF for some users, it seems that the TC makes the call not to permit AF if the aperture exceeds f 5.6. This must be in its chip and thus its the law, even if current and future bodies would be able to AF reasonably well with smaller f stops than 5.6.
So far it seems that some of the current Kenko converters work with my current body/lens combo, and I would expect that this would be the case with other current Nikon bodies as well. In the end it appears to be something one has to try out, and one could only hope that in the future the manufacturers will eliminate the f5.6 limit, or at least bring it up to a value that current camera bodies can do.

Ulli

P.S. If I can get my hands on a Kenko 1.4x I will give it a try.

Roger Clark
10-14-2012, 11:37 AM
Piecing the big picture together, that with some other brand TCs my current body/lens combo AF for some users, it seems that the TC makes the call not to permit AF if the aperture exceeds f 5.6. This must be in its chip and thus its the law....

Or it could be simply that the TC doesn't supply some info that the body wants, and when it doesn't get it, the body stops AF.

Roger