Lens not designed for CDAF

[Matijaa]

I was wondering, what does it mechanically mean when lens is designed for CDAF? Or in other words, why do legacy 4/3 lenses have slow AF on m4/3 bodies? And how slow is it exactly, a bit slower or twice as slow?

 

[Voldenuit]

I was wondering, what does it mechanically mean when lens is designed for CDAF? Or in other words, why do legacy 4/3 lenses have slow AF on m4/3 bodies? And how slow is it exactly, a bit slower or twice as slow?

Not an optical engineer, but I imagine it has to do with the way CDAF works. CDAF has to find a local maximum in contrast, and that means racking the focus back and forth to find the maximum. This means, the motors, gears and lens elements have to be optimised for this behaviour (good damping, quick reversal of focusing direction, etc).

In the good old manual focus days, it didn't matter what size or where your focusing elements were. In fact, many designs had multiple elements that moved during focus. When AF was introduced, designers were constrained to moving smaller numbers of elements at a time, and had to use lighter/smaller elements so the motors didn't get overburdened, and to preserve AF speed.

So I imagine a similar shift is occurring with CDAF-ready lenses.

I've only tried a few Oly 4/3 lenses on my GF1, most of which wouldn't focus at all. The ones that did were rather slow (compared to native MFT lenses). Possibly to protect the gears and motors in the lens?

 

[Matijaa]

I've only tried a few Oly 4/3 lenses on my GF1, most of which wouldn't focus at all. The ones that did were rather slow (compared to native MFT lenses). Possibly to protect the gears and motors in the lens?

Yeah... my guess was that PDAF lenses might have more glass to move, but I was wondering if that's all there is to it. Seeing how fast PDAF lenses focus (on their native bodies) made me think their motors are very capable of high speed moves and stops, but I'm curious to find out what exactly the differences might be.

Thanks for your thoughts!

 

[EXR]

I was wondering, what does it mechanically mean when lens is designed for CDAF? Or in other words, why do legacy 4/3 lenses have slow AF on m4/3 bodies?

Because CD-AF needs a constant measurement of contrast and constantly adjusts the focus-distance to find the maximum contrast (which equals maximum sharpness).

To start the AF-cycle the camera has to move the Focus-element in one direction, but it doesn´t know in which direction. It has a 50%-chance to guess right (i guess their are some methods to improve this with the pre-AF-feature), if it is wrong it has to change the direction. Then the contrast improves, but to find the maximum point the focus-element has to drive over it. Then the AF-processor notices the contrast-drop and moves the focus-element back to the maximum point, again it has to change the direction.

Phase-difference AF only needs one initial measurement to get the direction and how much to drive the focus element. Then it immediately gives the lens the instruction to move to the calculated position.

So to work good with CD a lens must have a strong motor, that can switch the direction very quickly. Of course a light focusing-element also helps a lot.

And how slow is it exactly, a bit slower or twice as slow?

It depends, on the lens, but twice as slow is really fast for a non-optimized lens.

You have to keep in mind, that a lot of the 4/3-lenses have been designed at least with CD-AF in mind, for the live-view-DSLRs of Olympus. But even these are usually at least 2-3x slower.

You can see what happens when you use a lens without CD in mind when you use one of the new DSLRs in liveview. Focus-times of a few seconds aren´t rare.

Maybe it is also not good for an old lens to be used with CD-AF, I can imagine that the quick changes of the direction which is needed for CD-AF will wear out the focus motor much faster.

 

[thinkfat]

CDAF also needs to know where the lens was when the local maximum was found. PDAF lenses move fast when they know where to travel to. But they don't report their positions while they move.

--
Everybody loves gadgets, until they try to make them
http://www.flickr.com/photos/thinkfat
http://thinkfat.blogspot.com

 

[Matijaa]

Thanks for your extensive reply! It's mostly what I suspected, although I can't help but wonder about experiences I had with dSLRs. For example, in low light my former EOS used to hunt focus a bit or a lot, depending on conditions. It wasn't much faster nor slower than CDAF when struggling. So the motor was certainly capable of such movement.

Also, lenses with fast AF seem very capable of quick and sudden motor movements. Even though it might just be in one direction, it would seem logical that the same fast speed could be used to hunt for maximum contrast, even if it was be reduced a bit. (Durability is a anther question which doesn't concern me in this theoretical discussion.)

So I can't help but wonder if motor is everything there is to it...

 

[Tord S Eriksson]

Bought a two Tamron lenses, both widezoom and telezoom, and both are very slow, or will give up focusing altogether on my Pentax K-x. Now I know the reason why, thanks!
--
tord (at) mindless (dot) com