Focus Speed

<p> After all the fun with requests for a 'Fast Focussing' lens on recent threads, it occured to me again that there is no standard for this. WHY?? We all know this depends on such things as Camera Body, Lens Aperture, AF-S or Screw-drive, available light etc, but it CANNOT be that hard to get something better than 'faster than a Nikon 200mm f4 or slower than a Sigma 50mm 1.4 HSM'.<br>

I'm suprised the camera manufacturing industry hasn't got hold of this, it just loves numbers to sell us stuff. Shutter delay of 0.01 seconds, 9fps, 24million pixels, 16 bit RAW, x25 zoom, 128000 ISO etc etc.. How about time to lock-on from infinity to minimum focus at an ISO Mk IIb Target under 1000 LUX daylight balanced illumination? This should be lens specific and not reliant on the focussing module.<br>

The Auto industry has (been) organised the specifications for 'standardised' fuel economy (you've all heard of the Urban Cycle??), acceleration, top speed etc and there can be no less competition between eg Ford, BMW or Fiat.<br>

The computer industry is surrounded by Benchmarks for EVERYTHING, CPUs, Mobos, RAM, HDDs etc etc and they're even more competitive for sales.. AMD v INTEL or ATi v Nvidia anyone?<br>

There will no doubt be some tinkering by Manufacturer X's lens to work better on the tests than 'real-life' but they'd soon be ratted out.<br>

What do you guys think??</p>

<p>It probably would be helpful. As you said, they have numbers for everything else. My guess is that there are so many facors affecting focus speed that it would be hard to have one measure that would apply to every camera and every situation.</p>

<p>Good points and I mostly agree. Lens focusing "speed" is among the most nebulous concepts around and obviously subject the vagaries of web lore and anecdotes. In many anecdotes I've read regarding perceptions of autofocusing speed, uses are attributing perceived problems to a lens or "the camera" in general when it's clear they don't understand how to fully take advantage of the various options available with their equipment, and in some cases don't grasp some of the fundamentals of photography.</p>

<p>However...</p>

<blockquote>

<p>This should be lens specific and not reliant on the focussing module.</p>

</blockquote>

<p>...would not be practical. It would be akin to determining a films ISO speed independently of a developer. Instead, a standard based on a specific camera or cameras might be feasible, or a testing device developed from existing technology that would satisfy the intended purpose for years to come.</p>

<p>And it's worth noting that the computer industry has changed benchmark tests over the years several times. So it may not be practical to develop something like an ISO standard for evaluating the potential or optimal autofocusing speed for a lens without having to revise the benchmarks every few years.</p>

<p>Should we invent a standard? Since all measurements are arbitrary, we could simply say, (for our purposes), the standard is:<br>

AF-D 50 1.4 - since everyone and their cousin has one, or should.<br>

-------<br>

And then develop a testing methodology that mimics likely photographic scenarios, something like:<br>

* Target is a 3d object with a high degree of contrast against a neutral background.<br>

* Camera is 10 feet from target.<br>

* lighting is natural sunlight & tungsten indoors. (might need to be more exact on lighting conditions such as ambient light meter reading suggests 60 f4 ISO 400) but for "forum reviews" telling folks to test during daytime, in a room with windows with the lights on, might be sufficient.<br>

* Camera is set to point focus, with focus point on the object.<br>

* Scenario 1: set lens to minimum focus, and focus on object.<br>

* Scenario 2: set lens to maximum focus distance and focus on object<br>

* Scenario 3: Set lens to middle focus distance and focus on object.<br>

* with lens focused on object (from prior test), move the object 3 feet backwards refocus<br>

* with lens focused on object and object returned to netural test position (10' from camera) move the object 3 feet closer to lens, and refocus.<br>

------------<br>

Against a known baseline, even arbitrary measures if consistent provide value.</p>

 

<p>As noted, auto focusing is so situation dependent that it would be hard for it to be measurable. On the other hand, it is extremely important, so I usually just turn off auto focus and focus manually. I have had way too many shots blown by too slow, wrong, or hesitant machine focusing. I'd rather make my own mistakes. But, Catch - 22, most auto focus lenses aren't really manual focus friendly; lacking visual focus scale, focus past infinity, etc.<br>

In the early days of auto cameras, it was just ignored by makers and photo magazines, like a red headed step child. Still true, to a large extent. Don't get me started, this is my favorite rant. It's the biggest "emperor ain't got no clothes" issue in photography.</p>

<p>help me out here... I understand the "situational dependencies" of auto-focus, or at least I think I do.<br>

if we were to devise a test for lens auto-focus speed under "normal" (read: not ideal) conditions, and lens A was 30% faster than Lens B. Would 'Lens A' not continue to be faster under say difficult auto-focus conditions (like low light). (although perhaps not 30% faster).<br>

---------<br>

Lets rewind a little bit... what factors do we suppose influence auto-focus speed the most:<br>

* Subject Contrast<br>

* Lighting<br>

* Subject Movement<br>

* ???</p>