Roger Cicala on Canon Autofocus Reality

[drmikeinpdx]

http://www.lensrentals.com/blog/2012/08/autofocus-reality-part-3b-canon-cameras#more-8458

I think this is a very important article. It addresses my concerns about the Canon autofocus system better than any other article I've seen and will influence my future buying decisions.

What do you think of it?

Mike in Portland

 

[Drizzt321]

A great series of articles looking at some of the rumors and facts, and showing pretty convincingly the truth of how things are right now. Best to read the entire set of articles first, for those who haven't. Links are at the top of the OPs linked article.

 

[Loswr]

Definitely a great series!

Very interesting that it's a combination of new camera and newer lens design that results in the most accurate phase AF.

Regardless, there's still a range of sharpness in a series of multiple shots, albeit narrower with the more accurate combinations. That's one reason I love the 12 fps frame rate of the 1D X - I make it a common practice to shoot in bursts of ~4 shots, and pick the sharpest in post, and with the 1D X, that takes the same about of time as a single shot on the 5DII, or two shots on the 5DIII. It also means I'm more likely to get a shot with no one blinking, etc.

As a side note, Roger's article quantitatively confirms a subjective impression I had developed when shooting with the 1D X - combined with the 40/2.8 pancake lens, it was harder to pick the 'best' of the 4-shot bursts based on sharpness, whereas with other lenses, such as the 24-105mm, it was easier because the shot-to-shot differences were greater.

...will influence my future buying decisions.

This is probably true for me, as well. I'm planning to get a supertele witin a few months (500 II or 600 II, not sure yet), but I assume the new 24-70 II will also focus more accurately, and that will make that hard to resist as a walkaround lens, which would delay the supertele purchase a bit...

 

[tron]

Definitely a great series!

Very interesting that it's a combination of new camera and newer lens design that results in the most accurate phase AF.

Regardless, there's still a range of sharpness in a series of multiple shots, albeit narrower with the more accurate combinations. That's one reason I love the 12 fps frame rate of the 1D X - I make it a common practice to shoot in bursts of ~4 shots, and pick the sharpest in post, and with the 1D X, that takes the same about of time as a single shot on the 5DII, or two shots on the 5DIII. It also means I'm more likely to get a shot with no one blinking, etc.

As a side note, Roger's article quantitatively confirms a subjective impression I had developed when shooting with the 1D X - combined with the 40/2.8 pancake lens, it was harder to pick the 'best' of the 4-shot bursts based on sharpness, whereas with other lenses, such as the 24-105mm, it was easier because the shot-to-shot differences were greater.

...will influence my future buying decisions.

This is probably true for me, as well. I'm planning to get a supertele witin a few months (500 II or 600 II, not sure yet), but I assume the new 24-70 II will also focus more accurately, and that will make that hard to resist as a walkaround lens, which would delay the supertele purchase a bit...

May I suggest to wait (if possible) in order to increase the possibility that the new white super tele lenses will come already updated with the new 1.1.1 firmware? This may be important for 1Dx owners. On the other hand you could buy the most expensive lens first (and optionally have it upgraded) because you are always running the risk of finding something interesting and postponing the purchase of the super tele lens.

 

[Jamesy]

Rogers findings surprise me a bit with respect to the 70-200 f2.8 IS II:

"The Canon 70-200 f2.8 IS II, oddly enough, seemed not quite as good as the above lenses, but more accurate than the older ones. Whether this was oddness in my measurements or a real finding, I wasn’t sure."

I would have thought this was one of the holy grail lenses in terms of Engineering prowess.

 

[drmikeinpdx]

This series of articles goes a long way to explaining why the autofocus system in Canon DSLRs is so mysterious and unpredictable. Both the body and the lens have to be in perfect calibration, unless they both happen to be off in a way that cancels out the error, I guess. But then there are issues of focus shift within the zoom range too.

I'm in the market for a new full frame camera to replace my 5D classic. It will be interesting to find out if the rumored entry-level FF camera will have the new focus system like the 5D Mark III and the 1DX.

A year ago I was tearing my hair out trying to find a zoom lens that would work well with my 5D Classic, when even my prime lenses would not focus as predictably as I wanted. Then I picked up a cheap and very well used Tamron 28-75 which happens to focus perfectly. Very sharp lens too. I was shocked. But now I'm a happy camper and can afford to wait while the new FF models are introduced and tested.

 

[Jamesy]

A year ago I was tearing my hair out trying to find a zoom lens that would work well with my 5D Classic, when even my prime lenses would not focus as predictably as I wanted. Then I picked up a cheap and very well used Tamron 28-75 which happens to focus perfectly. Very sharp lens too. I was shocked. But now I'm a happy camper and can afford to wait while the new FF models are introduced and tested.

A good buddy of mine shoots the 5Dc/28-75 combo and he loves it.

 

[Tcapp]

http://www.lensrentals.com/blog/2012/08/autofocus-reality-part-3b-canon-cameras#more-8458

I think this is a very important article. It addresses my concerns about the Canon autofocus system better than any other article I've seen and will influence my future buying decisions.

What do you think of it?

Mike in Portland

Hey! A fellow portlander! Wooot!

 

[peederj]

The article made me more interested in the 24-70/2.8 II and less interested in the 70-200/2.8 II.

I suppose they know only the super-enthusiast would read such articles (and therefore salivate over the new lenses) or else Roger would by now be photographing the fishes...

 

[Jamesy]

I too am more intrigued by the newer glass now to match my 5D3. It appears that the addition of the rotation sensor is the key piece of technology that has been introduced to increase accuracy. It seems that the newest lenses are as accurate with PDAF as with CDAF.

I don't know why the fact that the newer glass is better on the newer bodies would be a secret, like most things technical there are always improvements and advances.

I am bullish again about the 24-70 Mk.II after having read the article. I use the 24-105 as a walk-around and have not been all that happy with the results. It is going into to Canon on Monday along with the 5D3 (light leak repair) it came with to see if it is in fact the lens or just me.

 

[peederj]

It's secret because no one wants to know they just spent $5000 on a set of lenses built with obsolete technology. Once Canon has the range covered with the new technology they will probably make quite a bit of noise about it and the resale value of the old lenses will drop a bit.

This is the best argument yet for Canon's increasing premiums on revised lenses. I am hoping the 24-70II is blazing fast and tack accurate; the MTFs certainly indicate the potential, and with the new bodies that potential can I hope be repeatable in practice.

 

[bdunbar79]

I agree to an extent. However, obsolete technology or not, how on earth can it be worth to spend $7200 and $11,400 on the new 300mm and 400mm lenses respectively, when you can get BOTH like-new used version I's for $9500 and the IQ and AF is no different at all? What are you really paying for? I've shot with both versions of the 300mm and the version 2 had absolutely no higher hit rate, sharpness, or IQ for sports. Definitely not worth it in my mind. It is certainly lighter but is it worth $2500 more for weight?

 

[Skulker]

The bits of this report that caught my eye are these bits.

In this example, the SD of the 5D Mk III samples was 17 lp/ih, while the SD of the 5D Mk II was 38.5. Those are similar to the numbers we’ve seen over and over—accurate focusing combinations have SDs in the teens while less accurate ones have SDs in the 30s.

But clearly the 5DIII and 1Dx do better than the other cameras, while the 1D Mk IV seems to be a bit between those two and the rest of the pack

the need for both a new camera and a new lens to get accurate autofocus began to make sense

The two newest Canon cameras have more accurate phase-detection sensors than their previous cameras. The newest lenses have more accurate focus movement (or provide more accurate focus movement feedback, or both) that takes advantage of those sensors.

From that I take it that to get the best results you need the latest kit. ;D So no great surprise there. It's nice to get confirmation that Canon are making progress! Of course some folks will be convinced they don't know what they are doing and couldn't make a decent (insert what ever stupid comparison you want) never mind build a decent camera and lens, and then go on to improve them a few years later.

The question to me is "Are you prepared to spend the extra money for the improvement you will get?"
To me it goes without saying that if you want to spend loads on a new camera but have one of the old lenses then the question should be read as including the cost of upgrading the lens as well at some point if you want the full benefit. One other thing that is self evident to me, IMHO, is that not everyone will come to the same conclusion. Some people will be very happy with the results that they get from their old camera / lens combination. Thats why I still use my old 30D some times, but I can honestly say I don't think I have ever put one of the latest lenses on it. Today it got used along side the new stuff. Guess what it still takes great pictures.

If I had an Italian super car (said that cos I can't spell Ferrari ;D) I can't see me putting the same tyres on it as I put on the Volvo (hides head in shame :-[), If I had that Ferrary (there I tried that time) with its supper dupper tyres most of the time it would go no faster than the old heap. The supper sticky tyres and engine would hardly ever work to the limit. Many people would say what a waste he has spent loads of dosh and he hardly ever gets anywhere any faster. They would be absolutly right, and also be totally missing the point of buying a Masaratie ( :-[ must improve my spelling)

It makes me think of something he said at the begining of his report "But I got tired of explaining that to people"



PS now that I think of it the camera and lens in the back of the Volvo today were worth about twice what the car is worth. Not many people would make that choice, but it's fine for me.

 

[LetTheRightLensIn]

Rogers findings surprise me a bit with respect to the 70-200 f2.8 IS II:

"The Canon 70-200 f2.8 IS II, oddly enough, seemed not quite as good as the above lenses, but more accurate than the older ones. Whether this was oddness in my measurements or a real finding, I wasn’t sure."

I would have thought this was one of the holy grail lenses in terms of Engineering prowess.

yeah but it also came out earlier than all of the others than had the full super precision

 

[LetTheRightLensIn]

http://www.lensrentals.com/blog/2012/08/autofocus-reality-part-3b-canon-cameras#more-8458

I think this is a very important article. It addresses my concerns about the Canon autofocus system better than any other article I've seen and will influence my future buying decisions.

What do you think of it?

Mike in Portland

I think of it what I wrote here:
http://www.canonrumors.com/forum/index.php?topic=8397.0

interesting stuff and kinda fits