Forthright Review on 1DIII : 1DIIN remains king of sports focus!

[Asher Kelman]

Forthright Report by Rob Galbraith on the 1DIII: When Rob first reviewed the apparently groundbreaking built from scratch Eos 1D Mark III auto focus tracking sports and action camera, the report practically was made in 14 carat gold foil except for the focus, not doing too well and attributed to the issues that occur in pre-production models.

The low light performance with lenses such as a 50 1.4 was found to be stellar. All that is still true. The pictures at ISO 6400 are still acceptablle and ISO 1600 is very good. That's still so with production models.

What is staggering is that several other 1DIII copies tested by Rob and Mike Sturk, both well known sports/news photographers all had the same issues with maintaining or getting focus with long lenses and movement or even no movement some times! UGH! You must use the link below to get the whole story.

So the best current relaible camera on the market today for getting the shot in sports or presumablely for bird and wild life too, is the 1DIIN or the 1DII.

This is so sad, a disappointment and a letdown.

What does it mean?

Since presumably the much tested 1DsIII (or whtever they will call it) likley has the same autofocus circuitry and firmware, it's likely in trouble. I expected the camera to appear no earlier than the end of the year. However, the 1DIII will be the test bed to get the AF right and so the 1DsIII might, IMHO, be delayed if the issue is not sorted out.

I have trouble getting the 5D to get focus in dim light. The 1DII is far better but has failed me in weddings in dim light unless I use flash.

I was so looking forward to the 1DIII to be my new miracle cam to deeal with both issues. Well, not just yet! No doubt this is something Canon will now address. If they aknowledge it to Canon users, as Lecia did in there M8 fiasco in Decmeber 2006, is another matter!

Read more here (http://www.robgalbraith.com/bins/content_page.asp?cid=7-8740-9006).

[Asher Kelman]

Prophotohome, Drew's versioning of the RobGalbraithfroum has a Full Review of the 1DIII and shows a bunch of tiny vsequential shots to prove that the focus was fine. here (http://www.prophotohome.com/forum/pro-photo-reviews-articles/73606-canon-1d-mark-iii-full-review.html?garpg=8#content_start)

I'm so unimpressed! The pictures show no detail and there is no way to decide if they are in focus or not.

Asher

[KrisCarnmarker]

This is certainly a worrying issue. Mine should be arriving today or tomorrow, and I am now starting to think I should wait.

I have tried to read everything I can about this issue since last night, and my conclusion is that there is no consensus. Rob is a very trusted reviewer, but then, several trusted photographers have reported no similar issues.

What is most worrying is Canon's complete silence on the matter.

[Steve Saunders]

Still waiting on mine. I'd heard a few weeks back from the Canon rep that the delay was due to the AF being tweaked or a new AF module fitted, I can't remember now.

[Cem_Usakligil]

... What is most worrying is Canon's complete silence on the matter.
Hi Kris,

This is a difficult decision for any company to make, to react or not to react. And if you'd react, then how would one do it properly? Deny that there are issues at hand? People will become suspicious anyway. Accept that there are indeed issues? This would be a marketing suicide.


Right now, I can imagine that being silent for a while is the better route for Canon. Gradually, more and more users will work with the camera and statistically the severity of this issue will diminish, unless of course there really is an irrefutable issue. Then this might mean a huge flop for the camera and Canon in the making. Let's just wait and see..

Just my $ 0.02

Cheers,

[Asher Kelman]

I'd expect inscrutable silence!

There is a big thing about saving face.

I'd go with Rob. He has no reason not to be utterly truthful. I really commend his honesty. This is rare! Lesser people pander to the overwhelming advantage of being on good terms with powerful people. We can even rationalize what we are doing so we delude ourselves to thinking we are pretty fair.

However, a reviewer as serious as Rob has to be heard and resepcted. I cannot imagine he does not know how to use the camera. That's the only possibilitiy left!

Otherwise I'd say the others are not as careful. Remeber, Rob is concerned with tracking. That is a must for sports but also for wildlife, bird photography and news.

If I had to shoot a newly wed bride and groom coming toward me in Hawai by the ocean using a 300mm lens, I'd want to use my 1DII.

At least for now.

I see no reason to go nuts. Let things get sorted out. Then you are not asking for the camera to be retrofitted as they did with the very early 1D Mark I models.Who knows, maybe the distributer is sending Rob only dropped cameras of different production runs (they cannot dislike him that much) and save the perfect one for Drew?

In any case, I would let things sort out. Perhaps a simple firmware update might appear. :)

Asher

[Steve Saunders]

Okay, I've just read RG's report. As Rob is regarded as more of a Canon camera user than a Nikon one, I'd say that it must have been hard for him to be critical of the brand he normally uses. That makes his report more credible, as it would be very easy to be critical of a brand of camera that he doesn't normally use. We've all read reports over the years by people who just gloss over potential issues, but RG generally says it as he sees it.
Having said that, I really need to try the camera myself. The sample shots on the RG report don't particularly alarm me as I often get a similar number of slightly OOF shots with my D2X and D200 and in very similar shooting conditions, so this wouldn't bother me too much. The number of days we see sun and hot at the same in my part of the world you could usually count on your two hands (although May this year was lovely for a change) and it's been drizzling rain and dull here for the past couple of weeks, so I'm not going to worry about a camera overheating. Now you know why I don't like white lenses...

[Asher Kelman]

Steve, white lenses are cool! Were are you that has such mild weather. Here it's up in the 80-90 degree F range in May!

The Canon 1DII does a perfect job of such shots. Why upgrade right now when a month or so we'll know more. companies have a bad habit of not paying their beta testers! If they gave these cameras in alpha to fashion and beauty photographers, it's not surprising to find there are issues uncovered only now!!

Asher

[Steve Saunders]

I live in Dublin, Ireland Asher and it's another wet and cold summers day today. I'm swapping from Nikon to Canon after about 20 years or more using Nikon. The thing is that I've already paid for the MkIII in full and several lenses (300L IS, 70-200L IS and 24-70L) and I don't have another Canon body, which means my lenses are at home awaiting a body. The plan was to buy a second MkIII when they become more freely available. I would have liked to have the MkIII for the Dublin football games (the finals) this coming Sunday, and for the one last Sunday and the one two weeks before that. But it just isn't happening. An hour using this camera at a football game will reveal it's strengths and weaknesses to me I think.

[Ben Rubinstein]

The best thing Canon could do is fix it and quickly then offer a firmware upgrade or take back the cameras to do it themselves. Canon's pro rep was based on their sports cameras, albeit Nikon has nothing much out there in competition and hasn't had for a while, but ignoring a problem like this on their premier 1 series, with it's most important mechanical/technological part, the most crucial part for a sports camera, that would be suicide, especially when it's reported by someone as respected as RG.

[Steve Saunders]

Good point Ben. I wonder if the AF module is the same as in the MKIIn. If so, then surely just a firmware tweak is all that is needed?

[Scott B. Hughes]

If I had to shoot a newly wed bride and groom coming toward me in Hawai by the ocean using a 300mm lens, I'd want to use my 1DII.

At least for now.

That's us... we have been holding off on placing an order and now with this recent news, we are likely to continue to wait (longer than planned).

-Scott

[Asher Kelman]

Scott,

I've observed that weddings on an Hawai beach front hotel lawn are not quite predictable. There may be a path between two blocks of guests, but the wedding group and the official "Marrying officer" are against bright waves with glistening water, the sun is higher than one would like or lower than one can tolerate and people seem to be less disciplined. Time for things may also be very relative. Without warning the bride and her father can appear, or not!

So I'd say that Hawai wedding are a good test for the 1DIII. If it works in bright sun on a hot day, I'm impressed.

Asher

[Scott B. Hughes]

We rarely have focus issues with either the 1Ds or ID (M2s), but I'm looking forward to better quailty at higher ISO. A candlelight ceremony can be rough.... nothing slower than f2, and even then shutters at 1/15 - 1/80. ISO1600

-Scott

[Steve Saunders]

Well my MkIII arrived today. I collected it and the new 300 IS USM f2.8 and brought them home, then had to leave for work. I'm just about to open the boxes now...

[Bart_van_der_Wolf]

Well my MkIII arrived today. I collected it and the new 300 IS USM f2.8 and brought them home, then had to leave for work. I'm just about to open the boxes now...

Congrat's, hope you have a very nice weekend, and that everything (that needs to be) is in focus.

Bart

[KrisCarnmarker]

Mine arrived yesterday as well. Didn't do any AI Servo shooting, but did test quite a bit. It was bright sunshine (as always here) and 43 degrees out, so plenty of heat. Shooting with the 70-200 f2.8, I had no problems focusing. I will try with some back-lit subjects as well later.

[Asher Kelman]

This is a good start Kris and Steve,

I do hope we can find suitable subjects like birds in flight on a hot beach or approaching cars, joggers and so forth.

Asher

[KrisCarnmarker]

This is not intended as a repudiation of RG's article. It is just for my own curiosity.

I shot advancing/receding cars from around 200m, on a hot sunny day. I followed RG's instructions for the conversion; i.e. use DPP, sharpening at 3, no other params touched. I shot several cars in both high speed burst and low speed burst. All shots where with the 70-200 f2.8 @ f2.8 and 1/3200. Here is one from the high speed trials:

http://kris.carnmarker.com/forumimages/opf/m3ft/s1.jpg


Here's one of the low speed trials:

http://kris.carnmarker.com/forumimages/opf/m3ft/s2.jpg


In all instances I kept the center focus point on the license plate (as much as possible). Obviously the inserts are 100% crops.

[Steve Saunders]

Looks like a decent keeper rate there when you consider that the cars must have been moving faster than a footballer would.
I used my MkIII Sunday morning at a football game for a few minutes. It was raining and I had the 300 f2.8 attached. I only took a few shots as it was my first outing with Canon and I used my Nikon D2X and 300VR for the rest of the time. This weekend it will be in at the deep end so I'll post some sports shots then. My initial impressions are quietly confident that I made the right move.

[Asher Kelman]

Here's one of the low speed trials:

http://kris.carnmarker.com/forumimages/opf/m3ft/s2.jpg

In all instances I kept the center focus point on the license plate (as much as possible). Obviously the inserts are 100% crops.

Kris,

Thanks so much for the generous effort and for posting here! This is the sort of study that will eventually help us. One always needs a control camera as a baseline which would be the 1DII or 1DIIN. without a baseline, we are assuming but not knowing. Ideally we'd get of each, but the minumum is one unless the result is startling with the 1DIII.

Here the focus appears variable. Numbering the pics from left to right the degree of focus can be roughly guestimated from 1-5, 1 being focus is just fair and 5 is excellent.

This is of course merely my own quick estimate and subject to great error as no measurement are made. One could plot the density over the lines and look for the highest peak in density or the second derivative.

Anyway, just for my amusement, here's my results:

Frame Grade (5 is best, 1 is fair)

1....... 3
2....... 2
3....... 2
4....... 3
5....... 3.5
6....... 2
7 ....... 3
8 ....... 3
9 ....... 4
10....... 5
11....... 5
12....... 4.5
13....... 3.5
14....... 3

Now please don't get mad at me about my grading as I didn't measure, just looked and the side by side positioning might well alter perception. In any case, for sure, density of the letters varies!

To me, ony 4 of the 14 are im excellent focus!!!! Not to good.

So I am not overly impressed with the camera but very gratified with the test.

Asher

[Steve Saunders]

The thing about Kris's test is that it must have been difficult to keep the number plate in the focus point with the car beng so far away and the plate so small in the frame because of that. It would have been very easy to wander to the area just behind the plate, especially if a monopod or tripod wasn't used. I'm not saying that this was the case, but I'd probably have the car closer for that test, or use a longer lens or both. I'll be covering sports this weekend and I'll get someone to run towards me similar to the test RG did, that way the shooting conditions should be similar. Except of course that we haven't any sun here at present...

[Asher Kelman]

Good points Steve!

Any chance someone has a 1DII for you to take along too? That would be the greatest help since we'd have a side by side comparison!

Good shooting. We love the sports pics anyway!

Asher

[Steve Saunders]

Some of my fellow shooters that I meet regularly at games use MkII's, so if I bump into them this weekend I'll ask them to sit beside me and do the same test on the same subject with one of my spare cards in their camera. That might be the best way to do this.

[Ray West]

Hi Kris,

What size is the af point, relative to the images of the number plates? I suspect, even if you can get the sensitive area over the plate for the more distant views, it will pick up an area with higher contrast e.g. windscreen/body.

If you get another chance, with your friend, after you've done your limit pushing tests, repeat with settings you'd normally use for each camera - i.e. try to get the best from each camera, instead of the worst.

Best wishes,

Ray

[Cem_Usakligil]

.... i.e. try to get the best from each camera, instead of the worst.
Hi Ray,

I kind of like what you've said, may I quote it elsewhere if the situation calls for it?

This is not to say that Steve hasn't been trying to get the best from his camera, I am sure he has been doing just that.

Cheers,

[Ray West]

Hi Cem,

quote whatever, I can always deny it ....;-)

I am not being critical of Kris's/Steve's efforts, or RG's or whoever - I just wanted more input. I am very good at breaking things, finding faults. That's why I'm an engineer. But, sometimes, it is easy to get misled into 'testing the wrong things'. In much of the testing that goes on, the effects of the uncontrollable outweigh the controlled. 'Cars in the road' is not a controlled environment, but it is an interesting real world situation. (for example, in the last number plate, why did it not focus on the railing, or somewhere between?) Testing like Kris and others are undertaking bring real world understanding to more precise scientific methods, which are generally inappropriate outside of the test laboratories' controlled environment.

Best wishes,

Ray

[Cem_Usakligil]

....I am not being critical of Kris's/Steve's efforts, or RG's or whoever - I just wanted more input. I am very good at breaking things, finding faults. That's why I'm an engineer. But, sometimes, it is easy to get misled into 'testing the wrong things'. In much of the testing that goes on, the effects of the uncontrollable outweigh the controlled. 'Cars in the road' is not a controlled environment, but it is an interesting real world situation. (for example, in the last number plate, why did it not focus on the railing, or somewhere between?) Testing like Kris and others are undertaking bring real world understanding to more precise scientific methods, which are generally inappropriate outside of the test laboratories' controlled environment...
Well said. As a fellow electrical/electronics engineer, I can only agree fully! :-).

Cheers,

[Asher Kelman]

In much of the testing that goes on, the effects of the uncontrollable outweigh the controlled. 'Cars in the road' is not a controlled environment, but it is an interesting real world situation. (for example, in the last number plate, why did it not focus on the railing, or somewhere between?) Testing like Kris and others are undertaking bring real world understanding to more precise scientific methods, which are generally inappropriate outside of the test laboratories' controlled environment.


One can use just the central focus point.

That simplifies things.

Asher

[KrisCarnmarker]

Looks like I need to clear up some things :)

My test was not intended to be a scientific test, it was just a quick test for my own curiosity. Before my life as a software engineer, I worked in scientific research (acoustics), so if I wanted to set up some elaborate test I could probably manage. I'm just not interested :) That does not mean that my test was completely thoughtless either. I was aware of the limitations, but as has been said, in real life there are always limitations and compromises. The intention was not to get the worst of the camera, but then neither was it meant to get the best. It was meant as real life test, albeit with some caveats maybe (see rest of post). Some explicit decisions I did make:

-As high speed as possible to limit camera shake (1/3200)
-As large an aperture as possible to limit DOF (f2.8)
-Use central focusing point only
-Use fast moving cars (relatively anyway; this was not a highway)
-Stay far away. I did not want them to see me. This was not ideal, but in this culture you have to be a bit careful.

What size is the af point, relative to the images of the number plates? I suspect, even if you can get the sensitive area over the plate for the more distant views, it will pick up an area with higher contrast e.g. windscreen/body.


I did not make a mental note of exactly how large the AF point/rectangle was relative to the car, but I think you are right. Certainly in the most distant shots, the AF rectangle covered more than the plate and grill. As the car approached, I could keep the complete focus rectangle on the plate/grill.


I am not being critical of Kris's/Steve's efforts, or RG's or whoever - I just wanted more input. I am very good at breaking things, finding faults. That's why I'm an engineer. But, sometimes, it is easy to get misled into 'testing the wrong things'. In much of the testing that goes on, the effects of the uncontrollable outweigh the controlled. 'Cars in the road' is not a controlled environment, but it is an interesting real world situation. (for example, in the last number plate, why did it not focus on the railing, or somewhere between?) Testing like Kris and others are undertaking bring real world understanding to more precise scientific methods, which are generally inappropriate outside of the test laboratories' controlled environment.
Ray

I agree and I did not take your comments as criticism of my efforts. I hope I am clearing things up a bit. As for the last number plate, I explicitly avoided getting the focus rectangle across the railing, which means more of the hood instead.

Ideally, I would have set up an experiment with an object on a rail, with controlled acceleration and speed, patterned walls close by to establish exact focus plane and DOF, tripod mounted, etc., etc. :) But that's not what it was all about.

Lets see what Steve comes up with, it would be great if he could provide comparison with the MkII(N). Steve, maybe you could even temporarily switch cameras with your fellow shooter, so that you shoot the MkII and he/she shoots with your MkIII. That way both shooters shoot with both cameras.

[Steve Saunders]

...Steve, maybe you could even temporarily switch cameras with your fellow shooter, so that you shoot the MkII and he/she shoots with your MkIII. That way both shooters shoot with both cameras.


Yes that should be possible, once I know I will be meeting up with them. I get my weekend assignments on Friday, so I'll make a few calls then.

[Steve Saunders]

Well I'll be shooting two games today, between 6pm and 7pm. The rain is holding off so far, but is forcast for later...

[KrisCarnmarker]

Great, looking forward to your impressions!

[Steve Saunders]

Sorry but I couldn't remember the thread where I said I'd do a test, so I've started a new one. I didn't get to meet up with any Canon shooters using the MkII last weekend, only 30D and similar and lots of Nikons. I mean when I was shooting Nikon, I was always outnumbered by the Canon guys. Where were they all when I needed them last Saturday?

Anyway, this sequence is the closest I got to a test. I have to hand it to Rob Galbraith for putting the time into his tests, this just isn't easy unless you dedicate lots of time specifically to it. These are 600x350 pixel crops, the original image that you can see in the small film strip to the left of the crops. I saved the Jpeg at quality 9 so the crops look almost like they did before I saved the whole test file, sorry it's 432kbs as that's as low as I could go without it looking terrible.

The centre point of each crop is about where I had the centre focus point (as far as I remember anyway) and from that you can probably gather that the players were all over the place. Lens was the 300IS USM L shot wide open at f2.8 and the relevant camera settings were;

AV mode.
ISO 320.
Jpeg fine with no in-camera or PS sharpening.
AI servo, centre focus point, 5fps.
Shutter speed was;
1/2500sec on shots 1,2,3 and 9.
1/2000sec on shots 4,5,6 and 8.
1/1600sec on shot 7.
The whole sequence was shot over 2 seconds at I think 4fps.

Not very scientific I'm afraid, but I have to say I'm quite happy with it all and have had no problems.

http://www.goldwingfacts.com/images/nonWingimages/Mk3test.jpg

[Asher Kelman]

Great effort, Steve!

Glad you set this test up. Now there's an idea that you might test, and that is to add more focus points so the camera has no need to re-choose if the cetner focus point target is lost.

The depth of focus of course is limited. What distance were you from the players?

What exaclly did you focus on? The face is more detail rich, as is the ball. The arm and yellow band in the first image is not really in sharo focus, but the sleeve on the left seems to be.

So what is being tracked We need to know that to judge the success or not of the focus tracking here.

This is the kind of sequence that we need to see to disclose what is possible.

Thanks,

Asher

[Steve Saunders]

Hi Asher. The centre of each crop is roughly where I focussed. I say roughly because these guys were struggling and zigging and zagging all over the place, typical of Gaelic Football. This of course makes it very difficult to hold a particular focus target.

The second crop is a bit wrong, I should have cropped further to the left, so the focus point would have been more left as well. Only excuse I have is that it was late last night when I did the editing in PS, and sure I'm just as human as the next guy or girl after all. But I'd say when I do another test tomorrow, I'll have learned from all this and will know how to do it properly. If time allows, I'll get one of my biker friends to rid towards me with a fixed point of focus for me to target. Then I'll do the crops again.
First shot was about 150 feet away and the last about 70 feet, that's guessing of course. If temperature is a factor in how the AF behaves, it was dull and cloudy but warm at 14 degrees C.

I should mention that I shoot these same teams at this same venue and in almost identical conditions quite often, always previously with Nikon D2X and D200 and their super 300VR f2.8 lens. The bottom line is that I just can't see any real downside with the Canon, if anything I have a higher keeper rate as the MKIII seems to hunt less and finds initial focus faster than the Nikons.

[KrisCarnmarker]

Thanks for the update Steve. Looks pretty good to me.

[Steve Saunders]

"Last edited by Ray West : Today at 11:08 AM. Reason: I merged these threads, hope that's OK"


That's cool with me Ray. I think this is where I meant to post it in the first place!

[Asher Kelman]

Steve,

You're work is much appreciated! now have you seen any mahor improvement grabbing and maintaining focus in low light? That's a major peeve with the 5D and 1DII, for me at least.

Asher

[Steve Saunders]

I haven't taken any pics indoors yet, but I've toyed with the AF indoors and it seems to grab focus very fast and without hunting.

BTW, I just installed the Canon software and I see that Zoombrowser EX can be set to display the focus points on the image, so you know exactly where you focussed. Very useful for the testing.