I got back a 1D3 with the "fixed" assembly this week. Today I got a chance to go out in the backyard and take a few shots of our trotting along. Very simple, not too strenuous. Cloudy and dark but that's in the 1D3's favor, right? So here are some screen captures from Image Browser so you can see the focus point. This was done with single AF point with surrounding expansion points on. These are from one set of a seven shot burst. First two frames were similar to this:
So my conclusion is not good. I'm willing to say that I stink as a photographer, but this is pretty simple and it shows where the AF point was. Additionally I took a similar sequence about 5 days ago with a D3 and they were 100% in focus on the eye. I shot with a slower shutter speed so about 25% had a small amount of motion blur. But the focus was clearly on the right spot. Not sure about what to do with the 1D3, but this fix is looking like anything but.
As an interested observer who will never own a top of the range Pro camera I am beginning to suspect that they (The Canon engineers) have not identified the problem.
There are obviously 1D3 cameras out there that work. we see it continuously from people on the nature and wildlife site. (Jody, Chad, Bonnie, Jim Victory etc) and then there are cameras like Brians and Jeffs that dont. Even after being fixed.
maybe Canon need to identify one of those which work 100% and strip every component of the focus mechanism down and measure everything and compare to these to the individual parts in those that dont.
To me Brian and Jeff have behaved in a manner that could only be described as saintly. (That is comparing their behaviour to what I suspect mine would be if i were in their position)
If I were Canon, I would be giving them a full refund and a 1d2n to tide them over till a new "fixed" model comes out.
This whole situation might be a disaster for Canon but relative to the size of the Canon Corp, it is nothing to the heartburn and disapointment they are causing their loyal customers
Tim
Looking at the specs, I would love a 1DMkIII but I'm sticking with my 1DMkII. I was first on the waiting list at my dealer but cancelled the order and there's no way I'm jumping in until it's 100% fixed.
I've also heard that the "ring of fire" expansion points aren't necessarily the best way to go as the points can jump around more. Did you try it without the expansion points on?
Kim P wrote:
I've also heard that the "ring of fire" expansion points aren't necessarily the best way to go as the points can jump around more. Did you try it without the expansion points on?
I am not sure 'jump around more' is fully correct. On the 1D3 with centre AF point selected and surround expand enabled, the expand points only become active if focus lock is lost on the centre primary point. But even so, with or without expand, if focus lock is not achieved and you fire the shutter all you end up with is a set of nicely out of focus pictures.
I would start with One Shot AF and see how the camera operates. I believe you said in another post your MKIII had problems with One Shot. See if there is any impronement there.
With Servo AF, start with the center point and no expansion, and as Alistair pointed out make sure you have the focus locked before you hit the shutter. Check the CF's and make sure focus is given priority over frame rate, and set the micro-adjustment to -1 for all lenses.
I am in no way implying you do not know how the camera works, so please don't go there. My main interest is simply does the sub-mirror fix work. Right now there seems to be both positive and negative views.
Alistair and Richard,
Your pointers are right on the money. No questions about it, the AF system of the 1D MarkIII is different than that of its predecessor. You almost have to start from scratch to familiarize yourself with it.
Alistair Watson wrote:
I am not sure 'jump around more' is fully correct. On the 1D3 with centre AF point selected and surround expand enabled, the expand points only become active if focus lock is lost on the centre primary point. But even so, with or without expand, if focus lock is not achieved and you fire the shutter all you end up with is a set of nicely out of focus pictures.
It did obtain focus on the first two shots (I included the capture from DPP since Image Browser makes that first shot look blurry for some reason). The next 5 were out. This was with a single AF point with the expansion points on.
I would start with One Shot AF and see how the camera operates. I believe you said in another post your MKIII had problems with One Shot. See if there is any impronement there.
With Servo AF, start with the center point and no expansion, and as Alistair pointed out make sure you have the focus locked before you hit the shutter. Check the CF's and make sure focus is given priority over frame rate, and set the micro-adjustment to -1 for all lenses.
I am in no way implying you do not know how the camera works, so please don't go there. My main interest is simply does the sub-mirror fix work. Right now there seems to be both positive and negative views.
I understand, but unfortunately it did have lock before the first shot and the two shots of the sequence are definitely in focus.
I didn't include another 8 burst sequence that I took. In that one the first 2 are infocus, the next one soft, the next infocus, the next two way OOF, the next one infocus, the last one soft.
The one shots pictures (limited so far) seem mostly focused. The issue I have with one shot is that often it would jump to a higher contrast area even though it indicated my selected AF is infocus. Firmware 1.1.3 seems to help with that although I haven't done a huge amount of testing so far. I guess I'll try to call Canon to see what they say but they haven't been very helpful on the phone so far.
bcaslis wrote:
It did obtain focus on the first two shots (I included the capture from DPP since Image Browser makes that first shot look blurry for some reason). The next 5 were out. This was with a single AF point with the expansion points on.
How long were you tracking the dog before you fired?
How do you know the camera obtained focus?
In AI-Servo mode the focus confirmation light does not light up. I haven't used Zoom Browser for years but if you have a centre AF point selected and take a series of images, whether the camera has focussed or not, the red AF mark is displayed in ZoomBrowser yes?
I don't know how long you have to track before shooting but it should never be long. e.g. it should be far less time than IS takes to settle in. The reason is simply that the AF system is fast enough to reassess focus between shots at 10 fps and so it should not take a second to lock on in the first place. It would be nice of Canon to specify how much time is needed but I have never seen it.
Likewise, if the chosen AF sensor is not locked on when shooting begins then the camera should quickly lock onto something with one of the expanded AF sensors, even if it is not the intended target. The only excuse for a 20 shot burst of out of focus photos is that none of the sensors could find anything to lock onto within the first couple of frames. This is meant to be the speediest AF system ever.
AGeoJO wrote:
Alistair and Richard,
Your pointers are right on the money. No questions about it, the AF system of the 1D MarkIII is different than that of its predecessor. You almost have to start from scratch to familiarize yourself with it.
I sort of disagree. There's no question that there are cameras out there with AF problems but... the AF operation is not all that different from previous cameras - or at least it doesn't have to be.
I shoot primarily sports; field sports outdoors and the usual basketball and swimming indoors. I previously shot with two MkII and a MkIIn and in Sept. '07 I started shooting with two MkIII (both within the defective serial number range). I have found both of them to AF perfectly. However, when I unboxed both my MkIII's the first thing I did was to disable every option related to auto-select, auto-expand, etc. This is much the same way as I had my MkII and MkIIn optioned. Using the center AF point (and other manually selected AF points) I find the AF to perform quite satisfactory. By comparison, I find the auto-AF/auto-expand point option to be just as unreliable as every other 1-series camera I've ever owned (1V, 1D, 1DMkII and 1DMkIII).
Just hung up with a service rep in Virginia - said he hadn't heard anything but good news back from owners of fixed Mark IIIs. Said the forums are all talking about how great it is. I asked him if he read Fred Miranda - he hadn't!
Alistair Watson wrote:
How long were you tracking the dog before you fired?
How do you know the camera obtained focus?
In AI-Servo mode the focus confirmation light does not light up. I haven't used Zoom Browser for years but if you have a centre AF point selected and take a series of images, whether the camera has focussed or not, the red AF mark is displayed in ZoomBrowser yes?
I was tracking about 4 or 5 seconds before the very first shot.
I'm assuming focus was obtained because the first and second shot were correctly in focus. Is there some other way to determine this?
bcaslis wrote:
I was tracking about 4 or 5 seconds before the very first shot.
I'm assuming focus was obtained because the first and second shot were correctly in focus. Is there some other way to determine this?
Only Canon has software which will tell you there actually was a focus 'lock' at the time the image was shot.
If the 1st and 2nd image are in focus then it is fair to assume lock was gained and then lost, my only guess is that since you were using AF expand and none of the points were able to get a lock despite the central AF point being on target then something is still wrong with the camera. Have you done anymore tests on different targets to discount (and I hate to say and don't take offence to this) user error?
I try to give the AF a full second to lock before I hit the shutter....longer if I can. You are correct about the AF following at 10 FPS, but the assumption is made that a lock has been established and now the lens only has to travel a small distance to keep up. I have never seen a number from Canon.
John--G,
I agree with you, that is why I suggested turning off the expansion. My camera works pretty well with the defaults, just likw yours. The -1 micro adjust comes from the RG update.
jbear2000,
There have been several people here at FM that reported better results after the fix, and some who are still having trouble. My MkIII is working better than most, but I haven't shot it in much 95 degree weather yet. I am trying to figure out if I should even send it in.
Brian,
I beleive some of the MKIII's have problems other than the sub-mirror issue and I not sure the VA site is testing the camera completely. Yours may be one of the really bad players....I know Jeff's falls in that category. I have my fingers crossed Canon is not through with the MKIII.
Alistair Watson wrote:
Only Canon has software which will tell you there actually was a focus 'lock' at the time the image was shot.
If the 1st and 2nd image are in focus then it is fair to assume lock was gained and then lost, my only guess is that since you were using AF expand and none of the points were able to get a lock despite the central AF point being on target then something is still wrong with the camera. Have you done anymore tests on different targets to discount (and I hate to say and don't take offence to this) user error? ...Show more →
No offense, user error is always possible. I did three bursts each about a second or so. One was like shown here, two were very mixed as I mentioned in a previous post (focus, soft, OOF, focus, soft, etc...). I agree that I should try another target and different lighting, but it's raining here for most of the past two weeks so it's been very difficult to get time for any tests. But my reason for posting was that I never got results as bad as this before sending it in for repair. Previously it would normally be about 50 - 60% in focus. Now it's more like 30% in focus.
If this is the only shooting you have done since the repair lets keep our fingers crossed that Canon haven't (searching for apt Americanism) screwed your pooch?
I must say, the results with my new blue dot 1D3 so far are quite positive. For the first time ever I was able to get a burst of 12 to 15 sharp shots of a bird in flight, I staggered back in surprise and if the rig hadn't been mounted to a Wimberley head it would have fallen to the ground!
