fredmiranda.com
Login

Moderated by: Fred Miranda
  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

       2       3       4       5       end
  

Archive 2007 · example of 40D focus drift in AI Servo.

  
 
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #1 · example of 40D focus drift in AI Servo.


The next 7 posts will contain images that are screen dumps taken from a BreezBrowser display of a set of photos that show the focus drift on my new 40D. They do not show the entire images - just the portion with the AF sensors and below that. The 7 that I have shown represent all 16 that were shot as some are almost identical and not worth posting.

I was using a Tamron 180mm f/3.5 macro lens handheld.

ISO 800
shutter speed 1/1000
aperture f/3.5

The card that I photographed had little interest other than to show how wel the focus worked. It was in a pot plant almost 3m away from me.

I should add that BreezeBrowser was set to show the photos in fit-to-window size rather than 1:1, and that image enhancement was on (increased sharpness, etc.). Therefore you should not take these samples as an indication of 40D image quality - just the focus problem.

- Alan


Edited by Alan321 on Sep 19, 2007 at 01:42 AM GMT



Sep 18, 2007 at 12:18 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #2 · example of 40D focus drift in AI Servo.


#1,2 & 3 all were OK...









Sep 18, 2007 at 12:20 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #3 · example of 40D focus drift in AI Servo.


#4 is pretty much OK but shows the first signs of drift. The aiming is still OK.









Sep 18, 2007 at 12:22 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #4 · example of 40D focus drift in AI Servo.


#5, 6, 7 & 8 are all way off. Notice the sand on the near surface of the pot is in focus, and yet that is way outside the AF sensor area. The aiming is still pretty much where it was, however.









Sep 18, 2007 at 12:24 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #5 · example of 40D focus drift in AI Servo.


#9 is coming back towards where it should be focused (and used to be).









Sep 18, 2007 at 12:25 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #6 · example of 40D focus drift in AI Servo.


but #10 overshot. #11, 12 were the same.










Sep 18, 2007 at 12:26 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #7 · example of 40D focus drift in AI Servo.


#13, 14 & 15 are back on target again. Still the aiming is almost identical with no appreciable aiming errors.









Sep 18, 2007 at 12:27 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #8 · example of 40D focus drift in AI Servo.


#16 is going south again, but this was the end of the sequence as the buffer was now full.









Sep 18, 2007 at 12:29 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #9 · example of 40D focus drift in AI Servo.


To my way of thinking this is definitely focus drift that has nothing to do with the size of the AF sensor (I was using just the central one, which is shown in red while the others are shown in black).

If it is not just focus drift then there is a problem wherein the other 8 AF sensors are being used when they should not be, and without any indication that they are being used. However, in a separate sequence I got it to focus on something that was covered by none of the sensors, so to me the problem is just AI Servo focus drift.


My 40D is now off to Sydney to get the AF tested and hopefully repaired.
So far my 1Ds2 has not returned from having its AF fixed.

I seem to be pretty good at buying cameras with AF problems but hopefully you'll agree with me that this series of photos shows a genuine problem and not just a case of imaginitus


- Alan



Sep 18, 2007 at 12:33 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #10 · example of 40D focus drift in AI Servo.


This next image just shows what the overall photo looked like. You can better see the size of the AF sensor grid relative to the frame size.









Sep 18, 2007 at 12:48 PM
timbop
Offline
• • • • •
Upload & Sell: Off
p.1 #11 · example of 40D focus drift in AI Servo.


OK, in this series you convinved with the other inherent AF problem: AI servo imagines movement with stationary targets. Again, this is also generally accepted as a "known issue" with the non-1's. It's why a lot of folks shift AF to the * button for AI servo: for stationary targets you hit the * to get AF, and then release to prevent drift. I use CF 4.2 to use the * to stop AF in AI servo, because I am too lazy and malcoordinated to have to hit 2 buttons for ordinary shooting.

So, I see your point and add another - but neither is unique to the 40D. For still subjects you have to go to oneshot, or at least adjust your thinking shooting still targets with AI servo. The 40D may not be the camera for you.



Sep 18, 2007 at 12:53 PM
Planetwide
Offline
• • • •
Upload & Sell: On
p.1 #12 · example of 40D focus drift in AI Servo.


timbop wrote:
OK, in this series you convinved with the other inherent AF problem: AI servo imagines movement with stationary targets. Again, this is also generally accepted as a "known issue" with the non-1's. It's why a lot of folks shift AF to the * button for AI servo: for stationary targets you hit the * to get AF, and then release to prevent drift. I use CF 4.2 to use the * to stop AF in AI servo, because I am too lazy and malcoordinated to have to hit 2 buttons for ordinary shooting.

So, I see your point and add another
...Show more

This is true for my 5D as well.



Sep 18, 2007 at 12:55 PM
XsigmaSD
Offline
• •
Upload & Sell: Off
p.1 #13 · example of 40D focus drift in AI Servo.


I'm not sure exactly what your test is supposed to prove.

The manual clearly explains the different focus modes, and why you would use one over the other. The mode you are using is expecting movement. You are using the camera incorrectly and its not working, is that Canon's problem?



Sep 18, 2007 at 01:06 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #14 · example of 40D focus drift in AI Servo.


firstly, the same thing happens with moving subjects too (people at walking pace at about 60m range).

Secondly, what if we want to focus on something close-ish such as a wildflower without using a tripod ? Your own movement can upset the focus unless you use AI Servo mode.

Or, what if we want to shoot a bird that may take flight ? Are we supposed to change modes as it takes off and still get the action, or do we use AI Servo all along ? One Shot will trash the flight shots. AI Focus will take too long thinking about it and trash several before it switches over.

What I am doing is not invalid or unreasonable, and so yes it is Canon's fault. My problem, but their fault.

The fact that AI Servo applies predictive focus does not preclude it from working with a stationary subject, and does not account for it taking several shots to get back on track. This is not first generation AI Servo, you know, and it should not behave like it is.

- Alan



Sep 18, 2007 at 01:24 PM
foghorn
Offline
• • •
Upload & Sell: Off
p.1 #15 · example of 40D focus drift in AI Servo.


My 30D does this, if it's standing still, switch to one of the other focus modes. geesh.


Sep 18, 2007 at 01:28 PM
BubbaJon
Offline
• • • •
Upload & Sell: Off
p.1 #16 · example of 40D focus drift in AI Servo.


I can see both sides and I think what folks are saying is that it's not a battle worth fighting.
I would agree that an ideal servo system should not creep. I also agree that if there are viable work-arounds then use them. T'would that we lived in a perfect world...



Sep 18, 2007 at 01:38 PM
Alan321
Offline
• • • • • •
Upload & Sell: Off
p.1 #17 · example of 40D focus drift in AI Servo.


Jon, you're probably right but I'll add this before I give up on this thread...

If someone had posted a thread like this one before I bought my 40D then I would not have bought it. Instead all the talk of a much improved AF was the main selling point of the 40D and yet it has let me down. At least now we all know that it still behaves badly in some situations.

As I said earlier, the problem is not just for static subjects. If you're photographing people who are moving then you need to use AI Servo and keep it active, and yet I found that it still mis-focuses during a sequence of shots in which the AF sensor remains correctly aimed.

Hopefully I got a dud that can and will be fixed. I'll let you know one day. The camera is already on the way to the service centre.

My 1D2 worked far better in similar circumstances, so that at least is good. Unfortunately it too needed to be serviced before it worked as I would expect, but at least Canon did confirm and fix that AF problem. It certainly highlighted the deficiency of my 40D AF.

If the problems with my camera are normal then anyone in need of better AF for action / sports / wildlife needs to consider a second hand 1D2(N) rather than a new 40D.


- Alan



Sep 18, 2007 at 02:18 PM
Ed Peters
Offline
• • • • •
Upload & Sell: Off
p.1 #18 · example of 40D focus drift in AI Servo.


Thanks for posting Alan, much appreciated..


Sep 18, 2007 at 02:38 PM
CarpeyBiggs
Offline
• •
Upload & Sell: Off
p.1 #19 · example of 40D focus drift in AI Servo.


Alan321 wrote:
firstly, the same thing happens with moving subjects too (people at walking pace at about 60m range).


Hi Alan, interesting post here, but I think your test is misleading you, and your camera is probably fine. It's kind of like driving a ferrari on the go-kart track. AI-Servo isn't designed for still, non-moving objects.

Alan321 wrote:
Secondly, what if we want to focus on something close-ish such as a wildflower without using a tripod ? Your own movement can upset the focus unless you use AI Servo mode.

Hmm... Solution? Use a tripod. If not possible, stop down and take a few images with your natural "sway," and hope one of them nails it. Again, this is not movement that AI-Servo is designed to predict. You natural body sway is somewhat random and difficult to predict, especially for the 40D's algorithms. The AI-Servo is designed to track subjects that are moving in predictable patterns. Once the pattern is disrupted, it has to regain it's focus points and predictions again.

Alan321 wrote:
Or, what if we want to shoot a bird that may take flight ? Are we supposed to change modes as it takes off and still get the action, or do we use AI Servo all along ? One Shot will trash the flight shots. AI Focus will take too long thinking about it and trash several before it switches over.


You use AI Servo here, but when the bird isn't moving, you stop using the focus. I use CF 4-3, where focus is moved to the * button. When the bird is still, I focus once, then stop focusing, otherwise the camera thinks I am telling it that the bird is moving, when in fact it isn't. The result will be focus hunting. If the bird takes off, engage the focus again as it starts moving, so the focus can track it.

Alan321 wrote:
What I am doing is not invalid or unreasonable, and so yes it is Canon's fault. My problem, but their fault.


I guess if you feel it is Canon's fault, you might want to try a different brand. But my feeling is that it is a technique issue, probably not a camera issue. Who knows though, I could be wrong.

Alan321 wrote:
The fact that AI Servo applies predictive focus does not preclude it from working with a stationary subject, and does not account for it taking several shots to get back on track. This is not first generation AI Servo, you know, and it should not behave like it is.

See above explanation.



Sep 18, 2007 at 03:05 PM
Will Patterson
Offline
• • • •
Upload & Sell: Off
p.1 #20 · example of 40D focus drift in AI Servo.


My 30D did the same exact thing.


Sep 18, 2007 at 03:07 PM
       2       3       4       5       end




FM Forums | Canon Forum | Join Upload & Sell

       2       3       4       5       end
    
 

Welcome back
Log in to your account