Based on this thread I'm guessing the MkIV discussions (once they actually roll out the camera) are going to be pretty similar to the MkIII discussions in their employment of diplomatic skills. I particularly love the blasting related to the "fact" that AI Servo can't be used on objects moving in plane parallel to the photog because they're "not moving towards or away" from the photographer. Good grief. I'm just a lowly social science type (in other words I never studied anything useful in school ) but it seems to me the fact that one body is moving in relation to another by definition means that it's moving either closer or further away from you, unless it's moving in a perfect circle around you which is a pretty rare occurrence.
Sahid, if you're going to do this test and some type of useful data I'd suggest you use just the center AF point in AI Servo. (Might as well also find a target that is actually moving pretty well to test the camera's capabilities in tracking since that's what everyone is interested in.) As others have said using lots of points in a busy background is asking for all kinds of trouble. And is generally not the way BIF shooters, anyway, use it. At least it wasn't prior to the MkIII. I have the impression that some photogs with well-working MkIIIs are incresasingly using that mode.
It's too bad you are pulling the "I got more experience than you" card.
Sure, I haven't shot tens of thousands of birds and I have absolutely no intentions of doing it in the future, not to discount the genre. Not at all. It just doesn't interest me.
Now I have shot birds, many of them in fact, that always seem to make it into the trash bin because I usually shoot them when I am testing and for me they don't hold too much artistic value other than to evaluate the camera. Again, that's just me. Not knocking it one bit and I can attest to how difficult it can be depending on the bird and its flight patterns. Believe it or not,we do have ponds, ducks, gulls, etc. In NYC.
I have found that a bird coming at you with consistent speed is fairly easy to shoot as opposed to say a bird flying erratically, just like shooting a car moving right at you. Consistent speed is the key. Bird, athlete, car, plane, whatever it is...
And honestly, though I realize there are lots of birders out there, that's hardly the only genre that require precision AF.
But hey, you are the 10k duck shot man. Who am I to argue?
PetKal wrote:
We need to look at this empirically a bit. And I'll try to be gentle.
Nick, I estimate I have shot flights like that, some easier, some more difficult, at least ten thousand times more than you have, in all kinds of camera-lens combinations. I must have also posted several hundreds of those shots on this board in the past year alone. Now, based on the shots I've seen you post, my judgement is that your photography experience is limited even with simple stationary birds like a seagull on a rail.
Now, you are telling us that the target type I consider difficult is in fact "the easiest thing to track" ? Hmm.....by golly, what have I been doing wrong all these years ...Show more →
Nick Nishizaka wrote:
In the OP's photos, the subject was clearly moving (look at the changing background) and I think it's pretty safe to assume the focusing distance was changing. However, there's no reason why AI servo should not work for static subjects...
The OP's problem was that he used automatic AF point selection, leaving the camera to make the decision as to where to focus in the frame.
But that is far too sweeping a generalization. Using this logic, automatic AF point selection should never be used, because it will just choose something at random to focus on.
The truth is, the automatic AF selection algorithm is strongly biased toward the closest subject. In the first sequence above, this is clearly the man walking; nothing else in the frame is even close to the same plane of focus.
And the camera appears not to be locking and holding focus on what is clearly the closest subject in the frame. So unless there are some mitigating circumstances we are not aware of with the pre-production camera, I think this qualifies as unexpected behavior. At the very least, it suggests that automatic AF point selection may not always be relied upon to do something predictable, which is certainly worth discussion.
garyvot wrote:
But that is far too sweeping a generalization. Using this logic, automatic AF point selection should never be used, because it will just choose something at random to focus on.
The truth is, the automatic AF selection algorithm is strongly biased toward the closest subject. In the first sequence above, this is clearly the man walking; nothing else in the frame is even close to the same plane of focus.
It may be strongly biased toward the closest subject, but if something nearby has greater contrast, that is what it will focus on. This is why I always use centre-point focus only for moving objects with busy backgrounds.
garyvot wrote:
But that is far too sweeping a generalization. Using this logic, automatic AF point selection should never be used, because it will just choose something at random to focus on.
The truth is, the automatic AF selection algorithm is strongly biased toward the closest subject. In the first sequence above, this is clearly the man walking; nothing else in the frame is even close to the same plane of focus.
And the camera appears not to be locking and holding focus on what is clearly the closest subject in the frame. So unless there are some mitigating circumstances we are not aware of with the pre-production camera, I think this qualifies as unexpected behavior. At the very least, it suggests that automatic AF point selection may not always be relied upon to do something predictable, which is certainly worth discussion....Show more →
What I meant to say was that using automatic AF selection isn't the best way to do controlled testing on a camera's AF system. I think short of seeing the actual AF point that the cemera decided to use and how it was shot technique-wise, we should take the OP's test with a grain of salt.
It's exactly this kind of agressive BS that dissuades people from posting the results of their tests on these boards. I know I'm certainly not. Let's just wait until RG gets a camera & we can all rain on him.
Nick Nishizaka wrote:
A duck flying straight at you? Honestly, is that more challenging that a runner coming right at you?
yep it is.
i routinely shoot people running at me in football, soccer, baseball, basketball, etc. but a duck or bird in flight is much different and more difficult to shoot.
i routinely shoot people running at me in football, soccer, baseball, basketball, etc. but a duck or bird in flight is much different and more difficult to shoot.
We are talking about bird in flight...as in flying directly at you, or gliding towards you in a near consistant speed right? That is much, much easier than a randomly flying bird...again, I can certainly appreciate the difficulty of the genre. Quite honestly catching BIF can be a mind numbingly difficult endevour with the exception of when they glide or move in one direction at a constant speed, which is the specific scenario I am talking about based on the duck sample posted. Perhaps I was wrong to compare to a athlete that fills the frame more, but relatively speaking, it's not that difficult to do. Sorry...don't need 10k duck shots to know that.
p.5 #10 · Pre-production 1Dmk4 tests with Ai-servo
According to the 30d manual AI servo is used for moving subjects when the focusing distance keeps changing...so people are wrong here since AI servo works the same in all canon DSLRs even had to look in my eos3 manual yep pretty much the same...still AI servo isnt what is recommended with the above shots no matter your opinions read the manuals kids...thanks
Precisely half you pixel peepers wouldn't have a **** clue how to use a camera in the field, your too busy discussing bit depth noise moire and anything else totally useless to the day to day use of a camera, precisely why websites go down the shitter, and you get posts on here day in day out is my lens soft is it front focusing where do I put a lens cap after I have taken it off?
If an object is moving accross the focal plane it is going to be very difficult for a camera set to AI to PREDICT movement as the subject is not moving at a constant pace, Thats is what Ai is designed to do, predict where the focus needs to be based on a constant rate of movement. A camera set to AI tells the camera to lock onto a subject predict its rate of movement then predict wherethe focus will be when the shutter button is pressed and teh image is no longer seen as the mirror has flipped up. That is what AI is for. Not some person walking sideways and rattling off 10 shots, how can the camera grab focus when it can detect no movement?
Now if you don't beleive that pick up that black thing called a camera and go and test this yourself half of you might actually learn something.
You tell us Einstein.
Post like this are half the reason I and others are done with this site. This will be my last post on this topic. Having shot motorsport and all types of sport for 10 years I think I have a better understanding of AF than most here.
p.5 #11 · Pre-production 1Dmk4 tests with Ai-servo
darryn patch wrote:
Precisely half you pixel peepers wouldn't have a **** clue how to use a camera in the field, your too busy discussing bit depth noise moire and anything else totally useless to the day to day use of a camera, precisely why websites go down the shitter, and you get posts on here day in day out is my lens soft is it front focusing where do I put a lens cap after I have taken it off?
If an object is moving accross the focal plane it is going to be very difficult for a camera set to AI to PREDICT movement as the subject is not moving at a constant pace, Thats is what Ai is designed to do, predict where the focus needs to be based on a constant rate of movement. A camera set to AI tells the camera to lock onto a subject predict its rate of movement then predict wherethe focus will be when the shutter button is pressed and teh image is no longer seen as the mirror has flipped up. That is what AI is for. Not some person walking sideways and rattling off 10 shots, how can the camera grab focus when it can detect no movement?
Now if you don't beleive that pick up that black thing called a camera and go and test this yourself half of you might actually learn something.
You tell us Einstein.
Post like this are half the reason I and others are done with this site. This will be my last post on this topic. Having shot motorsport and all types of sport for 10 years I think I have a better understanding of AF than most here....Show more →
ouch.. that's a DPResponse.
I think I'll step put of this thread now as well. Nothing wrong with a healthy discussion, but when the tone drops to this level.. we... that's it really.
Scroll down to EOS1D instructions and download the manual for 1 series body, Then go to page 62 AI Sevo for moving subjects, where it states exactly what I have written. NOTE THE KEY WORDS approaches or retreats doesn't say moves sideways does it.
Then perhaps you people will see that what I have said is correct, and you've now learnt something.
Before you comment on AF again read the information provided by Canon as to the correct use of AI.
p.5 #14 · Pre-production 1Dmk4 tests with Ai-servo
darryn patch wrote:
Scroll down to EOS1D instructions and download the manual for 1 series body, Then go to page 62 AI Sevo for moving subjects, where it states exactly what I have written. NOTE THE KEY WORDS approaches or retreats doesn't say moves sideways does it.
You're not thinking this through enough.
If a subject moves laterally (sideways), it is also changing its distance from you, and therefore is also approaching/retreating from you.
p.5 #15 · Pre-production 1Dmk4 tests with Ai-servo
DavidP wrote:
You're not thinking this through enough.
If a subject moves laterally (sideways), it is also changing its distance from you, and therefore is also approaching/retreating from you.
QED
You might not be thinking that thru enough either.
Sideways motions in polar coordinates with with constant radii means no focal distance change, as long as we reduce the camera to a fulcrum point.
p.5 #16 · Pre-production 1Dmk4 tests with Ai-servo
PetKal wrote:
You might not be thinking that thru enough either.
Sideways motions in polar coordinates with with constant radii means no focal distance change, as long as we reduce the camera to a fulcrum point.
So what is happening, AI wise, when a person moves is in polar coordinates with constant radius, and the focal plane is, a plane?
p.5 #17 · Pre-production 1Dmk4 tests with Ai-servo
darryn patch wrote:
NOTE THE KEY WORDS approaches or retreats doesn't say moves sideways does it.
Unless the subject is orbiting the camera, an object will approach and retreat the camera if it's traveling "sideways" (or perpendicular) to the camera. Maybe not directly, but it will get closer, then further.
Anyways, how frequently do football players travel toward the photographer? They go from side to side with a photographer on the sidelines. I would say that's "sideways" travel.
p.5 #19 · Pre-production 1Dmk4 tests with Ai-servo
IMHO using "ring of fire" with object/subject not filling the frame or the "ring" is not going to work. Let alone with clutter BG with higher contrast than the intended object/subject to be shot. I bet Canon has not changed their AF logic
p.5 #20 · Pre-production 1Dmk4 tests with Ai-servo
PetKal wrote:
Sideways motions in polar coordinates with with constant radii means no focal distance change, as long as we reduce the camera to a fulcrum point.
True. DUCKS might fly around you at constant radius.
you take your camera into the tub with you
? well as long as you only post the ducks 
