Ted - Bruce wrote:
All the AF points you mention are irrelevant. He's only using the single middle one. Back in the non-autofocus film days, this is how we shot everything. Manually focus on the point we wanted to have in sharp focus and then recompose.
The only issue I see is that if he locked the focus point incorrectly, then the focus will be off. That is the advantage of the BBF (back button focus) technique as it breaks the link between the shutter release and the camera's focus control so both can be operated independently of each other. Chances of nailing the focus by going to continuous focus, rather than single shot mode give the photographer a better chance of nailing the shot. It does take a bit of practice to learn the technique, but once it is automatic, the number of shots with a blown focus decrease substantially.
The problem here is that George argues in terse phrases with little clarity - he does have a point though, as can be seen in the conductor shot. The conductor is an irregular shape where all points are at different distances from the camera image plane.
On the other hand, the rebuttal seems to refer to a flat plate parallel to the image plane, where "a few degrees" (presumably at some distance from the camera) has no visible effect. Or is a curved plate, as implied by "exactly the same distance" from the subject.
It only takes a simple diagram to determine the change of focus caused by camera rotation and it only takes a quick DOF calculation to determine whether that change of focus is significant or not.
George is good at drawing - it should be easy enough to illustrate the change of focus with some numbers ...
... I like numbers ...
Which is cross-type, is it not?! So why did you put a post up talking only about horizontal and vertical AF as if they were somehow separate, Manfred? That is what confused me ... not how Bruce took his shot!
The whole paragraph is irrelevant to my question. I was referring to your simple description of PDAF which seemed not to account for cross-type AF sensors of which Bruce's D7000 has 9.Back in the non-autofocus film days, this is how we shot everything. Manually focus on the point we wanted to have in sharp focus and then recompose.
The only issue I see is that if he locked the focus point incorrectly, then the focus will be off. That is the advantage of the BBF (back button focus) technique as it breaks the link between the shutter release and the camera's focus control so both can be operated independently of each other. Chances of nailing the focus by going to continuous focus, rather than single shot mode give the photographer a better chance of nailing the shot. It does take a bit of practice to learn the technique, but once it is automatic, the number of shots with a blown focus decrease substantially.
Did your post account for cross-type AF sensors or did it not? I think not, in which case, it did not apply to Bruce's perfectly good camera.You said in post #35: With phase detect autofocus, what works best are vertical lines (i.e. contrasty areas) when looked at with the camera in horizontal / landscape orientation. Put the camera in vertical / portrait orientation, look for horizontal lines. Phase detect can have a hard time focusing on plain coloured (low texture) building siding if it is in horizontal orientation and the camera is in landscape orientation.
Last edited by xpatUSA; 18th December 2018 at 12:32 AM.
Ted -the sensor type cross or regular is largely irrelevant to the discussion. Cross type sensors can provide higher accuracy in certain situations than regular ones, that's it. They tend to be used only when shooting fairly wide open, a Bruce was. Stop the camera down a bit and the old-fashioned linear sensors. The discussion is about the techniques that best take advantage of the technology built into the camera.
Look at a high end medium format camera (Hasselblad HD6-100C, Leica S or Mamiya 645 DF); the first two have a single central cross autofocus while the Mamiya has 3 cross sensors. These are $US 20 000 - $US 30 000 cameras. None of these fancy arrays of sensors. Photographers shooting these either manually focus or use very simple autofocus mechanisms.
The D7000 has 9 cross and 30 vertical linear sensors. Bruce only used one of these cross sensors in his shots; he was shooting close to wide open. Stop down a couple of stops and the camera only uses the linear sensors, with the inherent issues I mentioned before.
The question is really about understanding the best way to use the technology, rather than the technology itself.
D is the change of distance. Yes, I've used this one before and probably will use it later too
It's real world, especial with f/2 and moving objects.
George
And I really thought you were going to come up with a figure such as (example),
if you shoot from a distance of 3 metres with an 85mm lens, focus on a persons eye when his ear is touching the left side of the frame, lock focus and recompose so he is in the centre of the frame your focus will be out by 'x' mm on the eye'
So here's the real world relevance,
Taking Bruce's image of the conductor, if the conductor had been almost far left of the frame and you had focused on his eye, and then recomposed so he is in the frame centre the distance 'sensor to eye' would have altered by 60mm.
At f/1.8 (shown in my exif but image shows f/2) at the 10 metre distance this was shot at (assuming it has not been cropped) the DoF is 970mm (at f/2 DoF is 1400mm).
This example is at the extreme as it's very unlikely anyone is going to recompose to that extent.
Last edited by Stagecoach; 18th December 2018 at 09:03 AM.
George to get the geometry in this diagram, the camera would literally have to be on the subject's nose while focusing on the eyeball.
In real life you will have two lines that are almost parallel moving at a few degrees apart. The distance from the sensor to the subject (using the sensor as the plane of reference) will essentially be the same length and any minor issues would be within the autofocus range of accuracy. From the camera sensor standpoint, the distance would be the same.
With Bruce's APS-C sensor is 24mm, and in reality, it would probably be in the order less than 6 mm in the example that he posted. If the subject was around 5m away, from the sensor perspective, both the original focus point and the recomposed image would be about 5000 mm away from the sensor.
Last edited by Manfred M; 18th December 2018 at 01:51 PM.
It's a diagram. The only values in this diagram are the angel of rotation and the angle of view. There's no way to see any distance. I don't understand the discussion about such a simple thing. It's something a photographer should know. And not something to be scared of.
Graham,
Just google on "focus and recomposing". One of the hits was https://digital-photography-school.c...ompose-method/. That gives a "real world" example. I'm sure you'll find more.
George
I'm beginning to understand, thank you Manfred. Bruce shot with a cross-type AF sensor but you changed the discussion to linear type which somehow made my question "largely irrelevant".
That I did not know. Hope it does not apply to my cameras! They're bad enough at AF already without that millstone ...Stop the camera down a bit and the old-fashioned linear sensors [?]. The discussion is about the techniques that best take advantage of the technology built into the camera.
Look at a high end medium format camera (Hasselblad HD6-100C, Leica S or Mamiya 645 DF); the first two have a single central cross autofocus while the Mamiya has 3 cross sensors. These are $US 20 000 - $US 30 000 cameras. None of these fancy arrays of sensors. Photographers shooting these either manually focus or use very simple autofocus mechanisms.
The D7000 has 9 cross and 30 vertical linear sensors. Bruce only used one of these cross sensors in his shots; he was shooting close to wide open. Stop down a couple of stops and the camera only uses the linear sensors, with the inherent issues I mentioned before.
My naïve question to your goodself was specifically about AF technology itself, trying to relate your description of linear (regular) sensors to Bruce's use of a cross-type sensor. As can sometimes occur, we appear to have drifted off from that sub-topic, so I'll bid thee farewell for now. :-)The question is really about understanding the best way to use the technology, rather than the technology itself.
Last edited by xpatUSA; 18th December 2018 at 04:55 PM.
When I focus with my camera on my window, the framework, then there is a difference in the cross type and the non cross-type. Forgive me for not remembering which one, but the non cross-type had problems with either the vertical lines or the horizontal lines. But in this discussion, it focusses or it doesn't. There's no such thing as bad focussing.
Maybe informative https://digital-photography-school.c...cusing-points/.
George
Last edited by george013; 18th December 2018 at 04:34 PM.
Indeed, that's how all my cameras work.
I wondered about that but I don't own a Nikon D7000.So stepping down a stop doesn't have any influence.
I wondered about that too ... seems like there's some mis-information in this thread; hope Bruce can figure it all out for himself, eh?And a cross-type doesn't change [into] a non cross-type.
Last edited by xpatUSA; 18th December 2018 at 06:52 PM.
I can only assume George that it was the stick figure diagram and the description of what he was doing that attracted you to that link
I found the funniest bit in it that he states he's no mathematician, perhaps that's why he chose the most obvious extreme shooting example to 'generalise' something
Indeed, Grahame, using f/1.4 at only 4 feet is applying a serious amount of "spin" in the opposite direction to the recently stated "If the subject was around 5m away, from the sensor perspective, both the original focus point and the recomposed image would be about 5000 mm away from the sensor"