Nikon maintains DSLR lead over Canon

Neil H. wrote:

"Chris Malcolm" wrote in message
...
Neil H. wrote:

Interesting. I wonder how (or even whether) anyone has studied exactly
which
directions of camera movement contribute what amount to camera shake
visible
in the photo. It seems to me it would be an extremely difficult thing to
evaluate, since the camera-moving part is human and presumably not very
repeatable.

But the geometry of the optics doesn't change, and there are huge
differences as a result of that. How close the object is matters as
well. For example, think of taking a macro shot. Shifting the camera
0.1mm vertically upwards will have a very large effect compared to the
negligible effect of shifting vertically upwards 0.1mm when
photographing a landscape. In the case of the landscape, and in
general, rotational movement is by far the most important. A very
small shift in the direction the camera is pointing will have a large
effect on the image.

Sure, but what I'm wondering about is the relative effects of the three axes
of rotation. I think that yaw and pitch would have the greatest effects on
blur, and roll would have much less. But how to quantify them?

If you quantify them as I did above then roll about the axis of sight
has far less effect than yaw and pitch on the image. It has comparable
effects to linear position shifts.

Suppose for example you're using a lens with a field of view of 30
degrees, and suppose you wish to produce an image which is sharp at
300 ppi on a 10 inch wide print. The span of your field of view is
divided into 3000 pixels. So a rotational shift of 30/3000 = 1/100th
of a degree will move the image over one pixel, and you'd probably
have to hold it to within 1/1000th of a degree to keep the softening
of the image due to shake negligible.

That's an interesting way of looking at it, but I'm not sure how useful it
is. In order to translate any of that into required shutter speed you'd have
to know the *speed* of camera movement (i.e., angular velocity) as well.
That I think would be difficult to arrive at, with the movement in human
hands.

It's not difficult to bound it with reasonable general maxima and
minima. Note too that for a given shakiness and strength of hands,
then the heavier the camera the more slowly it will move, up to a
maximum which is probably heavier than any hand-held camera unless
you're pathologically weak or shaky.

--
Chris Malcolm DoD #205
IPAB, Informatics, JCMB, King's Buildings, Edinburgh, EH9 3JZ, UK
[http://www.dai.ed.ac.uk/homes/cam/]

"Chris Malcolm" wrote in message
...
Neil H. wrote:

Interesting. I wonder how (or even whether) anyone has studied exactly
which
directions of camera movement contribute what amount to camera shake
visible
in the photo. It seems to me it would be an extremely difficult thing to
evaluate, since the camera-moving part is human and presumably not very
repeatable.

But the geometry of the optics doesn't change, and there are huge
differences as a result of that. How close the object is matters as
well. For example, think of taking a macro shot. Shifting the camera
0.1mm vertically upwards will have a very large effect compared to the
negligible effect of shifting vertically upwards 0.1mm when
photographing a landscape. In the case of the landscape, and in
general, rotational movement is by far the most important. A very
small shift in the direction the camera is pointing will have a large
effect on the image.

Sure, but what I'm wondering about is the relative effects of the three axes
of rotation. I think that yaw and pitch would have the greatest effects on
blur, and roll would have much less. But how to quantify them?


Suppose for example you're using a lens with a field of view of 30
degrees, and suppose you wish to produce an image which is sharp at
300 ppi on a 10 inch wide print. The span of your field of view is
divided into 3000 pixels. So a rotational shift of 30/3000 = 1/100th
of a degree will move the image over one pixel, and you'd probably
have to hold it to within 1/1000th of a degree to keep the softening
of the image due to shake negligible.

That's an interesting way of looking at it, but I'm not sure how useful it
is. In order to translate any of that into required shutter speed you'd have
to know the *speed* of camera movement (i.e., angular velocity) as well.
That I think would be difficult to arrive at, with the movement in human
hands.

Neil

On Jul 21, 8:03 am, Chris Malcolm wrote:
Neil H. wrote:
"Chris Malcolm" wrote in message
...
Neil H. wrote:

Interesting. I wonder how (or even whether) anyone has studied exactly
which
directions of camera movement contribute what amount to camera shake
visible
in the photo. It seems to me it would be an extremely difficult thing to
evaluate, since the camera-moving part is human and presumably not very
repeatable.

But the geometry of the optics doesn't change, and there are huge
differences as a result of that. How close the object is matters as
well. For example, think of taking a macro shot. Shifting the camera
0.1mm vertically upwards will have a very large effect compared to the
negligible effect of shifting vertically upwards 0.1mm when
photographing a landscape. In the case of the landscape, and in
general, rotational movement is by far the most important. A very
small shift in the direction the camera is pointing will have a large
effect on the image.
Sure, but what I'm wondering about is the relative effects of the three axes
of rotation. I think that yaw and pitch would have the greatest effects on
blur, and roll would have much less. But how to quantify them?

If you quantify them as I did above then roll about the axis of sight
has far less effect than yaw and pitch on the image. It has comparable
effects to linear position shifts.

Suppose for example you're using a lens with a field of view of 30
degrees, and suppose you wish to produce an image which is sharp at
300 ppi on a 10 inch wide print. The span of your field of view is
divided into 3000 pixels. So a rotational shift of 30/3000 = 1/100th
of a degree will move the image over one pixel, and you'd probably
have to hold it to within 1/1000th of a degree to keep the softening
of the image due to shake negligible.
That's an interesting way of looking at it, but I'm not sure how useful it
is. In order to translate any of that into required shutter speed you'd have
to know the *speed* of camera movement (i.e., angular velocity) as well.
That I think would be difficult to arrive at, with the movement in human
hands.

It's not difficult to bound it with reasonable general maxima and
minima. Note too that for a given shakiness and strength of hands,
then the heavier the camera the more slowly it will move, up to a
maximum which is probably heavier than any hand-held camera unless
you're pathologically weak or shaky.

As anyone who shoots a gun can tell you, there is a vast difference in
movement characteristics depending on how you control things like
breathing, your stance, you hand position on the camera/lens, etc.
Consistency is the key, once you've isolated all the variables.

Philip Homburg wrote:
Wolfgang Weisselberg wrote:
Jeremy Nixon wrote:

I'd rather have no metering, than have the lens stopped down to shooting
aperture while trying to compose the shot.

First compose, then stop down, then meter. Like, you know,
first pillage, *then* burn.

So instead of getting a D200 that meters just fine with non-cpu Nikkors,
you get a camera that forces you to use stop-down metering?

Wow, way to go.

I show a way how to compose w/o stopping down and you want to
saddle me with non-cpu nikkors and no D200 --- which, being a
Canon user, I have no need for?

-Wolfgang