[K David]

I can't find this in the K-1 manual or in the camera's menus (maybe I'm not using the right terms or looking well enough), but does anyone know if it's possible for manually adjust the K-1's sensor articulation? I know I could in my K-7 to correct for verticals, for instance, and I think I could in my K-3. But I have want to do some work with my home-made tilt-shift bellows and being able to adjust the sensor would be like having an additional level of rear tilt.

[photoptimist]

Look for "composition adjustment" on page 71 of the manual. But it's only a shift & rotation adjustment. No IBIS system currently on the market provides a way of tilting the sensor.

[LensBeginner]

I'll add that this kind of adjustment can mimick a shift lens only if the lens image circle covers the whole sensor and then some - it's perfect for vintage FF lenses on APS-C, for instance.
I don't have the option to correct for verticals on my K-30, are you sure you have that on the K-7?

[funktionsfrei]

No IBIS system currently on the market provides a way of tilting the sensor.

Interesting in that aspect is how they manage to achieve "5-axis stabilization" with only three axes of movement for the sensor ...

[photoptimist]

Interesting in that aspect is how they manage to achieve "5-axis stabilization" with only three axes of movement for the sensor ...

Yes it is interesting.

What's happening is that the IBIS is compensating for the effects of camera tilt on the location of the image on the sensor.

Think of it this way: if the camera and lens tilt down just bit, the scene in the viewfinder (and on the sensor) shifts upwards just a bit, shifting the sensor up to track the upward shift of the scene then keeps the scene stable in the image.

The sensor does not have to be tilted to correct for a tilt of the camera-and-lens. In fact, a tilt of the sensor can't correct for a tilt of the camera-and-lens.

[reh321]

Interesting in that aspect is how they manage to achieve "5-axis stabilization" with only three axes of movement for the sensor ...

We already know that is short-hand for some other construct. We can "operate in" {control} only three dimensions - any physicist or mathematician will tell you that any motion or position can be provided by three axes. They are using this language to explain something else.

[lsimpkins]

We already know that is short-hand for some other construct. We can "operate in" {control} only three dimensions - any physicist or mathematician will tell you that any motion or position can be provided by three axes. They are using this language to explain something else.

Yes, it seems that in marketing land, rotation around the lens axis and yaw of the body are newly discovered spatial axes.

[texandrews]

We already know that is short-hand for some other construct. We can "operate in" {control} only three dimensions - any physicist or mathematician will tell you that any motion or position can be provided by three axes.....

Well, not in the 4th or 5th (or more) dimensions, surely. Any science fiction reader will tell you that... ;-} It's why no one but Pentaxians understand Pentax---because we are from another dimension, natch. Should be a T-shirt...

[LensBeginner]

We already know that is short-hand for some other construct. We can "operate in" {control} only three dimensions - any physicist or mathematician will tell you that any motion or position can be provided by three axes. They are using this language to explain something else.

But you have to have three linear and three rotational axes to be able to move in every possible way in a three-dimensional space (6DoF - that's not Depth of Field! ).

[photoptimist]

The whole "number of axes" issue really isn't "marketing," it's basic geometry.

As LensBeginner said, there really are six dimensions that define both the camera's location in space (x,y,z) and the camera's orientation in space (pitch, yaw, roll). That implies there are six different kinds of shake or vibrational effects that might afflict the camera during exposure. And it also means that different stabilizer technologies might compensate for some types of shake but not others.

For example, the easiest dimensions to stabilize are pitch and yaw tilts because it's relatively easy to measure tilting motion and relatively easy to estimate the compensating adjustment (which depends on focal length).

Stabilizing roll is extremely easy for an in-body stabilizer but it's impossible to do with an in-lens stabilizer.

Stabilizing up-down or left-right shifts is harder both because accurately measuring shift motions are tricky and computing the compensating motion is hard (it depends on BOTH focal length and subject distance).

Stabilizing closer-further shifts is also hard because accurately measuring shift motions are tricky and compensating for them requires adjustment in focus.

[reh321]

The whole "number of axes" issue really isn't "marketing," it's basic geometry.

As LensBeginner said, there really are six dimensions that define both the camera's location in space (x,y,z) and the camera's orientation in space (pitch, yaw, roll). That implies there are six different kinds of shake or vibrational effects that might afflict the camera during exposure. And it also means that different stabilizer technologies might compensate for some types of shake but not others.

The question I was responding to originally was how they could have so many axes in the world we inhabit. The answer is that they are re-defining the word "axis". In basic geometry "pitch", "yaw", and "roll" are actually rotation around those same x, y, z axes that define location in space. Yes, there are six effects that they have to deal with, three of "translation" and three of "rotation", but those effects involve just three axes. Instead of saying "5-1/2 effects" or some such thing, marketing guys choose to mis-talk about "5-1/2 axes"

[SteveinSLC]

Plus it also has to account for the 4th dimension...time. More specifically movement over time or acceleration. So not only which direction the camera is moving, but how fast it is moving and is that changing?

[photoptimist]

The question I was responding to originally was how they could have so many axes in the world we inhabit. The answer is that they are re-defining the word "axis". In basic geometry "pitch", "yaw", and "roll" are actually rotation around those same x, y, z axes that define location in space. Yes, there are six effects that they have to deal with, three of "translation" and three of "rotation", but those effects involve just three axes. Instead of saying "5-1/2 effects" or some such thing, marketing guys choose to mis-talk about "5-1/2 axes"

The concept of 6-axes (3 for position, 3 for orientation) is prevalent in aerospace engineering, robotics, CNC machining, and control theory. It's not something invented by camera marketing guys.

And if you think about a graph having different independent axes, then the graph defining all possible camera positions and orientations has six axes.

You are right that "5-1/2 axes" is a bit nonsensical!

[reh321]

The concept of 6-axes (3 for position, 3 for orientation) is prevalent in aerospace engineering, robotics, CNC machining, and control theory. It's not something invented by camera marketing guys.

And if you think about a graph having different independent axes, then the graph defining all possible camera positions and orientations has six axes.

You are right that "5-1/2 axes" is a bit nonsensical!

Then the answer to the original question is "it is notational" - the world we are familiar with has just three dimensions and can be "spanned" {math talk for "described"} by just three axes.

[photoptimist]

Then the answer to the original question is "it is notational" - the world we are familiar with has just three dimensions and can be "spanned" {math talk for "described"} by just three axes.

Yes, it is notational.

In math, engineering, and science, the number of "axes" is often a synonym for the number of different variables, especially the number of independent variables (often called "degrees of freedom") that are being controlled by the experimenter or user. Similarly the number of "dimensions" in a math problem is the number of variables which might include more than just the three dimensions of physical space (or the four of space-time).

If someone wants to replicate an image, they need to know both the three values of the camera's position variables and the three values of the camera's orientation variables because a change in any of those six variables, axes, or dimensions changes the image. And if someone wants to stabilize an image, they need to know the changes of up to six different variables or axes. As a mathematical problem in control theory, stabilizing an image has up to six dimensions.

P.S. Another way to look at this in the context of having only three physical dimensions is that one must know or control BOTH the 3-D location of the camera body and the 3-D location of the lens (for a total of six dimensions or degrees of freedom) to describe what the image will look like.