[K David]

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.

Thank you, that's it!

are you sure you have that on the K-7?

I was thinking of shift, and of correcting for framing. I mis-spoke.

[LensBeginner]

I was thinking of shift, and of correcting for framing. I mis-spoke.

Ah ok! at least they didn't take an useful feature away... we could do with tilt perspective correction... especially with UWAs...

[funktionsfrei]

A tilting sensor would rather likely collide with the focal plane shutter and the mirror mechanism which are integral parts of a SLR.

It would be possible if used only in LV mode with full electronic shutter.

[victormeldrew]

Well, not in the 4th or 5th (or more) dimensions, surely. Any science fiction reader will tell you that...

Any Quantum Physicist will tell you that there are at least 10 dimensions...maybe as many as 26. But most of them being only really relevant if you are a subatomic particle, compensating for them all in-camera is not entirely necessary.

But at least 4 dimensions (x,y,z,t) are essential to describe linear movement and 3 axes (plus time again) for rotation - you can use any 3 axes, not necessarily the same x,y,z. For a body moving both linearly and rotationally (as any handheld object will be) at least 7 dimensions are required to adequately describe (and therefore compensate for) it, 4 if you're using the same axes for both.

[photoptimist]

A tilting sensor would rather likely collide with the focal plane shutter and the mirror mechanism which are integral parts of a SLR.

It would be possible if used only in LV mode with full electronic shutter.

Quite true!

A tilting sensor for a DSLR could be built to tilt away from shutter, mirror, and lens. This would tilt the plane of focus closer to the camera.

A mere 5 millimeters of sensor tilt would add huge functionality. For example, a 50mm lens focused at infinity and one side of the sensor tilted back 5mm would have a plane of focus extending from about 1 meter to infinity.

[funktionsfrei]

A tilting sensor for a DSLR could be built to tilt away from shutter, mirror, and lens. This would tilt the plane of focus closer to the camera.

Yes, that would be possible, but somewhat complicated, and you'd lose the ability to focus to infinity with anything but the sensor's edges.

That's somewhat different to what optical bench cameras do. When they tilt the sensor (film, that is), they rotate it around a horizontal or vertical axis that runs through the sensor's centre, thus moving one half of the sensor back from the lens but the other half toward it. The axis of rotaion of the sensor remains in the focal plane.


Your concept would mean a rotation around one of the edges, thus moving the whole sensor back and only the edge around which the sensor is being rotated remaining in focal plane.


Interesting as the idea is, I'm afraid it would be somewhat complicated to achieve, because you'd have to implement four separate axes of rotation (lower/upper edge and left/right edge, though but one of each pair would be used for obvious reasons). That would add to the camera's depth, because all of this would have to happen behind the current plane of movement for the sensor (towards the photographer), thus making the camera even deeper (or fatter) than the K-1 already is.

But having the sensor tilt through the focal plane shutter would also be highly complicated and critical, beginning that the aperture of the shutter would have to be much larger than the usual 24x36mm to accomodate the sensor's outer dimensions ...


That doesn't seem feasible either.

Another approach would be tilting the combined shutter/sensor assembly, here the severest constraint would be the mirror's mechanics which would impede the sensors upper edges' movement towards the lens. That problem might be solved by increasing the dimensions of the mirror box. With such a construction, I doubt that it would be possible to use the tilt mechanism for image stabilization (due to the severely increased mass by adding the shutter) and ... I'm afraid that the resulting camera would be somewhat in the dimensions of the Pentax 67. Good to throw at a stampeding elephant herd ...

Maybe sensor tilt would be something where those evil mirrorless cameras are definitively at an advantage.

Until then, I just use my Samyang 24TS.

[reh321]

Any Quantum Physicist will tell you that there are at least 10 dimensions...maybe as many as 26. But most of them being only really relevant if you are a subatomic particle, compensating for them all in-camera is not entirely necessary.

But at least 4 dimensions (x,y,z,t) are essential to describe linear movement and 3 axes (plus time again) for rotation - you can use any 3 axes, not necessarily the same x,y,z. For a body moving both linearly and rotationally (as any handheld object will be) at least 7 dimensions are required to adequately describe (and therefore compensate for) it, 4 if you're using the same axes for both.

I hope you don't have a degree in physics.

In physics only three axes are required to determine location and rotation in the reality we inhabit.

The is the genesis of the discussion that started this subdiscussion.

[victormeldrew]

In physics only three axes are required to determine location and rotation in the reality we inhabit.

SR needs to determine movement, not location. Therefore a minimum of 4 dimensions are required as I said.

I hope you don't have a degree in physics.

Not Pure Physics, no. My degree is in Mechanical Engineering...so essentially Applied Physics.

[reh321]

SR needs to determine movement, not location. Therefore a minimum of 4 dimensions are required as I said.

The discussion started off trying to explain "5-1/2 axes", not a mere 4 axes, to someone who lives in our world; in post #19, you were trying to shoe-horn seven axes in - that is what I was objecting to.

@photoptimist has already explained the "5-1/2 axes" as an engineering construct. There is no value for anything beyond that other than spreading confusion while increasing the post count.

[photoptimist]

Yes, that would be possible, but somewhat complicated, and you'd lose the ability to focus to infinity with anything but the sensor's edges.

That's somewhat different to what optical bench cameras do. When they tilt the sensor (film, that is), they rotate it around a horizontal or vertical axis that runs through the sensor's centre, thus moving one half of the sensor back from the lens but the other half toward it. The axis of rotaion of the sensor remains in the focal plane.


Your concept would mean a rotation around one of the edges, thus moving the whole sensor back and only the edge around which the sensor is being rotated remaining in focal plane.


Interesting as the idea is, I'm afraid it would be somewhat complicated to achieve, because you'd have to implement four separate axes of rotation (lower/upper edge and left/right edge, though but one of each pair would be used for obvious reasons). That would add to the camera's depth, because all of this would have to happen behind the current plane of movement for the sensor (towards the photographer), thus making the camera even deeper (or fatter) than the K-1 already is.

But having the sensor tilt through the focal plane shutter would also be highly complicated and critical, beginning that the aperture of the shutter would have to be much larger than the usual 24x36mm to accomodate the sensor's outer dimensions ...


That doesn't seem feasible either.

Another approach would be tilting the combined shutter/sensor assembly, here the severest constraint would be the mirror's mechanics which would impede the sensors upper edges' movement towards the lens. That problem might be solved by increasing the dimensions of the mirror box. With such a construction, I doubt that it would be possible to use the tilt mechanism for image stabilization (due to the severely increased mass by adding the shutter) and ... I'm afraid that the resulting camera would be somewhat in the dimensions of the Pentax 67. Good to throw at a stampeding elephant herd ...

Maybe sensor tilt would be something where those evil mirrorless cameras are definitively at an advantage.

Until then, I just use my Samyang 24TS.

Because three points determine a plane, the sensor tilt mechanism only needs three axes of motion. A short leadscrew motion system on ball joints would be attached to two corners and one opposite middle edge of the sensor assembly.

Infinity focus across the entire sensor is always possible just by setting all the motion axes to the zero tilt setting.

Each leadscrew motion mechanism doesn't need to be very big because it's only moving a short distance and does not need to be very fast. Just take apart a compact point-n-shoot camera and you'll often find a very compact leadscrew mechanism in the zoom lens. Each motor, gearbox and screw might be about 3 cubic centimeters and be folded to sit next the sensor assembly. Yes, the camera would need to be a bit thicker to accommodate the added volume of the tilting sensor mechanism (call it 8 cm x 4 cm x 0.7 cm = 17 cc. But this added thickness would be no greater than about 7 or 8 mm and could be reduced if more circuitry were moved from behind the sensor to beside the sensor assembly. The current central mass of the K-1 is over 450 cc and this mechanism probably adds only about 40 cc to that.

It does not have to be that complex or bulky.

[funktionsfrei]

Because three points determine a plane, the sensor tilt mechanism only needs three axes of motion. A short leadscrew motion system on ball joints would be attached to two corners and one opposite middle edge of the sensor assembly.

You're obviously quite right, I was thinking a mite too complicated.

I'm afraid though, that the loss of infinity focus when tilting might still be a limitation, at least for some applications.

[victormeldrew]

The discussion started off trying to explain "5-1/2 axes", not a mere 4 axes, to someone who lives in our world; in post #19, you were trying to shoe-horn seven axes in - that is what I was objecting to.

@photoptimist has already explained the "5-1/2 axes" as an engineering construct. There is no value for anything beyond that other than spreading confusion while increasing the post count.

If you look at my original post, I was quoting (and correcting) the assertion by texandrews that the 4th and 5th spacial dimensions were the preserve of science fiction writers. The 4th is most certainly not and the 5th and up are the preserve of quantum mechanics, so if you subscribe to string theory they are as factual as any other.

There's nothing wrong with correcting misconceptions as they arise during a discussion...especially if it increases the post count.


In terms of SR, the question has already been answered that the marketeers have come up with a term (axes) that really means 'Degreees of Freedom' - there can be as many of those as you like to engineer in...a robot arm can have a dozen or more. The human hand alone has 27 degrees of freedom (or axes if you're in the Pentax marketing department).

[photoptimist]

You're obviously quite right, I was thinking a mite too complicated.

I'm afraid though, that the loss of infinity focus when tilting might still be a limitation, at least for some applications.

There's no loss of infinity focus if the sensor motion platform is in the zero-tilt setting. With the sensor pushed forward toward the lens as far as it can safely go (without hitting the shutter) the entire sensor surface is at the infinity focal plane location.

The entire tilting mechanism can designed to sit either behind or beside the sensor-shutter area so that the sensor can be as close the shutter as it is now.

[Gimbal]

The entire SR assembly (which is a lot larger than the actual sensor) must also tilt. That means that the sensor will move backwards from the infinity focus plane as soon as you tilt even the slightest. Not one pixel will be able to reach infinity focus unless you are in zero-tilt setting.

[reh321]

The entire SR assembly (which is a lot larger than the actual sensor) must also tilt. That means that the sensor will move backwards from the infinity focus plane as soon as you tilt even the slightest. Not one pixel will be able to reach infinity focus unless you are in zero-tilt setting.

Do the view cameras with tilt mechanism have leaf shutters or do they use "focal plane" shutters?