EM1 4.0 Firmware Upgrade

[scott kirkpatrick]

Delightful upgrade choices. I was wondering after playing with the HRes mode, in which they move the imaging chip around in the x and y directions, when would they think of moving the chip in the z-direction, to change the point of focus. Now they have, and it seems pretty significant, if 7% of the image around the edges will be sacrificed in order to shift the focal distance. (That should tell us how much movement they plan to make available, if anyone wants to do the calculation.) I suppose the reassembly uses the focus peaking logic to determine what parts of each of the 8 images to retain in the final result. Can't wait to try it out.

And for Bart, not only could you always set the red video button to "OFF, dammit!" , now you can remove all those nagging electronic displays from the viewfinder. What's not to like?

scott

 

[k-hawinkler]

Delightful upgrade choices. I was wondering after playing with the HRes mode, in which they move the imaging chip around in the x and y directions, when would they think of moving the chip in the z-direction, to change the point of focus. Now they have, and it seems pretty significant, if 7% of the image around the edges will be sacrificed in order to shift the focal distance. (That should tell us how much movement they plan to make available, if anyone wants to do the calculation.) I suppose the reassembly uses the focus peaking logic to determine what parts of each of the 8 images to retain in the final result. Can't wait to try it out.

And for Bart, not only could you always set the red video button to "OFF, dammit!" , now you can remove all those nagging electronic displays from the viewfinder. What's not to like?

scott

Thanks. So Scott, you think they will move the sensor in the z-direction as well? But no tilt for now, right? :grin:

 

[Brian Mosley]

Interesting... if depth information could be captured by shifting the sensor along the z axis, presumably it would be possible to simulate a more narrow depth of field?

Cheers

Brian

 

[dhsimmonds]

Very good news indeed, I'm particularly looking forward to the focus stacking ...
C U
Rafael

Focus stacking is a great addition...interesting to note the press release states it's only compatible with 60mm Macro, 12-40 & 40-150mm PRO lenses though.

 

[Godfrey]

Nice stuff.

G

 

[scott kirkpatrick]

Thanks. So Scott, you think they will move the sensor in the z-direction as well? But no tilt for now, right? :grin:

Well, how do you do focus stacking? You take a series of images with the camera focused at different distances. They could do it with the autofocus mechanism, which in the E-M1 can use the phase-sensitive detection. This mechanism has the advantage of knowing whether the focus has moved closer or further, while contrast detection uses an error signal that increases both ahead and behind of focus. The announcement is apparently only for the M-1, and only supported on three lenses, which does sound like they will actually refocus the lens. But maybe this is being restricted to the M-1 for marketing timing reasons, and they actually intend to make it available for all models. That would make the shift-in-z more plausible. I'm hoping that the second case holds.

It's surprising that they don't make the high resolution mode possible for the M-1 in this upgrade. Maybe the improved stabilization that is reported for the M5.2 is the result of better hardware, not just fancier firmware.

scott

 

[Vivek]

Focus stacking is a great addition...interesting to note the press release states it's only compatible with 60mm Macro, 12-40 & 40-150mm PRO lenses though.

That is 2 more lenses than I thought!

 

[Vivek]

Thanks Bart, just making sure I am not dreaming! :bugeyes:
This feels like having Christmas twice this year, the first one in November!
Am I glad that I still have all 3 OM-D models. :grin:

But it will be hard to tear myself away from the A7r2. :banghead: :ROTFL: :thumbs:

What if Sony offers the same feature (focus stacking) in the near future (A7rII, etc)? :bugeyes:

 

[k-hawinkler]

Well, how do you do focus stacking? You take a series of images with the camera focused at different distances. They could do it with the autofocus mechanism, which in the E-M1 can use the phase-sensitive detection. This mechanism has the advantage of knowing whether the focus has moved closer or further, while contrast detection uses an error signal that increases both ahead and behind of focus. The announcement is apparently only for the M-1, and only supported on three lenses, which does sound like they will actually refocus the lens. But maybe this is being restricted to the M-1 for marketing timing reasons, and they actually intend to make it available for all models. That would make the shift-in-z more plausible. I'm hoping that the second case holds.

It's surprising that they don't make the high resolution mode possible for the M-1 in this upgrade. Maybe the improved stabilization that is reported for the M5.2 is the result of better hardware, not just fancier firmware.

scott

Thanks Scott.
Do we have any information that their sensor actually is designed to move in the Z direction or only within the X and Y plane?

 

[scott kirkpatrick]

Thanks Scott.
Do we have any information that their sensor actually is designed to move in the Z direction or only within the X and Y plane?

I think it has to be able to do this. Starting with the E-P5, Olympus has offered "5-axis" motion compensation. x and y shifts are only two. Pitch and yaw come next. A pitch correction is made by pulling back on the top of the sensor and pushing on the bottom (from behind). Yaw is done by pulling back on one side of the sensor and pushing out on the other. So to create a z-move, you push on both sides at the same time. The interesting axis is the roll correction, which offsets any tendency you might have to rotate the camera, say when pushing down hard on the shutter button. (Lenses with optical correction can't do this.) That takes actuators pushing and pulling together, but at 45 degree angles to the z-direction, one pushing, one pulling. If you have seen a full motion flight simulator, the cabin sits on top of several of these angled pairs of piston actuators.

Doing all this on a micron scale is already pretty awesome.

scott

 

[Annna T]

Delightful upgrade choices. I was wondering after playing with the HRes mode, in which they move the imaging chip around in the x and y directions, when would they think of moving the chip in the z-direction, to change the point of focus. Now they have, and it seems pretty significant, if 7% of the image around the edges will be sacrificed in order to shift the focal distance. (That should tell us how much movement they plan to make available, if anyone wants to do the calculation.) I suppose the reassembly uses the focus peaking logic to determine what parts of each of the 8 images to retain in the final result. Can't wait to try it out.

And for Bart, not only could you always set the red video button to "OFF, dammit!" , now you can remove all those nagging electronic displays from the viewfinder. What's not to like?

scott

Do you know for sure that they are moving the sensor and not the lens focusing mechanism ? I'd have thought it is easier to change focus with the lens rather than by moving the sensor : that one already has to take care of IBIS etc. and adding a further z axe seems more complicated than what they achieved with the HIRes mode (where they just makes use of the sensor aptitude to move for IBIS). .

 

[raist3d]

Guys, where on Earth you get the idea they are moving the sensor in the z direction? ;-)

No, focus stacking is done how you would- using the lens, different focus points. Changing z direction would require a major hardware change. Note focus stacking works only with specific lenses

 

[Brian Mosley]

Yes, of course it's nonsense.

Focus stacking will be done by moving the lens.

Cheers

Brian

 

[k-hawinkler]

I think it has to be able to do this. Starting with the E-P5, Olympus has offered "5-axis" motion compensation. x and y shifts are only two. Pitch and yaw come next. A pitch correction is made by pulling back on the top of the sensor and pushing on the bottom (from behind). Yaw is done by pulling back on one side of the sensor and pushing out on the other. So to create a z-move, you push on both sides at the same time. The interesting axis is the roll correction, which offsets any tendency you might have to rotate the camera, say when pushing down hard on the shutter button. (Lenses with optical correction can't do this.) That takes actuators pushing and pulling together, but at 45 degree angles to the z-direction, one pushing, one pulling. If you have seen a full motion flight simulator, the cabin sits on top of several of these angled pairs of piston actuators.

Doing all this on a micron scale is already pretty awesome.
T
scott

Thanks Scott. I don't think so.
To use the word axis is somewhat misleading IMHO.

You could do those 5 measurements and with computing projections do 3 image stabilization actions, namely shift sensor in X and Y, rotate within X-Y plane.

That would imply that no motion of the sensor in the Z direction is contemplated.

So focus with the normal focusing mechanism at the different distances would remain, no?

 

[scott kirkpatrick]

Yes, of course it's nonsense.

Focus stacking will be done by moving the lens.

Cheers

Brian

Not completely obvious. Take a look at the motions that the Olympus sensor is capable of in this Imaging Resource video. So z-motion is possible. And this 8-shot option will require use of a tripod to get the 8 shots in registry, so normal image stabilization will be turned off, as it is for the HighRes option on the M5.2.

scott

 

[raist3d]

Not completely obvious. Take a look at the motions that the Olympus sensor is capable of in this Imaging Resource video. So z-motion is possible. And this 8-shot option will require use of a tripod to get the 8 shots in registry, so normal image stabilization will be turned off, as it is for the HighRes option on the M5.2.

scott

The sensor doesn't move in the z plane. The Olympus marketing video is off in this one (if it's the video I am thinking). Look for the video demonstration of the physical sensor from one of the Olympus trade shows and you see the sensor can't move in the Z plane. Also consider all the major consequences that would have for the focus plane where the lens projects the image.

No, the sensor doesn't move in Z. I understand the marketing video from Olympus is confusing in this regard because they sure show it as if it could but it doesn't.

 

[Vivek]

I think Ricardo is right (again).

This lens focus is quite risky for Olympus I would think, especially for the zooms, even if they are parfocal.

 

[scott kirkpatrick]

Guys, follow my link and look at the video that was done, not by Olympus, but by the fellow at M43 Rumors. The Olympus marketing video is an animation, but it also shows that the sensor can move out of plane when compensating for pitch and yaw. So z-axis motion is possible. Whether it is a good way to get focus shifts is another question.

scott

 

[k-hawinkler]

Guys, follow my link and look at the video that was done, not by Olympus, but by the fellow at M43 Rumors. The Olympus marketing video is an animation, but it also shows that the sensor can move out of plane when compensating for pitch and yaw. So z-axis motion is possible. Whether it is a good way to get focus shifts is another question.

scott

Thanks scott. Which link? TIA.

 

[scott kirkpatrick]

Thanks scott. Which link? TIA.

Mouse over the words "Imaging Resource video" in my first post, and the word "animation" in the second. They are quite nice videos. The BBCode embedding doesn't light up unless you put your cursor on the link title.

scott