|
Dpreview does away with formal resolution tests in their reviews
|
Re: Dpreview does away with formal resolution tests in their reviews
On Tue, 01 May 2012 15:51:17 -0400, Alan Browne
<alan.browne@FreelunchVideotron.ca> wrote: >On 2012-04-30 22:55 , Eric Stevens wrote: >> On Mon, 30 Apr 2012 19:13:56 -0400, Alan Browne >> <alan.browne@FreelunchVideotron.ca> wrote: >> >>> On 2012-04-30 11:58 , RichA wrote: >>>> http://www.dpreview.com/reviews/olympusem5/ >>> >>> Very nice stabilization system (pitch, roll, yaw, y, z). >> >> I suspect only the X and Y rotational axiis matter much. > >Translation (y,z) are a lot of it (same as the Sony system). > >The weakness of that is pitch and roll is also present. Yaw is a lesser >concern. When you depress the shutter you can easily roll the camera a >little. Along the lens axis, if not well balanced, is the other concern >(pitch). A less ambiguous way to describe pitch would be to say it is rotation about the axis running horizontally across the sensor. Or, more simply, tilting up and down. |
Re: Dpreview does away with formal resolution tests in their reviews
On Tue, 01 May 2012 16:49:21 -0400, Alan Browne
<alan.browne@FreelunchVideotron.ca> wrote: >On 2012-05-01 16:41 , John A. wrote: >> On Tue, 01 May 2012 15:51:17 -0400, Alan Browne >> <alan.browne@FreelunchVideotron.ca> wrote: >> >>> On 2012-04-30 22:55 , Eric Stevens wrote: >>>> On Mon, 30 Apr 2012 19:13:56 -0400, Alan Browne >>>> <alan.browne@FreelunchVideotron.ca> wrote: >>>> >>>>> On 2012-04-30 11:58 , RichA wrote: >>>>>> http://www.dpreview.com/reviews/olympusem5/ >>>>> >>>>> Very nice stabilization system (pitch, roll, yaw, y, z). >>>> >>>> I suspect only the X and Y rotational axiis matter much. >>> >>> Translation (y,z) are a lot of it (same as the Sony system). >>> >>> The weakness of that is pitch and roll is also present. Yaw is a lesser >>> concern. When you depress the shutter you can easily roll the camera a >>> little. Along the lens axis, if not well balanced, is the other concern >>> (pitch). >> >> A less ambiguous way to describe pitch would be to say it is rotation >> about the axis running horizontally across the sensor. Or, more >> simply, tilting up and down. > >As you like. Point being if "along the lens axis" describes pitch, it also describes yaw (swinging side to side). Or rotation about any arbitrary axis running parallel to the sensor plane and intersecting the lens axis, for that matter. |
Re: Dpreview does away with formal resolution tests in their reviews
On Wed, 02 May 2012 13:29:03 +1200, Eric Stevens
<eric.stevens@sum.co.nz> wrote: >On Tue, 01 May 2012 15:51:17 -0400, Alan Browne ><alan.browne@FreelunchVideotron.ca> wrote: > >>On 2012-04-30 22:55 , Eric Stevens wrote: >>> On Mon, 30 Apr 2012 19:13:56 -0400, Alan Browne >>> <alan.browne@FreelunchVideotron.ca> wrote: >>> >>>> On 2012-04-30 11:58 , RichA wrote: >>>>> http://www.dpreview.com/reviews/olympusem5/ >>>> >>>> Very nice stabilization system (pitch, roll, yaw, y, z). >>> >>> I suspect only the X and Y rotational axiis matter much. >> >>Translation (y,z) are a lot of it (same as the Sony system). >> >>The weakness of that is pitch and roll is also present. Yaw is a lesser >>concern. When you depress the shutter you can easily roll the camera a >>little. Along the lens axis, if not well balanced, is the other concern >>(pitch). > >Pitch movement of the lens axis is the same as roll around the X axis. > >My point is that roll only matters if it deflects (as distinct from >rotates) the lens axis. The lever effect of a long lens axis (to the >subject) makes such movements critical. So if the camera is spinning about the lens axis so as to turn, say, ten degrees during the exposure, that will make no difference to the shot so long as said axis itself doesn't move? :) |
Re: Dpreview does away with formal resolution tests in their reviews
On May 2, 5:52*am, Eric Stevens <eric.stev...@sum.co.nz> wrote:
> On Wed, 02 May 2012 01:02:35 -0400, John A. <j...@nowhere.invalid> > wrote: > > > > > > > > > > >On Wed, 02 May 2012 13:29:03 +1200, Eric Stevens > ><eric.stev...@sum.co.nz> wrote: > > >>On Tue, 01 May 2012 15:51:17 -0400, Alan Browne > >><alan.bro...@FreelunchVideotron.ca> wrote: > > >>>On 2012-04-30 22:55 , Eric Stevens wrote: > >>>> On Mon, 30 Apr 2012 19:13:56 -0400, Alan Browne > >>>> <alan.bro...@FreelunchVideotron.ca> wrote: > > >>>>> On 2012-04-30 11:58 , RichA wrote: > >>>>>>http://www.dpreview.com/reviews/olympusem5/ > > >>>>> Very nice stabilization system (pitch, roll, yaw, y, z). > > >>>> I suspect only the X and Y rotational axiis matter *much. > > >>>Translation (y,z) are a lot of it (same as the Sony system). > > >>>The weakness of that is pitch and roll is also present. *Yaw is a lesser > >>>concern. *When you depress the shutter you can easily roll the camera a > >>>little. *Along the lens axis, if not well balanced, is the other concern > >>>(pitch). > > >>Pitch movement of the lens axis is the same as roll around the X axis. > > >>My point is that roll only matters if it deflects (as distinct from > >>rotates) the lens axis. The lever effect of a long lens axis (to the > >>subject) makes such movements critical. > > >So if the camera is spinning about the lens axis so as to turn, say, > >ten degrees during the exposure, that will make no difference to the > >shot so long as said axis itself doesn't move? :) > > I was considering only the movements likely to be encountered in > normal use. I'm not quite sure how you could obtain 10 degrees > rotation during any exposure sufficiently short for any stabilisation > system to have any effect. In fact, I don't think any stabilisation > system could cope with 10 degrees of rotation. > > Regards, > > Eric Stevens It can't. You'd have to have something like they use in the fire control system of tanks to deal with that. I don't think it would fit in the camera. :) |
Re: Dpreview does away with formal resolution tests in their reviews
On Wed, 02 May 2012 21:52:47 +1200, Eric Stevens
<eric.stevens@sum.co.nz> wrote: >On Wed, 02 May 2012 01:02:35 -0400, John A. <john@nowhere.invalid> >wrote: > >>On Wed, 02 May 2012 13:29:03 +1200, Eric Stevens >><eric.stevens@sum.co.nz> wrote: >> >>>On Tue, 01 May 2012 15:51:17 -0400, Alan Browne >>><alan.browne@FreelunchVideotron.ca> wrote: >>> >>>>On 2012-04-30 22:55 , Eric Stevens wrote: >>>>> On Mon, 30 Apr 2012 19:13:56 -0400, Alan Browne >>>>> <alan.browne@FreelunchVideotron.ca> wrote: >>>>> >>>>>> On 2012-04-30 11:58 , RichA wrote: >>>>>>> http://www.dpreview.com/reviews/olympusem5/ >>>>>> >>>>>> Very nice stabilization system (pitch, roll, yaw, y, z). >>>>> >>>>> I suspect only the X and Y rotational axiis matter much. >>>> >>>>Translation (y,z) are a lot of it (same as the Sony system). >>>> >>>>The weakness of that is pitch and roll is also present. Yaw is a lesser >>>>concern. When you depress the shutter you can easily roll the camera a >>>>little. Along the lens axis, if not well balanced, is the other concern >>>>(pitch). >>> >>>Pitch movement of the lens axis is the same as roll around the X axis. >>> >>>My point is that roll only matters if it deflects (as distinct from >>>rotates) the lens axis. The lever effect of a long lens axis (to the >>>subject) makes such movements critical. >> >>So if the camera is spinning about the lens axis so as to turn, say, >>ten degrees during the exposure, that will make no difference to the >>shot so long as said axis itself doesn't move? :) > >I was considering only the movements likely to be encountered in >normal use. I'm not quite sure how you could obtain 10 degrees >rotation during any exposure sufficiently short for any stabilisation >system to have any effect. In fact, I don't think any stabilisation >system could cope with 10 degrees of rotation. 1/10 degree, then. That will give you about a pixel-width of rotational motion blur per ~573 pixels (1800/pi) out from the center of rotation. (So at about 1150 pixels from the center you get about 2 pixels of motion blur, and so on.) Hmmm... Presumably lateral (x & y) stabilization would take care some portion of that if the center of rotation is way off-center, say at your tripod mount or monopod foot. So I imagine that would that reduce the task of rotational stabilization to being centered mid-sensor. |
Re: Dpreview does away with formal resolution tests in their reviews
On Thu, 03 May 2012 12:32:40 +1200, Eric Stevens
<eric.stevens@sum.co.nz> wrote: >On Wed, 02 May 2012 17:43:29 -0400, Alan Browne ><alan.browne@FreelunchVideotron.ca> wrote: > >>On 2012-05-02 05:52 , Eric Stevens wrote: >> >>> I was considering only the movements likely to be encountered in >>> normal use. I'm not quite sure how you could obtain 10 degrees >>> rotation during any exposure sufficiently short for any stabilisation >>> system to have any effect. In fact, I don't think any stabilisation >>> system could cope with 10 degrees of rotation. >> >>In the limited space inside a camera, the compensation is probably no >>more than a degree (maybe more, probably less) in any axis. >> >>There are commercial and military stabilization systems for various >>cameras and sensors that cope with much more. They are larger and often >>extremely expensive. > >And how do you compensate for rotation about the axis of the lens? >Rotate the sensor (with all it's connections)? Yup. IIRC that's how Pentax did it in the K5 (and K7?) It can also compensate for slightly tilted horizons. |
Re: Dpreview does away with formal resolution tests in their reviews
On Fri, 04 May 2012 08:49:23 +1200, Eric Stevens
<eric.stevens@sum.co.nz> wrote: >On Thu, 03 May 2012 05:49:02 -0500, "C. Neil Ellwood" ><cral.elllwood2@btinternet.com> wrote: > >>On Wed, 02 May 2012 06:54:35 -0700, RichA wrote: >> >>> On May 2, 5:52*am, Eric Stevens <eric.stev...@sum.co.nz> wrote: >>>> On Wed, 02 May 2012 01:02:35 -0400, John A. <j...@nowhere.invalid> >>>> wrote: >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >On Wed, 02 May 2012 13:29:03 +1200, Eric Stevens >>>> ><eric.stev...@sum.co.nz> wrote: >>>> >>>> >>On Tue, 01 May 2012 15:51:17 -0400, Alan Browne >>>> >><alan.bro...@FreelunchVideotron.ca> wrote: >>>> >>>> >>>On 2012-04-30 22:55 , Eric Stevens wrote: >>>> >>>> On Mon, 30 Apr 2012 19:13:56 -0400, Alan Browne >>>> >>>> <alan.bro...@FreelunchVideotron.ca> wrote: >>>> >>>> >>>>> On 2012-04-30 11:58 , RichA wrote: >>>> >>>>>>http://www.dpreview.com/reviews/olympusem5/ >>>> >>>> >>>>> Very nice stabilization system (pitch, roll, yaw, y, z). >>>> >>>> >>>> I suspect only the X and Y rotational axiis matter *much. >>>> >>>> >>>Translation (y,z) are a lot of it (same as the Sony system). >>>> >>>> >>>The weakness of that is pitch and roll is also present. *Yaw is a >>>> >>>lesser concern. *When you depress the shutter you can easily roll >>>> >>>the camera a little. *Along the lens axis, if not well balanced, is >>>> >>>the other concern (pitch). >>>> >>>> >>Pitch movement of the lens axis is the same as roll around the X >>>> >>axis. >>>> >>>> >>My point is that roll only matters if it deflects (as distinct from >>>> >>rotates) the lens axis. The lever effect of a long lens axis (to the >>>> >>subject) makes such movements critical. >>>> >>>> >So if the camera is spinning about the lens axis so as to turn, say, >>>> >ten degrees during the exposure, that will make no difference to the >>>> >shot so long as said axis itself doesn't move? :) >>>> >>>> I was considering only the movements likely to be encountered in normal >>>> use. I'm not quite sure how you could obtain 10 degrees rotation during >>>> any exposure sufficiently short for any stabilisation system to have >>>> any effect. In fact, I don't think any stabilisation system could cope >>>> with 10 degrees of rotation. >>>> >>>> Regards, >>>> >>>> Eric Stevens >>> >>> It can't. You'd have to have something like they use in the fire >>> control system of tanks to deal with that. I don't think it would fit >>> in the camera. :) >> >>The gun (main armament) in a tank is a little larger and has to cope with >>much greater movement than that which occurs in a camera. > >And they don't worry about rotation about the axis of the barrel. IIRC they actively (or passively?) encourage it in the projectile. |
Re: Dpreview does away with formal resolution tests in their reviews
On Fri, 04 May 2012 17:15:26 -0400, Alan Browne
<alan.browne@FreelunchVideotron.ca> wrote: >On 2012-05-03 19:13 , Eric Stevens wrote: >> On Thu, 03 May 2012 17:05:40 -0400, Alan Browne > >>> In the Oly case, the amount of rotation is small. The entire sensor >>> package is mounted on the articulation system. It's the articulation >>> system that moves. The connection from that to the camera electronics >>> would most likely be flexible-printed circuit. >> >> But these don't normally bend sideways. I wonder how they do it? > >"bend"? Nothing about bending, it's just degrees of articulation. > >For example a z-y translation table mounted on a x-axis rotation table >mounted on a z, y axis tilt table. Probably driven by piezo actuators >which can make tiny, precise, repeatable movements. The magic is in >making it all very thin such that it fits and the two tilt axis' result >in the center of the sensor not changing the focal length when activated. I wonder how much range the tilt has and if the firmware can be hacked to set it manually. |
Re: Dpreview does away with formal resolution tests in their reviews
> "Alan Browne" wrote in message
> news:KOSdndYXeKMooz3SnZ2dnUVZ_vEAAAAA@giganews.com ... > Translation (y,z) are a lot of it (same as the Sony system). > The weakness of that is pitch and roll is also present. Yaw is a lesser > concern. When you depress the shutter you can easily roll the camera a > little. Along the lens axis, if not well balanced, is the other concern > (pitch). Translation of the sensor (as the Sony system) is used to counteract the yaw and pitch (rotation) of the camera. To my knowledge it is not used to counteract the translation of the camera itself. That would be usefull for closeup situations, but not for general usage. Ben |
| All times are GMT. The time now is 07:04 PM. |
Powered by vBulletin®. Copyright ©2000 - 2013, vBulletin Solutions, Inc.
SEO by vBSEO ©2010, Crawlability, Inc.