Dpreview decries the high cost of in-lens I.S.

In article
,
ransley wrote:

Why dont they put IS in the body like Sony and Panasonic. You could
then have double iS if your lens has IS.

no you couldn't. it doesn't work that way.

With the new EF 16-35 2.8
there is no IS offered on a 1500$ lens, kinda dumb in this day of
competition

stabilization is not as important at short focal lengths and it would
make the lens much bigger too.

On Sat, 14 Nov 2009 01:49:37 -0500, John A. wrote:

On Thu, 12 Nov 2009 11:51:01 -0800, nospam
wrote:

In article
,
ransley wrote:

Why dont they put IS in the body like Sony and Panasonic. You could
then have double iS if your lens has IS.

no you couldn't. it doesn't work that way.

Can in-lens IS twist light to compensate for rotational motion about
the through-the-lens axis? Even if there's a way, I'd imagine it would
be a rather exotic & expensive proposition, and it would be easier to
just counterrotate the sensor anyway. You wouldn't even have to take
the lens into account unless maybe you're using a tilt-shift (and I
imagine you're pretty much screwed as far as IS with those anyway.)

Even without rotation, it would be nice to have in-body IS that
automatically disabled itself when a lens with it was mounted.

Look up "dove prism". They've been used for altering longitudinal image
rotation for about 150 years.

On Sat, 14 Nov 2009 17:25:56 -0500, John A. wrote:

On Sat, 14 Nov 2009 10:15:34 -0600, Mr. Info
wrote:

On Sat, 14 Nov 2009 01:49:37 -0500, John A. wrote:

On Thu, 12 Nov 2009 11:51:01 -0800, nospam
wrote:

In article
,
ransley wrote:

Why dont they put IS in the body like Sony and Panasonic. You could
then have double iS if your lens has IS.

no you couldn't. it doesn't work that way.

Can in-lens IS twist light to compensate for rotational motion about
the through-the-lens axis? Even if there's a way, I'd imagine it would
be a rather exotic & expensive proposition, and it would be easier to
just counterrotate the sensor anyway. You wouldn't even have to take
the lens into account unless maybe you're using a tilt-shift (and I
imagine you're pretty much screwed as far as IS with those anyway.)

Even without rotation, it would be nice to have in-body IS that
automatically disabled itself when a lens with it was mounted.

Look up "dove prism". They've been used for altering longitudinal image
rotation for about 150 years.

Interesting. I imagine one of those in a lens would make it honking
big, though. Not to mention the motor it would need to move it quickly
and accurately. Still best to do it with the sensor.

Yes, but thought you'd be interested in the optical alternative.

With a folded light-path, putting the dove-prism alongside the main optical
axis, it wouldn't have to be all that large. If the light-path of the image
was reduced to a centimeter in circumference (or smaller), then expanded
again to the rest of the optical train, then it could be lying alongside an
internal focal-point within a normal sized lens barrel without changing the
diameter at all.

The upside of doing it optically, since all cameras are rapidly joining the
video-capable team, you could do some special video effects that wouldn't
be possible by shifting the sensor. As well as making portrait and
landscape modes a convenient button-press instead of tilting the whole
camera into a more awkward position. Instant auto-leveling no matter the
camera's orientation a possibility too.

Though I'm not sure how the polarizing effects of a dove-prism would figure
into things. If that makes it a non-contender. There might be other
methods, by finishing the reflecting (not refracting) face(s) into fine-cut
steps or otherwise changing the light-path within a similar type of prism,
so the axial reflection is not coming off of the reflecting-face at such an
obtuse angle as to impart polarization.

E.g. the reflecting face being cut into a /\/\/\/\/\ surface (though less
acute). This would redirect the light path to the opposite long face, (now
silvered), and back again to the reflecting surface, sending it on to the
exiting refracting face without polarization. No obtuse reflections to
impart polarization. Cutting an optical surface of that nature as finely
and accurately as might be needed could have been outside of the means
available in H. W. Dove's day, and likely, there was no need for a
non-polarizing version that long ago so it was never explored. If that
surface is cut into a master blank and using that for manufacturing acrylic
prisms it would make it very cost effective as well as light-weight for
allowing for more rapid and accurate orientation changes. Not to mention
less CA problems from the refractions involved.

We shall call this non-polarizing, non-CA acrylic, dove-prism the "Mr. Info
Prism". :-)

(fixed-pitch font for viewing)

_____(silvered)_______
/ \
/ \
^^^^^^^^^^^^^^^^^^^^^^^^^^

On Sat, 14 Nov 2009 17:25:56 -0500, John A. wrote:

On Sat, 14 Nov 2009 10:15:34 -0600, Mr. Info
wrote:

On Sat, 14 Nov 2009 01:49:37 -0500, John A. wrote:

On Thu, 12 Nov 2009 11:51:01 -0800, nospam
wrote:

In article
,
ransley wrote:

Why dont they put IS in the body like Sony and Panasonic. You could
then have double iS if your lens has IS.

no you couldn't. it doesn't work that way.

Can in-lens IS twist light to compensate for rotational motion about
the through-the-lens axis? Even if there's a way, I'd imagine it would
be a rather exotic & expensive proposition, and it would be easier to
just counterrotate the sensor anyway. You wouldn't even have to take
the lens into account unless maybe you're using a tilt-shift (and I
imagine you're pretty much screwed as far as IS with those anyway.)

Even without rotation, it would be nice to have in-body IS that
automatically disabled itself when a lens with it was mounted.

Look up "dove prism". They've been used for altering longitudinal image
rotation for about 150 years.

Interesting. I imagine one of those in a lens would make it honking
big, though. Not to mention the motor it would need to move it quickly
and accurately. Still best to do it with the sensor.

Yes, but thought you'd be interested in the optical alternative.

With a folded light-path, putting the dove-prism alongside the main optical
axis, it wouldn't have to be all that large. If the light-path of the image
was reduced to a centimeter in circumference (or smaller), then expanded
again to the rest of the optical train, then it could be lying alongside an
internal focal-point within a normal sized lens barrel without changing the
diameter at all.

The upside of doing it optically, since all cameras are rapidly joining the
video-capable team, you could do some special video effects that wouldn't
be possible by shifting the sensor. As well as making portrait and
landscape modes a convenient button-press instead of tilting the whole
camera into a more awkward position. Instant auto-leveling no matter the
camera's orientation a possibility too.

Though I'm not sure how the polarizing effects of a dove-prism would figure
into things. If that makes it a non-contender. There might be other
methods, by finishing the reflecting (not refracting) face(s) into fine-cut
steps or otherwise changing the light-path within a similar type of prism,
so the axial reflection is not coming off of the reflecting-face at such an
obtuse angle as to impart polarization.

E.g. the reflecting face being cut into a /\/\/\/\/\ surface (though less
acute). This would redirect the light path to the opposite long face, (now
silvered), and back again to the reflecting surface, sending it on to the
exiting refracting face without polarization. No obtuse reflections to
impart polarization. Cutting an optical surface of that nature as finely
and accurately as might be needed could have been outside of the means
available in H. W. Dove's day, and likely, there was no need for a
non-polarizing version that long ago so it was never explored. If that
surface is cut into a master blank and using that for manufacturing acrylic
prisms it would make it very cost effective as well as light-weight for
allowing for more rapid and accurate orientation changes. Not to mention
less CA problems from the refractions involved.

We shall call this non-polarizing, non-CA acrylic, dove-prism the "Mr. Info
Prism". :-)

(fixed-pitch font for viewing)

Correction. The fine-cut parallel-saw-tooth surface would also have to be
silvered.


_____(silvered)______
/ \
/ \
^^^^^^^^^^^^^^^^^^^^^^^^^
(silvered)