Canon's IS, Nikon's VR, Panasonic's Mega OIS, Pentax's SR - Image stabilization

Are they all different? or is it just a case of Coke vs. Pepsi?
Any advantage or superiority of one over the others?

 

Yes. In fact just for the Case of Nikon VR in slr lenses there are now
three different versions in the field.

 

I think they are all very different, and you cannot say one camera's IS
(VR) = another's.

I believe it's like saying all cars are the same because they have 4
wheels and get you where you are going.

I have not read up on all of them but I know Nikon's VR will distinguish
between many types of intentional and unintentional movement. I would
like to see some comparisons.

 

Indecision is the key to flexibility.

 

Yes.

 

Not quite. in just the slr lenses themselves there would appear to be at
least three different versions. There are at least two different versions
just denoted as VR and now a new VRII has been offered. You can see
differences in the original designation in different lenses by Nikon's
recommendation as whether to use or not use VR on a tripod.

http://support.nikontech.com/cgi-bi...2U9MSZwX3NlYXJjaF90ZXh0PXZy&p_li=&p_topview=1

There were some graphs I recently found, on the .jp site I think, which
also illustrated this, but I can't find the link just now.

 

That page says my lens should have VR off when on a tripod. The manual
that came with the lens says additionally it should be on if the tripod
head is not secured.

 

The point of these technologies is to reduce blurriness from camera
shake. If the camera can't shake, the VR/SR/whatever will compensate for
moving subjects, or subjects which the lens/camera thinks are moving.
You typically don't want it to do that.

 

I thought the VR/IS system responded strictly to camera movement, with
subject movement having no influence.

 

Yes, for most systems, but the electronic image stabilisation used in
movie cameras and some digital cameras (e.g. for video modes) does rely
purely on the image from the sensor, and therefore could be fooled by
subject movement.

David

 

Just what implementation of image stabilization in still cameras
compensates for subject motion? Nikon's does not.

 

Nikon Coolpix 8400 in video mode, for one. It uses what it calls
"electronic VR".

If you make a movie of a group of people walking in the same direction you
can see the VR sometimes moving with the people (so that the people are
steady) and then jumping back (as its stabilisation limit is reached).

Of course, you might call it a movie camera in that mode....

David

 

Turned out these were from the pdf copies of the manuals for my lenses I
recently dl'd from the Nikon USA site. The applicable pages have been
excised here:
http://edwardgruf.com/Digital_Photography/dSLR/pano/index.html

 

That's my point. If the camera's not shaking, then the IS system could
be fooled by moving subjects. That's why manufacturers tell you to turn
off IS if the camera's mounted on a tripod.

 

Then the next time your camera is mounted on a tripod, leave the IS on
and see what happens.

 

Since most of those systems work by using gyroscopes (or some sort of
accelerometer, let's not get hang up on terminology), they'll have a
hard time detecting subject motion, I think.

 

I did and nothing happened. Perhaps all IS systems are not alike.

 

Err, no. Your typical IS system works by having some sort of motion
sensor, an accelerometer or something, measure the motion of the
lens/camera, and move actuators to compensate. They would work just fine
with the lens cap on on the camera. Or the shutter closed.

Turning the IS off when tripod mounting is to prevent the IS system from
going into some sort of runaway feedback loop based on spurious
"vibrations" (from noise in the accelerometers or some such), and
_adding_ vibrations rather than taking them away.

-dms

 

How? The IS in most still cameras uses accelerometers and/or gyros to
sense camera motion. Unless the subject is an apatosaurus shaking the
ground or something there's no way for the subject to "fool" the IS.

More like the IS algorithms are predicated on a particular response curve
(there's always lag in any feedback control system--it may be small but
it's there) and the tripod by changing the system mass changes the response
curve, possibly leading to overcorrection.

 

It depends on the lens. With Canon, the first generation of IS needed to be
turned off when the lens/camera was on a tripod. The second (or third,
depending on how you counted) did not.
I have a 28-135 IS, part of the first gen. and I've seen the, ahem,
interesting results when IS is left on using a tripod.