On Fri, 21 Apr 2017 19:01:16 +0100, Rod Speed wrote:



"James Wilkinson Sword" wrote in message
news
On Fri, 21 Apr 2017 00:56:55 +0100, Rod Speed
wrote:



"James Wilkinson Sword" wrote in message
news On Wed, 19 Apr 2017 23:21:12 +0100, Rod Speed
wrote:

That's a different situation, everything the camera
can see is all moving in the same way all the time.

So the camera could just ignore the centre.

Nope, because if it makes the background stable the center
would move around just as much as the background currently
does

It should make them both move around half what they currently do.

Not possible for the reason I spelt out later and you have now deleted.

I never delete the last five levels, try to make your point quicker. It's very easy to move the whole image electronically so the background moves half of what it currently is, and that movement is passed onto the person's head.

and would be even worse to watch because what you
want to watch is whats in the center.

No you don't. For example in the video we just watched, it's the animal
we're interested in.

Yes, but when that moves around more
than it currently does, that's worse.

Its never going to be feasible to eliminate the problem
with optical stablisation, it has to be physical.

That wouldn't be any different to what you objected to above.

Yes it is, because there would be no background movement with
that approach. You only get background movement when the
camera moves.

With a perfect physical stabiliser, the background would be stationary,

Yes.

and the person would move around.

Not when it tracks the person and lion.

Not possible if the person is moving relative to the lion.

--
Confucius say: "Man who run in front of car get tired"