I like that idea - not as a replacement for what is already there, but perhaps as an alternative. I'm not the LogiSphere developer, but I could certainly write the JavaScript to implement your idea. The math involved wouldn't be too hard, but there's a catch...

In order to do the centering thing right, the script would have to know at all times where the camera is pointed and what zoom it's set at (call it "the TPZ coordinates"). Seems easy - just keep track, in the browser where the script is running, of where we last pointed the cam. But remember that LogiSphere is a server that quite happily serves multiple clients at once. "You" could be watching the cam at the same time "I" am moving it.

What it comes down to (or so it seems to me right now) is that every time the browser grabs a new video frame from LogiSphere, it would also have to also get the PTZ coordinates that were current at the instant the frame was grabbed. I'm not seeing a good way to make a script to get those coordinates. Well, I could have it follow every "GET /current.jpg" with a "GET /getpos", and then somehow deal with the text returned by "getpos", but that's nasty and would probably do terrible things to performance.

Maybe SSE has a better idea that I haven't thought of.

I don't mind spending the time coding it if people are interested, a solution turns up for the above problem, and SSE has no objections.

Thanks, SSE! I hadn't considered the "lens factor" nor the differences between the various supported tilt/pan devices. Those things certainly complicate matters!