Stop Camera From Going Behind On Sit

Okay. Here's the deal on me - I'm pretty good at scripting but I suck at vector and rotation math. And this whole thing is killing me.

In essence, I have an object where when I sit on it, it spins the camera around behind me and if my object is up against a wall - I'm looking at a wall - not myself sitting.

So, okay, I can offset the camera so it's in front of the object and looking at me from the front. Which is fine and dandy until I decide to rotate the object and put it against a different wall. If it's the "opposite" wall, I'm back to where I started. And if it's a side wall, I'm looking at it from the back.


So... MY QUESTION AT LAST:

Can anyone help me with either:

a) some math so that no matter what direction my object is facing, the camera will be looking at me from the front of it?

or

b) Just stop the camera from moving altogether because before I sit, it's generally looking at the object I'm sitting on in the first place.

(I realize that if I ALT_CLICK on an object before I sit on it, my camera won't move, but I really want it to just work the way it should in the first place).

Any help, any clues, and links to examples of things that do this would be greatly appreciated. Thanks so much!

Rygel

this is the method I use in my poseballs.

I'm sure it can be improved, but it works for me.

set_camera()
{
    vector cpos;
    if (camera_pos_relative)
    {
        cpos = llGetPos() + camera_pos*llGetRot();
    }
    else
    {
        cpos = camera_pos;
    }

//        CAMERA_DISTANCE, 2.5, // ( 0.5 to 10) meters
    list CAMERA_PARAMS = [
        CAMERA_ACTIVE, TRUE, // TRUE is active, FALSE is inactive
        CAMERA_BEHINDNESS_ANGLE, 0.0, // (0 to 180) degrees
        CAMERA_BEHINDNESS_LAG, 0.0, // (0 to 3) seconds
        CAMERA_FOCUS_LAG, 0.1 , // (0 to 3) seconds
        CAMERA_FOCUS_THRESHOLD, 0.0, // (0 to 4) meters
        CAMERA_POSITION, cpos, // region relative position
        CAMERA_POSITION_LOCKED, TRUE, // (TRUE or FALSE)
        CAMERA_POSITION_LAG, 0.1, // (0 to 3) seconds
        CAMERA_POSITION_THRESHOLD, 0.0, // (0 to 4) meters
        CAMERA_FOCUS_OFFSET, ZERO_VECTOR // <-10,-10,-10> to <10,10,10> meters
    ];

    if (camera_focus_enable)
    {
        vector cfocus;
        if (camera_focus_relative)
        {
            cfocus = llGetPos() + camera_focus*llGetRot();
        }
        else
        {
            cfocus = camera_focus;
        }
        CAMERA_PARAMS = CAMERA_PARAMS + [
            CAMERA_FOCUS, cfocus, // region relative position
            CAMERA_FOCUS_LOCKED, TRUE // (TRUE or FALSE) grr basically must be true, or camera will just pan back after moving
        ];
    }

    llSetCameraParams(CAMERA_PARAMS);
}

regards,

Core

In the above script, how are these variables defined / set:

camera_pos_relative
camera_pos
camera_focus_enable
camera_focus_relative

Don't llSetCameraAtOffset() and llSetCameraEyeOffset() do what's desired? They're specified relative to the scripted object (like llSitTarget), so if the object rotates, they just rotate along too. Or am I missing something?

You're not missing anything. They'd work just dandy.

I perfer llSetCameraParams over it only because you can change the camera position and focus later without having the user stand and re-sit their avatar. It's a bit more work but offers a lot more options.

For instance: I once made a chair that focused on the face of each avatar that spoke. I never put it out there because it ended up being far more annoying than useful. =^_^=