# Questions on camera matrices

I am getting very confused on camera matrices.

1) Does the extrinsic matrix describe the transformation from object to camera or camera to object ?

2) Does the extrinsic matrix apply the rotation before the translation or vice versa ? In either case, does it transform the points with respect to the original origin ? Like translate(p with respect to origin) then rotate(p with respect to origin) or is it like translate(p with respect to origin) then rotate(p with respect to new origin after the translation) ?

3) I read that the extrinsic matrix transforms points in world coordinates to camera coordinates. When do I use this ? Aren't my camera always the origin ?

4) Does the extrinsic matrix bring my camera to the world origin or does it bring the world origin to my camera ?

5) Does anyone know of any link or any simple example with images that show the above numerically ? I think being able to watch a correct simple example with numerical would greatly help my understanding. I am getting extremely confused by this extrinsic matrix. What happens before and after applying the matrix and where my objects and cameras are before and after.