I really just wanna say, seriously??? You’ve did like how much work on your little short film project, and you’re asking this???
But I won’t say that… (I typed it instead :p)
Anyhow, create a single bone armature in a blank blender session and turn on axis display in the armature panel. Enter pose mode and insert a keyframe. Advance 10 frames, move the bone along the world’s z-axis, and insert a keyframe, advance 10 frames, move the bone along the world’s y-axis and set a keyframe. Now look at your curves, for the 1st 10 frames the bone’s y curve changed, even though you moved the bone on the world’s z-axis. For the next 10 frames, when you moved the bone on the world’s y-axis, either the bone’s x or z-axis curve changed. So the bone’s curves are for the bone’s axis, not the world’s axis.
Now advance 10 frames, rotate the bone 90 degrees and insert a keyframe, the curves don’t change. advance 10 frames, move the bone back along the world’s y-axis to the same spot as before you moved it on the world’s y-axis, and insert a keyframe. Advance 10 keyframes, move the bone back along the world’s z-axis to it’s starting point and insert a keyframe. Looking at the new keyframes, they are the opposite of the first ones. So rotation has no effect on the bone’s location curves.
Start over with a single bone armature, and rotate the tip of the bone so it’s even with the tail along the world’s x-axis, and repeat the above steps. Looking at the curves, you’ll notice that since the orientation of the bone in edit mode has changed, the curves now refelct those changes.
So the bone’s curves show the changes from the bone’s rest position according to the bone’s axis and not the world’s axis.
Seriously, I wasn’t 100% sure about this, so I did what I just typed, and figured it out. I would assume un-connected child bones location curves would be relative to their parents, but I haven’t tested that. Additionally, rotation values (if using euler rotation) is in radians, not degrees…
hope this helps,