After Effects 7 Graph Editor - Separate Position Axis Velocity?

Discussion in 'Amateur Video Production' started by gmccreight, Apr 27, 2006.

1. gmccreightGuest

Perhaps I'm spoiled from using Blender (which is really really
amazing), but it seems that you ought to be able to have separate X, Y,
(and what the heck, Z) position velocity information in the Graph
Editor.

For example, let's say you want to simulate the throwing of a ball off
a tall building. X velocity decreases over time and Y velocity
increases. This would be extremely easy to do with two seperate
"channels" of velocity information, but a real pain otherwise.

I mean, do I really need to bust out expressions code for this, or nest
compositions?

-Gordon

gmccreight, Apr 27, 2006

2. gmccreightGuest

You're crazy. Oh wait. I'm crazy. Turns out, you can do this by
using a null object. Put one of the axes of position change in the
null object (say the x axis change). Now use that as a parent, and put
the y axis change in the child, which will now have both the x and y
change because of inheriting the x change from the parent.

This seems a little bit odd to me, but it's easy and it works.

--Gordon McCreight

gmccreight, May 1, 2006

3. Gene E. BlochGuest

But to be physically accurate, the x-axis speed should not change (or
not much, anyway).

You should end up with a parabola expressed as x = v*t and y =
0.5*g*t**2, where v is the original x velocity and g is the
acceleration of gravity, 32 feet (or 9.8 meters) per second squared.

If you don't like my choice of the y-coordinate (and who could blame
you!), we could rewrite the second equation as y = h - 0.5*g*t**2,
where h is the starting height.

Exception: if the object is not very dense (like a chunk of
light-weight plastic foam or a sky diver with an open parachute), both
formulas break down, because of air resistance. Actually, they always
break down for that reason, but it's only noticeable for lower density
objects or for long fall times.

Totally OT: The day of my first sky diving, I watched a sky diver fall
most of the way to the ground with an unopened chute. She got her
auxiliary chute open at the last minute and had a rough landing, but
managed not to be injured. Another diver that day bounced off some high
voltage power lines near the airport, also without injury. I went home
without jumping and never went back.

Gino

Gene E. Bloch, May 1, 2006
4. gmccreightGuest

But to be physically accurate, the x-axis speed should not change (or
not much, anyway).

As you say, it all depends on what the initial speed of the object is,
whether you're factoring in drag, etc. In my case I just wanted to
make something which looked semi-real. I can see how using Null
Objects and the expressions you provided could make a nice powerful
combination.

Cheers - G

gmccreight, May 2, 2006
5. Gene E. BlochGuest

At the very least, it looks to me like you're having a lot of fun!

Of course, in your extensive experimentation, you might discover the
perfect fall trajectory in perceptual terms, which might not relate
that closely to the physical perfection I'm obsessing over

I was thinking that if your editing software doesn't automate this
process enough, you could make up a spreadsheet using the appropriate
formulas. Each row in the spread sheet would end up with a frame
number, an x position, and a y position. That might help a lot if you
end up editing the image manually. All you need to do is figure out the
most convenient units to use, such as pixels or inches or mm.

Gino

Gene E. Bloch, May 2, 2006