See it here.

Next I'll work on squash & stretch- for now I want notes on my timing and spacing.

Looks good! My only crit would be that it seems to pause slightly when its at its peak in the air. Otherwise its looks great!

Photoman

It's an honorable start!

I just analyzed the first bounce, but the comments would apply to the other bounces as well.

Click to view attachment
(Frame thru it to read the notes.)

Robcat hit my crits in his notes on the video he posted. The simple, first pass ball has a constant horizontal velocity so it moves the same amount each frame in the x direction. Since the height of the the second bounce is about 2/3s of the the starting position, the vertical velocity right after the bounce should be 2/3s of the the velocity just before the bounce in the opposite direction of course. Each subsequent bounce should be about 2/3s of the previous bounce until it gets quite low when you can just roll it without making it bounce forever.

BTW the technical term for the ratio of drop height to bounce height is "Coefficient of Restitution." Use the tem often and you'll look over-educated. A rubber ball will have a large one like 0.85 or so and a cannonball will have something like .10 - .05 . Flubber has a C of R > 1!

Alan

Very interesting points. While I know the formula of a parabola, it certainly isn't an intuitive shape for me yet!- Is there any easier cheat, like using a rotoscope of a parabola to line your onion skins up on, or is grinding a parabola into your blood part of an animator's job description?

Also thanks Alano, pointing out that the CR gets its reduction from the imapct itself is priceless. While I knew the parabolas would be reducing by a consistent ratio, I still wasn't catching that the next bounce was the moment the ratio changed.

I'm pleased to hear my theory was correct, that the x-axis motion stays constant. I based my effort on this theory, and after creating a "start" and "end" extreme of x-motion, made all of my keyframes exclusively by modifying the y-axis, which leaves my sense of a parabola as the largest fault remaining.


Robcat- the method you used to give me advice is awesome. Looks like you're writing w/ a tablet for the single-frame notes, did you draw in each dot yourself, or do you have some fancy analysis software? In any case, with the dots drawn in, I could see where the arcs of my imperfect parabolas wobbled, much better than the full-shaded "onion skinning" feature of A:M allowed me to.

After you do enough bouncing things you get a sense of what the right motion is and get a sense for the approximate shape of graph that gives you that.

Another option is that it slows down a bit at each impact or that it gradually loses steam (from costant air resistance). It depends on the ball you are trying to represent.



manually drawn



Onion skin can work well too and is potentially more accurate than my hand-drawn dots. Ultimately you have to know what you are looking for either way.