View Full Version : Glass Breaking

04 April 2008, 04:56 AM
Pardon the quality,
futzing around with Youtube while I get the p's and q's of my website in order.

I was just wondering if I might be able to get a critique on this? A simple animation of a wineglass glass breaking on a marble floor.

04 April 2008, 07:19 AM
Not that my opinion counts for much, but:

I think the actual shattering, and the fragments are really very good-- the FX portion of the animation is nice. The faux-physics on the main body of the glass, however, could use some tweaking.

It's motion blurred a lot, it appears to be moving very fast before the impact, but it has very little rotation-- so it looks like it's been thrown (or it's extremely finely balanced and fell at a peculiar angle). Is this intentional?

There's also very little bounce-- it hits, and then it tips. It looks like it's being rotated around a pivot at the base of the glass. I'm not sure that reflects the location and angle of impact, the distribution of weight (the latter, of course, changing as the glass shatters). I'm iffy on this-- but something about the impact looks "wrong" to me, though I can't quite place my finger on it even after repeat viewing.

04 April 2008, 04:24 PM
Your right, maybe toning down the motion blur is a good idea.

To be perfectly frank, I didn't really consider *how* the glass has fallen. And perhaps that's sloppy on my part. You think more rotation and slowing it down a bit will give it a sense of being knocked off a counter top accidently?

Hmm, I'll experiment with the impact and see if I can't make it look more natural.

Thank you for the critique.

04 April 2008, 07:24 AM
First up, lose the marble texture from the floor. It makes diagnosing
the animation almost impossible.

The glass strikes the floor with a force that suggests somebody has
thrown it down in anger. So the fragments would be scattered over
at least a radius of 2-3meters, and the speed of the fragments would
be at say a meter a second maybe?

If you can, buy a cheap wine glass and film yourself smashing it in
a safe area, taking all relevent safety precautions. :)

To pull this off, it is a must that a physics fracture would be the
better option, instead of hand keying. Much respect to hand key all
this, but in a production environment it would be physics engine
all the way.

04 April 2008, 01:06 AM
I'll see what I can do to make a less busy background.
And I'll slow down the impact, for sure.

I have to admit, I'm fairly ignorant about Physics Engines.
But I was under the impression they simulated physics in lieu of animation, sort of replacing the animator. So, how does an animator use a physics engine?

04 April 2008, 02:30 AM
The idea is to take care of the fiddly bits of inanimate objects interacting so that the animator can focus on the things that need an aesthetic touch. Don't look at a physics engine as replacing the animator, just another tool he can bring to bear on a project.

04 April 2008, 12:39 AM
Grinnock has hit the nail on the head.

Don't get caught up in that whole physics engines do the animating
debate, it's just that to create a glass shatter like you are trying to do,
one would be better off using a physics engine.

To hand key a decent shatter with a high amount of fragments, all
behaving correctly and not interpenetrating by hand would be murder!

Believe me, I have done the same effect and even using a physics engine
it took quite a while, and was anything but easy.

Read up about it for your 3D app. It will be under "rigid body dynamics"
and prolly also "fracture".

It's not a case of select glass, make shatter. Once you have all your
boring set up done for the simulation, you are free to have wind and
gravity to name just a few, all having influence over the result. So it
could be a very standard type of shatter, or a highly stylized one. The
only limitation is your imagination, so the PE approach is not a purely
mechanical one, devoid of character and emotion.

Try it:)


CGTalk Moderation
04 April 2008, 12:39 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.