Here's an interesting one... A brick gets thrown through a window. Does anyone have any ideas on how to calculate the size/shape/velocity of the resulting shards of glass? Thanks, Neil ---------------------- www.3dPool.co.uk ---------------------- [edited by - n e i l on August 16, 2002 10:28:30 AM]

Well, if you were to accuratly calculate it in a game, you''d have a frame rate that''s like a sloth in glue. It doesn''t have to be completly accurate for it to work and look good. Just ask any first person shooter with breaking glass to verify that.

What you should do is for one frame of the breaking glass, just display a crack at the location of the hit and then the next frame on you just randomly cut up the glass wall. That will look fine.

Accuratly modeling it would have to take in the brick''s hardness,speed,angle of hit,angle of brick,air thickness,glass thickness,and more. It would just take too long. Good luck.

---
My Site-My Tetris Clone w/Source
Come join us on IRC in #directxdev @ irc.afternet.org

There was an interesting test done on glass by forensics investigators in a case that is very similar to what you''re asking. Not sure this will help, but it''s good to know anyway. We normally think glass will go in the direction along the axis of the object that broke it, but a bullet at close range was found to make the glass go 180 degrees opposite, ie towards the person. This may or may not be true for other projectiles such as a brick.
As for the size/shape/velocity of the resulting debris, there are a lot of unknowns. Size and shape are (mostly) related to the type of glass. Glass like in most store windows or the back and side windows of cars shatter into extremely small pieces that are safer than big shards, hence the name safety glass. Glass like you see in houses breaks into large pieces and long shards.
The velocity would depend mostly upon the striking angle and velocity of the projectile. Once the glass was broken and set into motion all the normal laws of physics would apply.
I think modeling a realistic breaking of glass would be possible, just require a lot of variables and guesswork. But that''s my opinion and I''m sure someone else knows a better way.

I''d suggest picking a few random points on the glass, then connecting each to all the others that are within a certain radius. This should give you a good pattern for the inner part of the glass. Maybe for the rest consider checking if any point is connected to only one other, if so connect it to a point along the edge.

heres a good link.. but its not realtime..

http://www.cs.berkeley.edu/~job/Papers/obrien-2002-GMA.pdf

quote:Original post by Dmzline
There was an interesting test done on glass by forensics investigators in a case that is very similar to what you''re asking. Not sure this will help, but it''s good to know anyway. We normally think glass will go in the direction along the axis of the object that broke it, but a bullet at close range was found to make the glass go 180 degrees opposite, ie towards the person.

This has a lot to do with the sort of glass being shot at and because, in general, normal glass is elastic. The bullet doesn''t actually shatter the glass... the glass does! As the bullet enters the glass it punches a neat hole at the point of impact. Friction against the surface of the bullet then drags the glass with the bullet as it expands the hole. If you watch slow motion video of it you can see the glass extended into a sort of sharp pointed bowl in the direction of motion. Once the bullet passes through the glass friction ceases and so the only forces acting on the glass are internal tensions in the crystaline structure. These force the glass to accelerate in the direction opposite to motion of the bullet. The glass was slightly stretched when the bullet dragged it out of shape and now it compresses again. Glass doesn''t like being compressed and so it shatters in the compression region as it tries to restore its shape. Since the glass has some momentum the shards continue to move back toward the firer. This is why, when you look at the hole made by a bullet through a thick piece of glass, the bigger section of glass missing is on the side of the firer and there is usually a ''neat'' hole on the side where the bullet came out.

If the glass is too thin then it shatters as it is dragged out of shape by the bullet since its internal tensions are insufficient (it is too brittle).

This reminds me of the tests they do to determine blood splash. They get a fresh pig corpse and fire live rounds into it, then measure where all of the blood went to figure out the trajectory properties of human blood (since pigs and humans have very similar blood properties)!

My stepfather is the senior specialist for the crime scene investigation unit for the forensic division of our state police. Needless to say I grew up with interesting dinner conversation!

Cheers,

Timkin

When I was talking about the glass coming back, I was referring to normal household glass (yes, there are lots of different kinds), but I should probably have explained that. And thanks for the explanation as to why that happens, Timkin. Had always thought it was something to do with the pressure defferential created by the gunshot itself.

I have heard that in very brittle forms of glass, the pressure wave in front of the bullet is enough to compress the glass and cause it to fracture. One would imagine that in such situations, the shockwave from the impact would also cause large fractures through the glass (since it is felt as an internal compression wave)... but I''m only speculating on this being possible.

Cheers,

Timkin