ragdoll..AGAIN .fla included

[NoSkills]

i've been playing around with the old ragdoll source by namzorg again and modified it so the collision detection isn't height based, now it's based on actual collisions...however... the little guy seems to have epilepsy attacks most of the time and his hat jumps around wildly too

if anyone could take a look at it and tell me how to fix this i'd greatly appreciate it

here's a demo:
http://ize.badne.ws/ragdoll/Ragdollv2mx.swf
mouse to drag the little guy
space + mousedistance = kick him
cursor left + right to change the scene

and here's the fla so you can take a look at it:
http://ize.badne.ws/ragdoll/Ragdollv2mx.fla

[OpethRockr55]

I think I know why, but I can't tell you for sure on my present computer.

If you wait till later today, then I might have an answer for you.

[NoSkills]

thanks alot *waits*

[OpethRockr55]

Whoa! Sorry! Way late on that one, but I think I'm pretty much right. Your method only tests that the background has been hit, not like tile methods which not only tell if it's been hit, but also the direction and the penetration distance of the hit. This can cause some unpleasant things to happen when your ragdoll hits in a weird place that causes the ragdoll to collide again, and thus comes the 'seizures'.

If you can find a way to not only determine a hit, but its penetration depth and correct it accordingly, then you might have better luck.

[Nialsh]

Opeth is right. You drew the stuff the guy collides with in authoring and just used hitTest to make it work. The only real solution I can think of is drawing the stuff with code; then you would have the equations of the lines and you could get something working.

[Santuria]

what about for flash 5?

[OpethRockr55]

What about it?

[Pugger]

Where would you get the original source code?

You know, namzorg's example.

[Santuria]

aye aye!

[swak]

for the depth checking thingy-mabober you can try making layers. Probably better ways. maybe this link will help http://ericlin2.tripod.com/collision/bb5.html . It involves an extreemly precise hitTest.

[Snipergen]

Holy crap! I needed this file and the link is broken!! I REALLY NEED THIS .FLA FILE!!

[NoSkills]

here you go:

http://ize.badne.ws/rag/Ragdollv2.fla
http://ize.badne.ws/rag/Ragdollv2mx.fla

i didn't think anyone would still be interested in this stuff since the collision still sucks and my mathematical skills are suboptimal...so i've given up on this

if you make any progress with it please let me know

[Snipergen]

Thank you so much! The collision indeed sucks but we can do something about that

[NoSkills]

many have tried...many have failed... will you be the one to fix it?
haha

[Pugger]

Thanks too, I've been looking for these. Vectors would be the way to fix it.

[Snipergen]

Nope, it is not even possible to create a perfect collicion hittest function. If you paint a surface the particles cannot find a way to fit on them perfectly.

[Pugger]

I agree. Cant be done using hittest, thats why I'm using simple vectors for collision detection.

I suppose I can post what I have done on my experiments so far. I'm trying to make a quick efficent rag doll. You can pick em up with the mouse. Although the mouse is a bit fast so they sort of stretch, which I can fix through iterations, but that means more CPU time so screw it.

linky: No longer

By the way thanks for posting those two ragdolls above NoSkills helped me HEAPS! The code is so much easier to understand than anything else I've seen.

[ozmic66]

very nice
that's a lot of ragdolls

(if thats what you want to move to flash 9, its possible, but you'd have to draw it by code and load pictures and move them around by code)

[Snipergen]

Very nice! VERY NICE INDEED!

PS: ozmic, how's your physicsengine? Almost done? Can't wait to see more examples in flash 9 of it!

[shedokan]

anyone has a working link to the files?