#Dynamic Motion Module
6 messages in this thread
Dave,
While in some instances it would be easy to calculate roughly how much time was
needed, it becomes much more complex when an object may encounter many others
on its way down. Also, for things that might be traveling across a ground
object, and maybe even heading uphill, the calculations become much too
complex.
Calculations in DMM are frame interdependent meaning that until we calculate a
given frame, we cannot calculate the following frame. DMM calculates motion at
1/30th of a second, although we added a sampling routine to calculate motion at
finer increments. Sometimes the distance travelled between two frames may allow
an object to penetrate another object. To allow for this, the user can select a
resolution of up to 1/480th of a second. This is not necessary for most
animations and does increase the scene generation time.
Yes, the first sample I uploaded to Compuserve is fairly simplistic. Within a
few days, I should be able to upload some complex object interactions. If you
compare the sample versus what the bounce kxp does if you rotate the slab on
its side, you should get a general idea of the differences between a bounding
box detection test versus a complex detection test like DMM performs.
Thanks for your interest and comments. DMM will only get better by users
requesting features they'd like to see.
Brent
Hello, Brent,
That is good news. Does that mean that you will be one of the special
few who will keep upgrading their IPAS programs?
What about horizontal impact and rebound, like two cars running into
each other. Is this motion covered and controllable? Also, can a motion be
calculated in reverse, so that an object can suddenly explode wwithout a slow
in motion. I'm not sure how you could do it so it explodes and then slams into
other things, gets more spin and careens away. You know—action stuff!
Regards, Sanford
> Does that mean that you will be one of the special few who will keep
upgrading their IPAS programs?
Yes. We have a long list of future features and enhancements. We have one
upgrade planned for January 1995 already.
>What about horizontal impact and rebound, like two cars running into each
other. Is this motion covered and controllable?
You can use DMM for simulating those types of motions, but there are some
missing variables. Because the intitial version doesn't include friction, the
resulting response will not be as accurate as it could be. Also, DMM does not
perform the object deformation, so even though the collision response would
look very close to what actually would happen with the set speeds, the vehicles
would look the same after the collision as they did before.
After the motion files were generated, you could probably set up an object
morph in 3DS to do the crunching at the right time.
Accident reconstruction is one area that DMM will continually improve as new
versions become available. There are so many factors though, that it is a very
difficult procedure. Consider the make up of the pavement, the weather
conditions (rain, snow), sand on the road, skidding and so on. DMM is now best
used as a replacement to the chalkboard or storyboard. It is a reasonably close
approximation of the events
>Also, can a motion be calculated in reverse, so that an object can suddenly
explode without a slow in motion. I'm not sure how you could do it so it
explodes and then slams into other things, gets more spin and careens away. You
know—action stuff!
I'll have to check on the motion in reverse thing. With our LightWave version,
we included a macro that would chop an object into a user defineable number of
pieces, save them as individual objects, and place them so that they were all
pasted back together. Then in DMM, you could smack the objects from behind with
a non-rendered sphere. Not only do the pieces go flying, they collide with
other pieces as a result. We have discussed adding a similar routine to the
IPAS DMM, but as of yet haven't done so.
Brent
I am looking forward to mashing, smashing, bouncing, slamming all kinds of
stuff! Got anybody to write a review for 3D Artist? I am already writing an
article on instance objects and character substitution in Animation Master. If
you read that mag, check out my article in issue 16 on the roadrunner …I did
the cover image. I know you are a writer, but you can't very well praise
yourself, can you? (g).
Sell lots and prosper,
Regards, Sanford.
No one has officially contacted me about the 3D Artist review, but I have
spoken many times to Wallace Jackson. If you want to line it up, you can
contact me.
You think people would notice if I wrote a bunch of reviews for it myself?
Yeah, you're probably right.
I think 3D Artist is a great magazine. It seems to be a good resource for 3D
Studio users.
Brent,
If that review is already being written, I understand. Perhaps
because my specialty is animals and their behavior, I could approach
motion from a different viewpoint. Just a thought.
I would enjoy reviewing it from that viewpoint, if you are up
for it. People are animals too, in my book. Interaction of "soft"
objects is different.