Rune wrote:
> That's impressive. :)

Thank you.

>
> I found that the hardest part in making this kind of automated walking,
> is when the character changes speed, for example from standing to
> walking, or from walking to running. I haven't yet actually tried the
> latter...

I'm sure I'll need to tackle those before I'm done.

>
> However, one thing I successfully implemented was an algorithm to
> preserve a somewhat realistic dependency between speed, step distance
> and step time. In short, the distance of one step length should be about
> inversely proportional with the time that it takes to take that step. So
> when the speed increases, it should be achieved partly by taking longer
> steps, and partly by taking quicker steps. However, as with most of my
> other work, this approach was based on a guess* of how these things
> work, rather than on actual scientific papers.

That is an interesting point.  I'll have to look into that.

Currently my Path spline is indexed on a time value, and I specify the
amount of time between each step.  I realize I can get into trouble with
that if I'm not careful (if I make him try to walk too fast), but it was
good enough for the present demonstration.

>
> * And I tried to do some walking and running at various velocities
> myself, which must have looked pretty silly, when at the same time I had
> to measure how long steps I took, and how fast. ;)

I, too, have spent several minutes walking up and down my hallway,
trying not to walk as if someone (me) is watching how I'm walking.

I've thought about taking a digital video camera to the gym to get
some footage of me and/or others on the treadmill (at various
speeds and gaits), except I don't have such a camera, and I don't
have a membership at a gym.

>
> Rune

Thanks for your comments.


Tom Bates.

--
#macro T(B)#local m=mod(B,8);#local B=floor(B/8);m#end#local C=126981491245202;#
local D=115694168992822;sphere_sweep{b_spline 16#local i=0;#while(i<16)<T(C)-4,T
(D)-4,10-i/10>.1#local i=i+1;#end pigment{rgb 1}}light_source{-10,1} //Tom Bates

Post a reply to this message