[Micheel Leavitt]

Sry for not responding earlier, this works like a charm. I changed it to happen every approx. 1.62 seconds,and when I interpolated it it kind of came out weird. Heres how I did it

phi = (1+Math.sqrt(5))/2;
count = Math.floor(time/phi);
x = count*phi;
x1 = (count+1)*phi;
y = thisComp.layer(“Null 1”).effect(“fib”)(“Slider”).valueAtTime(x);
y1 = thisComp.layer(“Null 1”).effect(“fib”)(“Slider”).valueAtTime(x1);
x = linear(time,x,x1,y,y1);
[x,x]

What is wrong with this? It does indeed interpolate, but it waits a little too long to begin interpolating the next values. BTW this is on the scale of an object. the effect “fib” is this:

s = Math.sqrt(5);
phi =(1+s)/2;
count =Math.floor(time/phi);
nphi = (1-s)/2;
x = Math.pow(phi,count);
y = Math.pow(nphi,count);
(x-y)/s

which works perfectly as far as I have seen, which is to this 100th recursion. So can anyone see why is the interpolation not properly following the values given to it? Thanks for any help!

[Micheel Leavitt]

well, it kinda goes like this….

I set up a formula (on paper) that scales out an object following a fibonacci pattern, and the recursion counts up every approx. PI seconds. The margin of error was about .05 on the frame desired, so I thought maybe the while loop would work to set up the recursion. I thought maybe the modulus time%Math.PI would be fine because it does change every frame, but I guess I’m wrong somewhere 🙁 Basically I have the formula all set up I just need a counter that will count up one every PI seconds x.x

I also figured out a formula to interpolate the values between each set of recursion, I just need the damn counter to test it. Thank you Todd for the help, I still don’t see why the modulus time%Math.PI won’t change, but then again my knowledge of loops is next to nil. Ah well.

[Micheel Leavitt]

So I should include a condition that omits t==0? I tried that by rewriting it:

while(t<.05&&t!=0) but it would still timeout 🙁 I figured time/PI at time=0 would equal 0 right? So maybe I'm setting this up wrong? Thanks for the help so far Todd!

[Micheel Leavitt]

Thats too bad 🙁 It is not nearly as easy to do quick synthetic elements such as particle generated smoke, powerfully recolored fractal noise, etc in PS or AI. Ah well, thanks Dave.

[Micheel Leavitt]

Will this work if there is a keyframe on every frame?

[Micheel Leavitt]

It really depends on what you are working with. My advise (This saves myself time) is to take the full image into AE and mask the first branch, duplicate the layer, delete the mask, and draw the next branch. If it’s a “blooming” object it’s easiest to just make a whole bunch of traced leaves or other elements in AI and slap them in wherever you need them. Using Stroke is easier on the CPU from my own experiments to animate the “vines.” Yet it still takes about two-three hours from the tracing on paper to getting the masked element to masking each individual item to animation. The really great thing is if you do in the highest resolution your CPU can handle you got an amazing animated flourish that can be used again and again in any size comp. Hope this helps!

[Micheel Leavitt]

Ah, I thought it looked like a bit of 3D shading… It does not have the streaks in ANY preview except Colorama applied. It was also done with Open GL on, which may be the problem. I’m using an ATI Radeon 2400 with 256mb of video memory, and it’s only animating a masked photoshop layer with stroke to create motion. I’ll try with GL turned off when I get home from school tonight and see if that helps. I am also using a premultiplied alpha channel, can that affect it? Thanks a lot for all the replies, and out of curiousity, if it is indeed a new graphics card that needs to be gotten, anyone know a good one that will work with a Dell slim model with Vista Premium 32 bit?

[Micheel Leavitt]

I found a post from a while back that was very similar to my problem, and I was able to figure it out:

fd=thisComp.frameDuration;
t=0;
x=0;
amp = 30;
while (time>t) {
accel = thisComp.layer(“Null 2”).effect(“Range 1”)(“Slider”).valueAtTime(t);
x+=accel*fd;
t+=fd;
}
value + amp*x;

I see what you mean by the needing to use the valueAtTime() to get the value, it has to be told what frame it needs to take the value from, which is one frame behind the calculation, right? This does what I was looking for, and I realized the linear() was freaking AE out when it was a part of the while() loop. Is it too much for it to handle? I am curious. Thank you Dan and Filip for the help!

[Micheel Leavitt]

I don’t know if this can help, but if nobody has a different solution and you can have control over the output use quicktime animation. It looks (to me) better than the comp’s motion blur, but I don’t really have a solution. Maybe you should say what effects you are using and what isn’t blurring.

[Micheel Leavitt]

hahaha, ya it does. << Feels stupid. Thanks again Dan.