View Full Version : FiberFX and Kink

05-07-2009, 04:06 PM
So I have tried to wrap my head around some of the FiberFx settings in layout and I am just not getting it.

The model has various squares and strands all of which are part of one model. The stands are proportionate to the squares in size.

The real problem it seems is that some of the settings for fiberfx are relative to the local size of the strands while others seem to be relative the object as a whole regardless of the size of the strands.

With this particular scene Swirl and Random length proportionately have the same effect regardless the length of the strand guide. However, Kink and Width are opposite; if the length of the strand guide is short, Kink and Width values seem to have a much greater effect than if the guides are long. The longer the guides are the more the effects are diminished.

So I resolved the width problem by creating a null and setting a gradient with object distance for the input.

Using the same method for Kink does not work. I assume this is because object distance is relative to the object as a whole and not the individual fibers, so why does this technique work with width?

It could be that I just don't understand the gradients properly, either way I could use some help.

There are two cameras, you cannot tell from the images the squares are different sizes, but those from camera one are about 1/3 the size they are from camera two.

The scene and object files are included, version 9.6.

05-09-2009, 07:46 PM
Anyone want to look at the scene, theres not much to it, Bueller?

I just need some direction

05-12-2009, 11:50 AM
Not sure what you're going for. Can you simplify the problem?

With my so-so experience with FiberFX, I've noticed the different settings are unusually dependent on things that I wouldn't think otherwise. One big note is the fibers are dependent on geometry.... making the same settings to display different results from object to object.

05-12-2009, 02:13 PM
I would like both objects to have the same kink as they do in camera 2 or the second pic.

The scale of the geometry is different between the two test subjects, hence the percentage of distortion between the two. The first pic the object is much smaller, so the distortion is greater.

So I tried using a gradient based on Z distance for the kink, the further away the object is down the Z axis the smaller the kink % gets.

However, that effects the entire gradient scale instead of just the value I am changing in the gradient...seems buggy since I used the same technique on fiber width.

In any case, if you disable the kink, the fibers look just right regardless of the scale of the object. It's not until I introduce the kink variable that things go awry or at least don't seem to follow the same rules as the other Fiber variables.

05-12-2009, 04:33 PM
I tried using polygon area on the gradient input, but didn't have much success.

Interestingly, turn off the gradients for kink and width for a moment, and set width to 200% numerically, fiber widths at both camera positions render much more closely.

Could get Kink to behave.

FiberFX is still very buggy to me. I hope NewTek scrap it and start again in CORE.

05-13-2009, 11:33 PM
Strange, well at least I dont feel like I am completely nuts. I thought I might be doing something wrong, although it doesn't mean there isn't a different way.

I guess I will hold off on this project and visit it sometime in the future, thanks for taking a look.