[Darby Edelen]

I’ve noticed that the last expression I posted behaves a little strangely so I’d no longer necessarily recommend it 🙂

Darby Edelen

[Darby Edelen]

Well one option to simplify your expression while adding some additional functionality might be to use the Math.pow() function and account for negative returned values when the exponent is evenly divisible by 2. For example:

sign = 1;
exp = 3;
w = wiggle(2,4);
if((w < 0)&&!(exp % 2)) sign = -1;
sign * Math.pow(w, exp);

The more you increase the ‘exp’ value the more extreme and infrequent the returned values will be.

Something to consider, however, is that this “more extreme and infrequent” behavior will only occur when the value is wiggling around 0. If you want to get the same behavior around other values it’ll take a slightly different approach:

sign = 1;
exp = 3;
w = wiggle(2,4) - value;
if((w < 0)&&!(exp % 2)) sign = -1;
(sign * Math.pow(w, exp)) + value;

Another thing to consider is that as the amplitude of your wiggle increases the Math.pow() function will make the return values exponentially more extreme. You could eliminate that behavior by doing something like this (this would be my preferred expression):

f = 2;
a = 4;
exp = 3;
sign = 1;
w = wiggle(f,1) - value;
if((w < 0)&&!(exp % 2)) sign = -1;
(sign * a * Math.pow(w, exp)) + value;

So now you have control over the spacing of the spikes in the wiggle with the ‘exp’ value and separate control over the base frequency with the ‘f’ value and the maximum/minimum intensity with the ‘a’ value.

Darby Edelen

[Darby Edelen]

It’s not entirely clear what you’re after from your description but here are a couple of things to try.

You could use the linear() or ease() expression to remap the wiggle():


frequency = 1;
magnitude = 2;
w = wiggle(frequency,magnitude) - value;
value + linear(w, -magnitude / 10, magnitude / 10, -magnitude, magnitude);

Something like the above would make the transition from ‘-magnitude’ to ‘magnitude’ much more extreme, and the expression would be likely to hold those values for a short amount of time before changing. The above example assumes a one dimensional property (like opacity or rotation) instead of a 2D or 3D property (like position).

Another interesting option for wiggle() is to use the octaves parameter of the wiggle() to add more layers of complexity:

wiggle(1,2,3);

The above essentially combines 3 unique wiggles together. This results in more high frequency wiggling without dramatically changing the overall curve of the original wiggle (it adds more detail).

This is conceptually similar to modifying the Complexity parameter of the Fractal Noise effect. The default wiggle() has a complexity of 1 (only 1 octave) which is not very detailed.

Darby Edelen

[Darby Edelen]

[stig olsen] “My issue is that when Im selecting (masking) a part, it actually selects an another part than Im selecting, when the image is flopped.”

I can’t recreate this issue. Can you provide a sample project file?

Darby Edelen

[Darby Edelen]

[Kevin Camp] “if you wanted to key it, i’d modify darby’s recipe slightly and use the channel combiner effect, setting ‘from’ to max rgb and ‘to’ to alpha, then add the remove color matting effect.

this will keep all the color saturation (using shift channels will lose some of it).”

Good point. I always work in 32 bits per channel where that doesn’t happen 🙂

Darby Edelen

[Darby Edelen]

[G. Scott Scribner] “Yes, so because the layer is actually moved, I have to roto with mb on and hope it renders as I see it on screen. Previously I’d roto with mb off and render with it on and everything was fine. Now I just don’t trust it.”

If you roto with the composition’s motion blur switch off (leave the layer switches as they are) then there shouldn’t be any problem.

Darby Edelen

[Darby Edelen]

Sorry, here’s a more helpful bit of info. Try turning the Clip Output Black property to “On” at least for the top layer.

Darby Edelen

[Darby Edelen]

[G. Scott Scribner] “Photoshop does the same thing in 32bit. How can black be “added” to anything and be darker than the original?”

When you add a negative number it’s called subtraction 🙂

Darby Edelen

[Darby Edelen]

I would extend the alpha of the matte layer outward some using either matte>simple choker or channel>minimax set to maximum operating on the alpha channel.

This may eat into the contact shadow some in which case you could roto a copy of the matted ambient occlusion pass to bring it back.

Darby Edelen

[Darby Edelen]

It might be possible to format your camera track data in such a way that you could copy it as text and paste it as key frames to a camera layer.

The formatting of AE key frame data is very straight forward. You could ask the AE artist to copy some key frames and paste them into plain text files for reference.

As far as I recall there will be some header data in the clipboard that identifies the following data as AE key frames, then some additional relevant data like frame rate, followed by the property name and line break delimited plain text values of the frame number and key frame value.

Darby Edelen