[Filip Vandueren]

I don’t think that’s the kind of thing I would use a regexp for.

Maybe a python script to change/normalize the JSON files ?

Or just write the max value to a separate data point in the JSON.

Or can the expression be written in a way that averages out the JSONs ?

Do the JSON files always have the same amount of data (keyframes) in them ?

If you give a bit more info on how exactly the data is formatted, and how it is used, it shouldn’t be too hard to figure something out.

[Filip Vandueren]

Correct, invisible characters (opacity=0) still count for the sourceRect, the fix is to scale them to 0% so they don’t take up any space.

You should set the smoothness of the scale animator to 0 so it doesn’t scale the letters up, they’re only ever 0 or 100% nothing inbetween.

[Filip Vandueren]

I would make it with 3D layers and animate a light source, so the trees can cast real shadows on eachother

[Filip Vandueren]

The Ram is going to be the limiting factor right now.

[Filip Vandueren]

Alas, can’t be done (yet ?) all styles you set through expressions are applied to the entire text-layer.

[Filip Vandueren]

That would probably mean you are either working in an older version of After Effects, or the current version’s expressions – engine is set to Legacy mode (in the project settings )

I can rewrite the code tomorrow so it works on older versions of After Effects tomorrow

[Filip Vandueren]

try this:

( gaussRandom(0,10) + 5 ) % 10;

[Filip Vandueren]

I haven’t tried it yet, but maybe this library can help:

https://github.com/motiondeveloper/eTractors

[Filip Vandueren]

Try this, it’s a kind of brute-force approach (it tries every possible line length and chooses the one that has the least difference in line-lengths), but it works:

You will have to manually input the ideal maxWidth-value though, as that depends on the font used and the width of the box.

If the value is too high, then the text-box will also wrapping, and we don’t want that.

But you might not even need to use text-boxes using this approach ?

var txt = "Lorem ipsum dolor sit amet, consectetur adipiscing elit. Etiam rutrum sapien ultrices sapien condimentum, et congue risus placerat.";
maxWidth = 110;
minWidth = 30;
var results=[];
for (i=maxWidth; i>=minWidth; i--) {
	const splitter = new RegExp('(.{'+i+'}[^ ]* )', 'g');
	const newTxt = txt.replace(splitter, '$1\n'); 
	const lines = newTxt.split('\n');
	const lineLengths = lines.map(a=> a.length);
	results[i] = {
		index: i,
		txt: newTxt,
		lines: lines.length,
		contrast: Math.max.apply(Math, lineLengths) - Math.min.apply(Math, lineLengths)
	}
}
// best match has the lowest contrast (diff betweens longest and shortest line) and the lowest amount of lines
results.reduce( (acc, curr) => 	((curr.contrast < acc.contrast) && (curr.lines<=acc.lines)) ? curr : acc , results[maxWidth]).txt;

One Last thing: if there are line-breaks in the input-text already, then it won’t work as expected. You might want to first filter them out.

[Filip Vandueren]

Make sure the Illustrator layers have “continuously rasterize” enabled.

https://www.schoolofmotion.com/blog/collapse-transformations-continuously-rasterize-after-effects