[Filip Vandueren]

I get how the shape can adapt, but how do the two texts adapt to each other? Just position wise ?

[Filip Vandueren]

Having the original text with asterisks in a separate layer is a good idea BTW.

Try this for your dual-styled layer:

• expression for sourceText:

t=thisComp.layer("inputTextLayer").text.sourceText;
st = t.style;
st.setText( t.replace(/\*/g,'') );

• add an animator with these properties:
  • Stroke Opacity: 100%
  • Stroke Color: expression: text.sourceText.style.fillColor + [0,0,0,1];
  • Stroke Width: expression: text.sourceText.style.fontSize/16;
    the smaller that last number, the fatter your Faux Bold, it depends on your font, anything from 8-30/40 could work (the thinner the font you start with, the higher this number will also likely be, it’s a design decision)
  • Optionally, you could add a tiny bit of tracking at about 1-2
• remove the default Range selector from the animator and add an expression selector.
  This is the expression for the selector amount:

t=thisComp.layer("inputTextLayer").text.sourceText.replace(/\n|\r|\u0003/g,"");
bold=false;
upTo = textIndex;
for (i=0; i<upTo; i++) {
    if (t[i]=="*") {
        bold=!bold;
        upTo++; i++;
    }
}
bold ? 100 : 0;

[Filip Vandueren]

Here’s a version that doesn’t try to do everything in 1 layer, and also a slightly different approach on how I stored the ranges.

– 1 guide-layer to enter the “html”

– 1 guide-layer to analyse what character-ranges are affected by what “tag”

– 1 layer that “renders” the html.

It should render about 2.5 times faster, but I’m sure there’s still room for improvement.

https://we.tl/t-UcUZZglNv7

[Filip Vandueren]

Hi Malcolm, as you know that solution was just a purely academic excercise to see if I could figure out some way to do it in just 1 layer.

I’ll see if I can come up with some optimisations, but expressions selectors always seem to bog down after effects. I’ll report back shortly.

[Filip Vandueren]

You are correct.

For example height gets interpreted as thisLayer.height and you can’t assign a value to that.

So it’s best to avoid variable names that are property and method names of thisComp, thisLayer, thisProperty

[Filip Vandueren]

A solutin without scripts:

When you rename your comp to “0001”, then each subsequent duplicate of the comp will increase their name 2,3,4…

so your text-layer’s expression could simply be:

thisComp.name;

or if you name it something like “Hologram Area #0001”,

you could do:

thisComp.name.split(“#”)[1];

to get the final part of the name that has it’s number.

You’ll manually have to duplicate the comp 500 times, and add to the render queue.

[Filip Vandueren]

What would you want the extra letters do ? Squish together until they are width 0? Fade out, then the remaining letters merge? Move out of the way in a different manner ? I think you need to design an example first, hand animate it, and then we can think about how to Automate that.

[Filip Vandueren]

Hi Evgueny,

The Speed Graph Curves are the Derivative (Gradient) of the Value Graph Curve.

The tangents (which are always horizontal) also have a relation to the tangents of the Value Graph Curve: the Y-position is the gradient of the Tangent of the bezier handle in the value graph, the X-Width is 1/2 of the X-width of the Bezier handle of the value Graph.

[Filip Vandueren]

Perhaps you can make a hidden copy of the layer with 3D Channel Extract, then use sampleImage() on the resulting grayscale depth matte

[Filip Vandueren]

You would start with a dataset, like a .csv.

Then to draw the line; this expression on a Path should get you started:

w=1600;
pts=[];
endF=1+timeToFrames(time);
for (i=0; i<=endF; i++) {
    data = parseFloat(footage("data.csv").dataValue([1,i]));
    pts.push([i*(w/endF),data]);
}
createPath(pts,[],[],false);

That will draw a line, that gets a new datapoint every frame from column 2 of data.csv, but the total width will always be 1600 pixels.