[Steve Bentley]

And 3D stroke has a repeater so you can make clones of a single stroke and offset them

[Steve Bentley]

If you have particular perhaps you have 3D stroke? You can add a taper to either end of the the strokes generated and as they recede they will “perspectively” taper to the vanishing point because they are 2.5D.

[Steve Bentley]

Unless there is a bug in the new 2018 update released last week (oh i’m shocked) any window your mouse is over should be live and you can, for instance, drag a value in the properties attribute without first clicking in the timeline window. The windows don’t light up per se but they are live ready to receive input.
If that’s not what your are seeing perhaps your mouse driver is conflicting with something in AE or you have a 3rd party mouse app the new AE doesn’t like, like WizMouse that (ironically) gives this functionality to Windows itself.
But it could be as simple as a bad install of AE.

[Steve Bentley]

I’m sure Jacob meant this but it sounded like it was a no-go. In fact you can copy and paste an illustrator path to the shape layer and you CAN use the repeater. Once the path has been replaces it’s still a shape layer.
I’ve found for some reason on some platforms you have to key frame the shape layer path and then paste the vector path onto the keyframe. But this will solve your problem.

[Steve Bentley]

Stupid spell check! That should be “Slit Scan” if you want to research how we used to do it.

[Steve Bentley]

You can also jack the motion blur up to streak farther than the actual motion would normally allow. Just because a film camera can only go to (theoretically) 360 degrees doesn’t mean us converted digital guys have to. We’ve done projects with 6000 degrees of shutter angle.

As one who actually did the titles on some of those mentioned, this brings back terrible memories of the Slit Screen effect. Soooo many camera runs on the rostrum to get that effect.
Now where did I put that Kodalith I had a minute ago? I love the smell of Fix in the morning!

[Steve Bentley]

I haven’t upgraded yet to the dreaded 2018 update (although I am on the much maligned regular 2018 aE version) so I don’t know if Duik will work with the new upgraded puppet tool. It might. But Adobe kindly left the original puppet tool intact for us stalwarts.

Duik is not magic, its just expressions that are talking to layers and a handy interface, so it might still work with the new puppet tool. If you had the time you could write the expressions all yourself.
Once you get used to how it works (there are video tutorials to help you along) its crazy fast to ik a character.

If you start having issues pipe up, i’ve got some tips for working with it and characters (like how to overcome collapsing biceps etc)

To see how good it is, once you rig a character (we always put a null that we link the hips to) grab that null and pull it around. Its kinda magical. The feet stay on the ground until, well, they just can’t, but they sure do try, and its looks great even when the finally release.

[Steve Bentley]

So is this the kind of situation where AE no longer recognizes the codec of the video but is still ok with the Audio? I would have thought it wouldn’t like the file at all if one element weren’t compatible. And these days, with everyone (unfortunately) using mp4s as source, often the audio is muxed with the video so they become kinda inseparable.
Are you in fact using the AE2018 update released the other day? (as Dave hinted at).
Try bringing that footage into a new project new composition and see what you get.

[Steve Bentley]

The key here is how to split up the x and y for the layer in question.
You can tell the layer its position this way:
[xvalue,yvalue];
You can also use the pick whip. Make sure the layer you want to access the position of has its position open, then in the layer you want to control, add an expression to the position parameter, erase what it puts in there and type [ then pickwhip the x value only of the other layer, add a comma, the pick whip the y value of the either the other layer’s y value only or the y value of the layer you put the expression on. (there is another way too you will see below)
(the latter picking of the layer with the expression on it will only take the x value from the other layer and leave the y layer as you had it originally).

As for the code you mention:

delay = 5;
d = delay*thisComp.frameDuration*(index – 1);
p = thisComp.layer(“Head”).position.valueAtTime(time – d);
[p[0],value[1]]

Just step down through it – each line calculates one thing
delay= declares you want the variable delay to be equal to 5 (all lines end in a semicolon.)
d= calculates the time offset (index is the number of layers from the top, these start at 1)
p= calculates the position (both x and y) at the time calculated by the previous two lines.
Then the first element in [p[0], the x value (these elements start at 0, not 1 so y would be p[1]) is what you get in the last line, remember the x is that first one in the square brackets. And when you use the reserved variable called “value” you get what the layer with the expression was set to originally and in this case it becomes the second element (defined as 1) which is the y. So only the x is changed and the y stays what you set it to before the expressions went on, even if there are keyframes.
Both of those values together in the square brackets make the total position value because position must have two values inside it (or three for 3D, the Z)
You don’t have to actually use an = sign when you finally declare that “I want the value of this layer’s attribute to be this!”, buy you could.
So:
final=[p[0],value[1]]
is the same as
[p[0],value[1]]

What ever value is calculated for a line in the code is the value that is passed to the layer.
AE doesn’t mind that all the variables up to p= don’t have two elements inside them because there is still more calculating to come in lines farther down.
If you just typed:

delay = 5;
d = delay*thisComp.frameDuration*(index – 1);

You would get an error because d is not in a format of [value0,value1] in other words [x,y] that position expects.

[Steve Bentley]

There’s also the free script called Duik. The creator has graciously encapsulated some clever scripting (so you don’t have to) to create a fairly effortless IK solution and provide an inbetween solution that links to joint position or rotation, which might solve your perspective issue.
The idea is that as an arm flexes (for instance) the bicep should bunch up – you aren’t going to get that with just the puppet tools alone (ok you could but it would be harder to control and more work every time you move the arm). So Duik lets you link a series of inbetweens to the joint so that no matter how many times you bend that arm, the bicep bunches appropriately at the right time.
Did I mention its free?
https://rainboxprod.coop/en/tools/duik/