[Darby Edelen]

Could you share your Preferences > Previews > GPU Information…?

Also do you have any lights in your composition? What kind of layer are you testing with?

Darby Edelen

[Darby Edelen]

Interesting. It looks like this uses Shape Layers in lieu of the beam effect? Or are there options that allow you to choose the method of drawing the line? Shape Layers have their benefits but it’s pretty undesirable to see the ‘edge’ of them in 3D.

This doesn’t look like a replacement for Plexus, as that feature set goes well beyond connecting objects with lines, but if you aren’t comfortable rolling your own expressions and you need some lines this looks pretty good 🙂

Darby Edelen

[Darby Edelen]

Your text is coplanar with the wall. You need to separate them some.

I’d recommend parenting the text elements to the wall layer and then entering something like 1 or -1 (depending on which way your wall is facing) in the z position values of the text layers. You could decrease this value quite a bit if you want the text to look like it’s on the wall, or increase it if you want it to float off the wall.

The old “Classic 3D” renderer would use simple depth sorting in addition to layer order to determine which layer resides “on top” of another, but with the new Raytraced Renderer coplanar layers will always intersect in this strange grainy way regardless of layer order.

Darby Edelen

[Darby Edelen]

There are a couple of ways to do this.

My recommendation would be to select your mask and press cmd-shift-m/ctrl-shift-m (Mac/PC). This will give you a dialog that has the offsets of the mask’s bounding box. In order to get the width/height you’ll need to do some math:

Height = Bottom – Top
Width = Right – Left

These pixels are in the layer’s transform space so if the layer has been scaled to 50% you’ll need to multiply your results by 0.5 to get the size of the mask in the composition.

This is pretty easy but requires some simple arithmetic.

The other option is to start by pre-composing the masked layer with “Move All Attributes” selected. This will move the mask into the pre-comp.

Then go into the pre-comp and draw out a Region of Interest (see the AE help) that bounds the area. Under the Composition menu you can then use Composition > Crop Comp to Region of Interest.

In your parent composition the size of the pre-comp layer should now be the size of the region of interest you drew out. Of course this technique depends on you drawing an accurate region of interest, but it’s pretty fast and mindless.

Darby Edelen

[Darby Edelen]

Matchmoving, or camera tracking, is the name of the game. There are lots of tutorials available for these techniques online.

There are also lots of tools. CS6 includes a camera tracker effect. The Foundry has a camera tracker plug in available for AE. The latest versions of Mocha Pro include camera tracking. There’s also Syntheyes and Bojou.

Any of these may work but they’ll all work differently. I like The Foundry’s solution personally. You could also attempt to manually match the camera move.

Regardless of which way you choose having some basic information about the camera that filmed the shot will help. Especially the focal length, camera height and distance to (a) subject.

Darby Edelen

[Darby Edelen]

[Adam Collins] “I am unsure what is going on because the pre-composed layers do not have a camera in them, nor do I have them in 3D inside of the nested comp.”

As Roland already stated you need to have the 3D switch enabled in the pre-comp. If you want motion blur on the layers you also need that enabled in the pre-comp. Layer switch states and material options are taken from the layers that are inside a pre-comp when it has collapse transformations enabled.

Furthermore, if you had a camera or lights in the collapsed pre-comp they will be ignored entirely in the parent composition. So that wouldn’t be your problem, but it might surprise you in the future. Best that you know now 🙂

Darby Edelen

[Darby Edelen]

My guess is that the artist created a matte for the foreground clouds and composited the text in behind the matted clouds.

As for the tracking, those shots are all very short and my guess is that they were manually match moved. Most camera trackers I’m familiar with would have a hell of a time trying to track clouds like that.

If your shot is not clouds then camera tracking might be possible for you. It all really depends on your source material and the result you want.

Darby Edelen

[Darby Edelen]

Another good thing to modify if you are using a wiggle() expression are the octaves.

The first two properties supplied to wiggle (frequency and amplitude) are required for the expression to work, the rest of them are optional. If you supply a value for the 3rd property in the wiggle you can increase the number of octaves from the default:

wiggle(1, 10, 3);

The above, for example, will wiggle with a frequency of 1 Hz, an amplitude of 10 but with 3 octaves. This is similar to the “Complexity” of the Fractal Noise effect. It will preserve the overall effect of the wiggle but create additional “sub-wiggles” that add detail.

Here’s an image of a wiggle() with 1 octave:

And here’s a wiggle() with 3 octaves:

Note that both curves follow the same basic path, but the one with 3 octaves has more “sub-wiggles.” If you want to reduce or increase the effect that these additional octaves have you can add a 4th property to wiggle():

wiggle(1, 10, 3, 0.2);

The default value is 0.5. A value of 0.2 would reduce the effect that the additional octaves have on the total wiggle.

Darby Edelen

[Darby Edelen]

You could try creating a circular ramp, blending some fractal noise into it, distorting it maybe with Turbulent Displace and using the Colorama effect to colorize it.

Darby Edelen

[Darby Edelen]

Auto Orient can cause a layer to turn on the x or z axis which may not be desirable. If it doesn’t give you the look you like try this expression from Dan Ebberts:

https://www.motionscript.com/design-guide/auto-orient-y-only.html

Darby Edelen