[Kalleheikki Kannisto]

With no selection, call up the Lens Blur dialog. In there, you’ll find an option “Depth map” where the source default is “nothing”. Select your alpha channel there and look at how the preview looks. If the wrong part is blurred (the object instead of the background), you can pull the “blur focal depth” setting to the other end. Or invert your alpha channel and try again.

PS. These might not be the exact names of the options, I have Finnish set as my interface language.

[Kalleheikki Kannisto]

Make a alpha channel that has the car masked, then apply Lens Blur using the alpha mask you created to control the blur area (option inside the Lens Blur effect).

[Kalleheikki Kannisto]

Short answer: After Effects is way more capable of handling moving images than Photoshop — by miles and miles.

You can remap time in After Effects on a single footage layer to speed up or slow down the typing.

You might be able to find a tutorial that talks about Time Remapping and achieve this. The main problem being that if you don’t know the basics of After Effects and something doesn’t seem to be working, it can be very difficult to figure out what went wrong. So you would want to get at least the basics of After Effects under your belt first. Interfacewise it is not that dissimilar to Photoshop, but on the other hand, some things work very differently.

[Kalleheikki Kannisto]

You can copy the masks one by one to separate shape layers. Just have to roll down all the way to the mask path, copy and then paste onto a shape layer.

[Kalleheikki Kannisto]

Try asking in the After Effects Expressions forum. At least Dan Ebberts will be able to tell if you can do this reading from an external text file.

You can copy the list onto a hidden text layer and read them onto a visible text layer with an expression, but I don’t know if this amount of text causes a problem. Probably not.

In either case, it is an expression or scripting question.

[Kalleheikki Kannisto]

You can zoom out of the other views in the same way you can zoom out of the perspective camera view, pushing the view close or further out with the “zoom camera” tool.

[Kalleheikki Kannisto]

I made a simplified sample file:

9208_delayedripple.aep.zip

[Kalleheikki Kannisto]

I assume you already know how to record and run actions and the main problem is selecting the gradient background?

If it is always the same exact gradient and size, you can use the difference blend mode to help create the selection. Overlay the gradient and the darkest blacks will be your selection. Or at least a rough selection that you need to smooth out some.

If the sizes and the gradients vary, it won’t be a simple matter. It is possible to create a rather complicated action that will select two colors from the image and create a gradient with them and then do the difference blend to get the background areas selection.

In any case, you probably want to do it so that you can adjust the results manually, i.e. use layers that you can modify if the action doesn’t dial it in perfectly. I suggest using layers masks, as you will likely need to paint in some details in the dresses that get mistaken for the background.

[Kalleheikki Kannisto]

Another strategy could be to blur them out so you essentially just get the effect of the colored lights. I blurred the screens (on a separate, masked layer) by 50 pixels, gaussian, and then curved them lighter. If you need some texture in addition to footage on the layers, you can fake it with an overlay texture layer.

[Kalleheikki Kannisto]

Lots going on there but the basic method revolves around using a track matte of footage of splattering paint to mask out other footage. Then repeat that about 250 times and compose them into sets of pre-composed 3D layers with camera moves and finally add high contrast color correction.