I am using Astute Graphics Texturino to add raster textures on top of my illustrator vectors. Then I import them to After effects (for Character rigging in duik)
I'm just starting this process, so I haven't hit any snags yet, but I'd rather import these the correct way and not deal with slow or crashing after effects when I have many characters on screen. Is it better to conver these Illustrator files to Photoshop. Is there an easy way to do this maintain the layers?
As far as AE is concerned this is pretty much irrelevant. the more interesting question is indeed whether the underlying AI engine will crash eventually, being that AI itself is terrible at dealing with extensive amounts of raster data. and that's kind of the problem: A conversion with layers may already fail because the AI PSD export runs out of memory. feel free to try, though. I'd definitely consider using even the largest PSD a better alternative than the AI file, especially given that you already deal with raster data and applying any effects would also rasterize layers even if they are vector-based.
As soon as you add a Raster image to a vector shape in Illustrator it loses all vector information so collapse transformations will not let you scale up the layer without artifacts. You just complicated the vector file.
A much better option would be to convert the layers with the Raster image textures to PSD files. The challenge is the size. You need to figure out what size these files need to be. If you start out with an AI canvas that is 1920 X 1080 because you are working on an HD project, and the full body of your character fills the frame you are only OK if you are always going to show your character's full body. If you want to push in on the character at some point to create a head and shoulders shot then you need a bigger artboard.
The easiest workflow is to create your Illustration on an artboard that is the same size as the video frame, then scale up the illustration so that it looks like the closeup you need in your final edited animation. When the character is properly framed for its closeup you resize the artboard to include the entire character, separate the illustration into layers so it is easily animated, then, in almost every case involving bitmap textures, export the AI file to Photoshop so that the vector data that can foul up AE if it gets too complicated, is turned into pixels.
Then you import the PSD as a composition retaining layer size and let AE create the master comp for you. You then open up the Comp AE creates and resize the comp, add a new null to the timeline that you can use as the parent for all of the layers, then resize all layers by scaling the comp to set everything up for character rigging.
You have to live with the increased render time because of the large images that are scaled-down but at least each image will be smaller than the oversized PSD you need to do this kind of work.
That is why I avoid any pixels, raster images, or raster effects in AI files that I need to animate. I just keep it simple, never convert vector layers to shape layers unless I need a shape animator on a specific vector or need to extrude a specific shape.
Depending on the texture you are applying to the layer it might be more efficient to just import the textures, use a blend mode and use the AI layer as a track matte. I've done that kind of thing a lot. I'm working on a vector animation right now for a documentary that has gradient fills on a lot of vector layers and I'm just adding the textures using blend modes and Set Matte or Track matte to get them to overlay properly on the graphics.