Sorry…still getting to grips with the terminology.
I cheated in the end by creating another comp with just the layer I wanted to zoom into (a coffee mug in this case that was part of another comp). The new comp I made 1920×1080 in size and increased the coffee mug to the full comp size, turned on continuous rasterisation and then dropped that into the main comp with the original comp (lower resolution coffee cup with other layers). Then just resized the new higher sized comp with the coffee mug and laid over the lower quality coffee cup and reduced the opacity of the layer underneath when I got to the bit in the animation where I zoomed in.
This is no doubt a very messy way of doing things, but it did mean I didn’t have to fuss about animating the coffee cup again with the new larger comp.
In future, I guess it is best to drop each layer into a separate new comp and make as large as I need to (preferably comp size) with the layer being comp size. Then nest each of those separate comps into a comp that I want for that particular set of animation then nest that comp into the main comp with all the other animated comps.
Or bring each comp (the separate illustrator files within the comps), animate, then pre-comp the lot to tidy up.
The reason for doing all this is that I understand that 3D space transformations don’t affect continuously rasterised layers, but by making the illustrator files large and turning on continuous rasterisation within the comp, but when that comp is nested seems to solve this problem. I still have great quality and all of the ability to animate the nested, non continuously rasterised layer.
I guess asking the illustrator to create each illustration the full size of the comp I’m working to (in this case 1920×1080) would be a good start!
Happy for someone to tell me i’m doing it all wrong and there is a better way!
Olly Lawer