-
Double checking : AE raster-based ? If yes, how to preserve appearance of small contents ?
Hi everyone,
So I am animating some sprites on a large world map. Each sprite is a pixel graphics inside its own composition. And even though my whole target composition (the world map) has a high resolution (5000*5000 or so) those elements are quite small – under 100*100 pixels. When I translate or rotate then around inside their own composition I notice they lose a lot of their sharpness – unless they are perfectly aligned with the target composition raster matrix. This also happens when rotating/translating the sprite composition within the main world map composition.
My guess is that this is somehow due to the fact that Atter Effects is not vector based but pixel/raster-based (right ? I also wanted to double-check).
But then I should be able to solve it by increasing the object size so that one single pixel of the object is represented by many pixels in the target composition, thus reducing that effect.
Through one example I can see this is not the case : I have a sprite of a boat whose mast is one pixel wide – it’s a column of pixels going up. When I increase the size of this sprite a LOT in the target composition the mast is still blurry on the edges, while when viewing it in its own composition it’s fine, sharp and crisp along the pixel edges.
So I feel I don’t understand what is going on or how I can improve the situation. Any help greatly appreciated !
Regards,
Charles -
7 Replies
-
You need to use the “collapse transforms” checkbox. What that does is hold off rasterising the nested comps until the outer comp is rasterised. There’s a good tutorial about collapse transformations here, and you can read the adobe docs here. It can be a little bit tricky to get your head around.
To answer your first question, yes, ultimately AE is ultimately raster based, because it produces video, and video is a raster based format (unless of course you’re exporting to flash). But within that it does have support for vectors, both natively with its built-in vector tools, and externally such as support for illustrator files. What you need to manage is when the rasterisation occurs, which is what the collapse transformations switch is for.
Say you have your vector-based boat layer nested within a comp (we’ll call it the outer comp). If you have collapse transformations off (as it is by default) and you scale the boat comp in the outer comp, it will treat the nested comp as if it were a rendered raster image, so the pixels in the mast will be interpolated, meaning fuzzies. If however you turn collapse transformations on the boat will stay a vector layer until after the scaling done in the outer comp meaning that the lines will be as crisp as the outer comp’s resolution allows. Does that make sense? Rule of thumb: most of the time if you’re using vector images in precomps, turn on collapse transformations.
Stephen Dixon
Editor, Animator, Motionographer
Museum Victoria -
Oh, BTW the collapse transformations switch is the little sun icon between the shy person and the quality slash in the layer switches panel.
Stephen Dixon
Editor, Animator, Motionographer
Museum Victoria -
Hi Steve, hi Dave
Thanks for the heads up on raster vs. vectors.
To answer the general question : I will rendering a video out of this animation which will be part of a movie I’m making in Adobe Premiere Pro. So it will end up being 1080p video.
Actually there are multiple views of the boat, which I alternate between depending on its orientation on the path. Therefore I have simple pixel graphics for front view, side view, back view … each of which I have isolated in its own composition (back_comp, front_comp, …), in case I need to make a modification later. In a separate composition called Boat I order those comps for each view in chronological order so it changes view at the moment I need. This composition called “Boat” goes in the main composition which contains the large base map. This is where I first see the fuzziness.
I have activated the option you suggested for all the individual view compositions within the Boat animation comp, and also for that Boat comp in the main comp – in fact I now have i activated everywhere – to no avail. What I am missing ?
Thanks again and in advance,
Charles -
I think part of the problem is that it sounds like the assets aren’t really suitable for video. A 1 pixel line will only look good when it is aligned perfectly with the pixel grid. Any other position and the edges will have to be interpolated, and since it’s only 1 pixel that means all of it will be interpolated.
Think about if it’s moved on X half a pixel. How do you rasterise a 1 pixel line that’s halfway between two columns of the pixel grid?
Can you really not afford to make it 2 pixels wide?
Stephen Dixon
Editor, Animator, Motionographer
Museum Victoria -
Hi again,
So no I cannot afford to make any of my content wider in terms of pixels. One possibility here, which I would accept, is to take my main comp and increase its resolution (ie, its canvas size) by a large factor, then blow up everything until it fits back on the canvas (ie by that same factor).
But I do have comps that are using this main comp (mostly to create some camera travelling effect where only a part of the main comp is seen at any time). How do prevent that blowing up from messing too much with that ?
Thanks in advance,
Charles -
Hi everyone,
So the magic button “Collapse transformations” needs to be activated all the way up and down the comps. Now that I’ve done this it looks (a lot) nicer. Will this also look like this when rendering the final comp ?
-
I don’t understand how this would help. If your final resolution is fixed, blowing up the comp will just create extra render time, and then the pixels will have to be reinterpolated back to the final resolution.
Stephen Dixon
Editor, Animator, Motionographer
Museum Victoria
Reply to this Discussion! Login or Sign Up