Some questions about creating web animations with lighter data and exporting to vector Flash format.

Question

Hi, I will be creating animations in After Effects mostly using Illustrator vector files and will be exporting to SWF or XFL for using the animations on the web.
One of the priorities is light data as possible to provide viewers with slower connections with better experience. However the animation should also impress and I need to find a balance between performance and possible imagery.  The animations may have sound and voice narrative.
I need help with understanding which type of non-vector data is heavier than others. Non-vector image data or sound and voice? And from non-vector image data is some of it lighter or heavier than other - for example motion blur vs smooth gradient vs many colors?
I would greatly appreciate any suggestions, tips, pointers, etc.,

Answer 1

Hi, I will be creating animations in After Effects mostly using Illustrator vector files and will be exporting to SWF or XFL for using the animations on the web.
One of the priorities is light data as possible to provide viewers with slower connections with better experience. However the animation should also impress and I need to find a balance between performance and possible imagery.  The animations may have sound and voice narrative.
I need help with understanding which type of non-vector data is heavier than others. Non-vector image data or sound and voice? And from non-vector image data is some of it lighter or heavier than other - for example motion blur vs smooth gradient vs many colors?
Let's do a little "Myth Busters" on that one, shall we?
Vector data is per se not "lighter" than pixel data. You could have complex paths for which storing the information of the individual nodes takes up more space than a pixel image of the object.
Dynamic rasterization of vector data costs CPU performance and may in effect make things more slow than just playing a series of pixel images. That would apply to anything from path strokes to gradients.
All your pretty AI vectors don't mean anything if they are constructed in such a way that either AE or Flash need to rasterize them on import. Anything from converting Bezier paths to paths consisting of dense linear segments to converting fills, patterns, gradient meshes and so on to pixel data
Pixels are pixels. It does not matter whether they are gradients, mothionblur or just non-antialiased shapes. Flash doesn't care and neither image format that may be used in Flash stores any more or less info. They only differ in their methods, but in the end it's for you to decide whether you use JPEG, PNG8 or PNG24. They all have their pros and cons.
It makes a difference, if data is embedded in an SWF or if it is externally linked just as it makes a difference whether you enforce audio sync or not, playing every frame or not.
So for all intends and purposes, it's still up to you to make up your own mind and find the best solution. And Rick has a point - if you're really that concerned about efficiency, then do it in Flash or Flash Catalyst and consider the options there that may help like GPU accelerated layer styles for drop shadows etc., dynamic text or your vector data remaining vector in teh first place.
Mylenium