I'm building a flash site (as3) with large videos that act as transitions from section to section. The videos don't scale; in order to accommodate users with large monitors, we're planning on playing back HD video and only showing users the center of the frame (depending on their screen size). That is, there will be a central area of roughly 1024x768 that we'll assume that all users will be able to view and where most of the information will reside, but we'll want users with big screens to be able to view all of the information and the entire scene around it as well.
My question is this: does it make more sense from a performance perspective to encode the video as hi-def FLV and play it back without modification, or to encode it at a reasonable resolution and then scale it to hi-def (with smoothing) on the fly? Loading time is not an issue as all clips will only be a few seconds long and will preload completely prior to playback, so the bottleneck will likely be on the playback side of things.
Many thanks!
Ultimately, we needed to play back video at 1680 x 1050, which is too hi-res for most of our audience's computers to play back natively. What we ended up doing is encoding the video in HD using On2's VP6-S codec's 720p presets and scaling to the desired resolution. The quality is more than acceptable and playback is decent on most computers, even when the whole video is visible (for users with smaller screens, the video is centered, not scaled to their screen size, so they just see a smaller portion of the video).
Anyway, ours is a fairly unusual set of circumstances, but hopefully this information will be helpful to someone else down the road.