klinky wrote:
With this he can bloat his posts up 2 times possible 3 - 5 times larger than they have to be.
That's why he uses Microsoft products
(ok, I'll admit...that was mean)
Though there's still four big problems:
(1) The source files for an AMV generally take up much more space than the end product itself. (Pre-rendered sections alone are usually bigger than the Internet-distro version.) So much for bandwidth savings.
(2) The chance of someone having the same configuration -- filesystem, program configuration, and so forth -- as the author of an AMV is quite small.
(3) The proposed codec is precisely what Flash is, and would be subject to the same problems as Flash: while you could conceivably capture a video using vectors and mathematical transforms, you'd have to render it out sometime. This gives you two possibilities for playback:
(a) Realtime rendering. Way too slow, especially for trying to precisely recreate bitmapped images.
(b) Rendering to a file and playing back that file. This would work, but require gigantic amounts of disk space and I/O throughput to maintain full frame rate with perfect quality. Basically you'd be trying to play a HuffYUV or uncompressed file at full frame rate. Guess how many PCs have that kind of throughput: none. A compositing workstation at ILM might, but not Joe Average's Dell Dimension XPS.
(4) Given the resolution you'd need to not make the image look like crap, breaking a sequence of bitmaps down into vectors wouldn't result in much of a space savings anyway.
The idea is simply not feasible.