Author Topic: About the efficency of new models  (Read 1701 times)

Wedge

  • Hydlaa Notable
  • *
  • Posts: 619
    • View Profile
(No subject)
« Reply #15 on: March 29, 2004, 11:29:11 pm »
Some people think that for some reason... I dunno.  Of course advanced graphical matters are handled on the client side.  It\'s things like advanced-physics that can be a pain in the arse for bandwidth.  As long as the shaders and extra graphics are optional, with auto-fall backs to lower level things, I\'m fine.  Since as long as I have this MX 420, I don\'t think I\'ll be supporting much in the way of shaders.
Ninjas have feelings too.  Mostly they feel like dancing.



dfryer

  • Veteran
  • *
  • Posts: 1070
    • View Profile
(No subject)
« Reply #16 on: March 30, 2004, 01:51:24 am »
Wedge: As far as I understand it, skeletal animation is more CPU-intensive because it must compute the vertex displacements from the computed bone position, (including weighting and blending) instead of just reading off a list of coordinates from a frame of vertex-based animation.

This does reduce the RAM requirements though, which might help speed up RAM-limited systems.  It also allows more flexible characters, which is a Good Thing.

I think it would be theoretically possible to temporarily cache the results of the vertex calculations so as to make the trade-off back to low-CPU, high-RAM usage, but it would be difficult to get right and probably wouldn\'t provide much of a performance boost.
Quidquid latine dictum sit, altum sonatur.