[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [pygame] Anyone use Blender to create sprite sequences?
On Friday, September 24, 2004, at 07:09 PM, Jasper Phillips wrote:
Sure sounds like optimization to me, which is exactly why I've held off
moving to "filmstrips". Plus the optimization seems fairly minor for
pure pygame rendering.
On the other hand, perhaps it's a big win when rendering through openGL?
I can see how if you put the filmstrip into texture memory their _might_
be a way to change the drawn subrect, which I think would allow you to
simply change the "Frame Index" and avoid the cost of blitting.
Only problem is that looking at the pyopenGL docs nothing like this leaps
out at me. :-/ Anyone know off hand if something like this works?
-Jasper
On Fri, 24 Sep 2004, [utf-8] Bruno MartÃnez Aguerre wrote:
You don't have to split them up to use them. You can just blit the exact
frame from the surface. Also, this may save memory and simplify the
managment of the surfaces, as you have fewer.
Bruno
On Fri, 24 Sep 2004 15:10:21 -0400 (EDT), <shandy@geeky.net> wrote:
what does it matter? it's up to the coder to split them up into
animation frames anyways, right?
this is the reason why i don't bother with putting all the frames into
one image. I consider that optimization, something to be done when the
rest of your game is finished.
-sjbrown
Correct me if I'm wrong, but I believe the sprites in solarwolf are
assembled the wrong way. Each .png is a horizontally large image, when
it should be a vertical strip, to improve memory coherence.
What am I missing?
Bruno
Hello all.
I've used a variety of art programs to create art for a number of games and the latest way we've been using to organize our images is to create 256x256 sheets of images which are laid out on one or more grids for each animation.
The reason for the 256x256 max size for each of these is that some crummy video cards make anything larger than this look terrible.
The reason we have multiple frames of a single animation on one sheet is that apparently there is some reduction in the overhead one would have if you had to load each image individually from a separate file. This is especially true if you are loading these images over a network of some sort.
And for the sake of saving space, we tend to cram onto these sheets as much as possible so we end up often having multiple animations on one sheet if there is room.
One reason there is often extra room is that we try to keep the size of the large sheets to powers of two (16x16, 32x128, 256x64.. and so on...) is that various cards, or platforms tend to prefer that and will treat you nice if you treat them nice.
Then we organize our animations with a variety xml files which we usually create by hand but when someone makes a GUI tool, we use those.
I've found Blender to be very nice (especially after the UI revision), but I really love Maya, though that's way too spendy for personal projects. Photoshop is my preferred 2d platform (though gimp is workable) for taking stuff rendered by Blender or Maya and giving them some individual love and attention. I've got a batch file setup in photoshop to grab all the files in a directory and assemble them onto a sheet and do various other image processing tasks.