The simple problem here is that SDL does not have full access to all your graphics card's hardware acceleration abilities, even when you ask for a hardware surface. You cannot approach the same speeds you get in OpenGL or Direct3D/DirectGraphics using SDL, sadly.

Well, perhaps it's time to change that. Does anybody around here have connections with the SDL development team?

If you read the article Grellin posted, it explains why software surfaces work faster. I have found that using alpha in SDL isn't so bad if you use it sparingly. I intend to switch to OpenGL someday though cus hardware alpha seems like much less work than finding all of these "tricks" to speed up the process. At any rate, depending on the amount you "need" to use alpha blending you might find that switching to OpenGL might be more work at this time than you intended causing you to lose interest in your current project which is never a good thing :) I say finish it one way, then make it better later or rewrite it using OpenGL. What is the worst that could happen? You would already have the game code written so you can spend more time on effects and just improve the overall look of the game.

Quote: Original post by PnP Bios
Well, perhaps it's time to change that. Does anybody around here have connections with the SDL development team?

Post on their mailing list. Basically all this sort of thing is pencilled in for SDL 2.0, which has been in development for years now. I think that a lot of the people who use SDL use it for retro-style games and don't care much about upgrading it. Someone was developing an OpenGL backend for it, but it doesn't seem to be ready yet. You can get some sort of OpenGL wrapper for SDL here (the most up-to-date version appears to be glSDL-20040602.tar.gz).

Quote: Q: What's up with SDL 2.0? Can I help?
A: SDL 2.0 will be a full redesign of the SDL functionality, based on what we've learned over the past four years. The architecture design is partially done, and we'll start prototyping the design soon. As soon as there's a working framework, we'll make it publicly available for comment and contributions. This new framework has about a year or so before we anticipate it being ready for stable release.

I wonder how they are going to impliment this...