[Freeciv-Dev] Re: different iso drawing algorithms

[Raimar Falke <rf13@xxxxxxxxxxxxxxxxx>]

On Sun, Dec 08, 2002 at 02:06:51PM +0100, Rafa³ Bursig wrote:
> >> SDL can pack transparent pixel in sprite but this is usefull only
> >> for read whole pixel array.
> >
> >Sounds like it uses something like a run-length encoding to minimize
> >memory footprint.
> >
> not only memory footprint but that sprites has less pexel to copy.
> That depend on numbers of transparent pixels and aspec of sprites ( 
> sprites where width >> high are faster/better compres )
> 
> This is called RLEACCEL

It is run-length encoding per line. See the top of
<http://www.libsdl.org/cgi/cvsweb.cgi/SDL12/src/video/SDL_RLEaccel.c?rev=1.5&content-type=text/x-cvsweb-markup&sortby=date>

> >> If you want have partial acces to pixel ( first metod )you must
> >> uncompres sprite and then read those pixels.
> >
> >It wouldn't matter if SDL supports a mask as a clipping
> >object. However it looks like it only support rectangulars.
> >
> Yes but that isn't problem here.
> If you want use that compresion all direct acces to pixels force 
> decompess sequence.
> And when you want get only part of pixels ( clipping rect mask ) from 
> sprite you must done directly acces.
> 
> SDL can support both metod but IMO secound has better performance 
> becouse can use RLEACCEL.

Can you qualify this? How much time do you need for a complete map
canvas redraw if you use RLEACCEL and how without?

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "Make it idiot-proof and someone will make a better idiot."