Hi all,
I’m thinking of making a game in a similar vein to “Baldur’s Gate”,
which has a non-repetitive background (ie: it’s not tiled as far as the
eye can tell). I was wondering if there was any advantage to having a
tiled system anyway, or if getting rid of it would be a benefit…
The game map will probably be a maximum of 6000x4800 pixels (10x10
screens, worst case), with a fog-of-war overlay of the same (although
this can be 1 bit deep, assuming SDL allows a 1-bit alpha to act as a
mask in a 1-5-5-5 display format). Now admittedly, this is 57Mb of
background, but I’m not sure I care - it’ll be a s/w surface and RAM is
cheap.
I can see the reasoning behind ground-level animations being limited
in size, but I don’t require a tile system for this, water animations or
trees waving in the wind can be superimposed on the background, and have
their co-ordinates matched to the correct area.
The procedure (it seems to me) would then just be
- blit background from (background: x,y) to (screen: 0,0)
- blit an ground-level animations (water etc.)
* blit ground-level objects (players & monsters, start of high-level
objects)
* blit higher-level objects (trees, anything taller than the
characters, etc.)
(*) needs to be done top-left to bottom-right as usual.
I can see there might be a small advantage to using tiles if I want to
break the map up into "this chunk is visible now, this chunk isn’t) so I
could optimise the memory use - I’m just not sure it’s worth it. In BG,
any explored area of a map is fair game to be viewed at any moment by
jumping to it from the overall-map screen. Sure I could load up the
tiles on-demand, but it seems to me there’d be a lot of disk access if
you’re scrolling fast across the map…
So, am I missing anything ?
Cheers,
Simon.