I read here that sprites were usually 8x8 or 8x16. However, I downloaded the sprites here: and opened up the image in paint and mario is typically in dimensions not divisible by 8 pixels.įor example: the mario with the racoon tail is 21x28px and the regular big mario is 14x27px. #SUPER MARIO WORLD SPRITES POWER UPS FULL#Įven Mario running at full speed in 19x26px. I'm guessing the sprites were 24x32px to hold the big mario characters. How did they do collision detection where there was no pixels in the sprite? Would Mario actually hit something a few pixels early? Or would they calculate it appropriately? Why did they not use up the whole dimensions? I would think they would want to use up as much as possible, at least maxing out the length or width. I have looked on various wikis, but I did not find anything mentioning how the character sprite breakdown worked for the SMB3. There's usually a common misunderstanding that sprites always tile up together to form a solid divisible-by-8 formations on all axes, but there is absolutely no reason for this. It's applying limitations of the background grid ("nametable" on NES) to the sprite system, which is a completely different beast. The entire strength of a hardware sprite is its ability to go "off the grid", and every 8x8 or 8x16 sprite on the system behaves as a completely individual entity - at least on the hardware. The programming of the game will probably handle them in bunches known as "metasprites" (multiple hardware sprites making up a bigger sprite as you're describing).Įspecially on the NES it's very common to let sprites overlap - maybe to save graphics data for animations, or simply to allow more colors - a popular example being Mega Man's face, which uses a different palette than his armor. But this is actually more often the rule than the exception when it comes to NES games. #SUPER MARIO WORLD SPRITES POWER UPS SOFTWARE#įor an example of really clever and safe conserving use of metasprites, check out this post about how Blaster Master does it:Īs for the collision detection, that's entirely a software thing! The graphics have no relation to collisions at all, so never trust what you see. Usually, well-made games will always have a hitbox on your character that is smaller than the size of its sprite. The NES isn't even able to do hardware per-pixel collision detection (except on a single sprite, a feature that is never used for gameplay purposes) - that kind of design kind of died out with the Atari 2600. :)įinally, I'd like to point out you are right - any transparent pixels padding the edge of an individual 8x8 or 8x16 sprite IS a waste of data, but working outside of the 8/16 pixel constraints usually helps giving the game a more natural, dynamic look, which became even more common in the 16bit generation. This is probably one of the reasons SMB3 is usually considered one of the better looking games of the NES, some times appearing more like a SNES game with fewer colors.You'd think after 23 years society would have mined a game as popular as Super Mario World of all its secrets, but you'd be wrong. #SUPER MARIO WORLD SPRITES POWER UPS SOFTWARE#.#SUPER MARIO WORLD SPRITES POWER UPS FULL#.
0 Comments
Leave a Reply. |