( Let's pretend this activity started remotely near any sort of Bloodstained announcement so I can seem clever and coordinated.)
Welcome back to CC guys! This quarter we are tackling Castlevania, a title dear to many. It's creation planted the seeds for the later PSX title: Symphony of the night (with many installments in between), whose RPG elements, when mixed with Metroid's shortcut centric, interlaced maps, would go on to create the "Metroidvania" style of games that are so popular in indie development today. But let's not get too ahead of ourselves, as we are here to go back to the basics and reflect on not only what made this game great, but what could perhaps be improved graphically with a modern critical eye.
The purpose of this CC is to analyze what does and doesn't work about this game from a visual standpoint. Once the discussion get's going, take this information over to the activity where you are given the task of revamping a map for a more interesting, organic look!
Feel free to add more screenshots in the thread if you like. I'm a pleb that got too frustrated with Medusa heads to capture much of any later levels.
For those looking for some more in-depth information about the game, here's some analysis from our resident NES expert, Kasumi:
Castlevania 1 is UNROM. It does not have CHR ROM. This means the PPU (NES' video card basically) is reading from RAM, not ROM when it displays Castlevania's sprite tiles and background tiles. The game must copy tile data from its PRG ROM (a separate eprom tied to the CPU not the PPU) to the CHR RAM. Graphics can be compressed in CHR RAM games (since the CPU can transform the data to what the PPU expects before it copies it to RAM), but Castlevania does not appear to do this. CHR ROM on the other hand must be stored in the right format, because it's ROM and can't be modified.
NES has two sets of 256 eight by eight tiles in memory at any given time. Usually one set is used for the background, and the other is used for sprites. NES can display 64 sprites at any given time. Castlevania takes advantage of an 8x16 sprite mode. This lets one cover more of the screen with sprites than 8x8 mode (there is always a limit of 64 sprites, but 8x16 sprites are bigger). It also allows one to use both of the 256 tile sets to display sprites, but there's a caveat. Two adjacent tiles MUST be used, even for a sprite that is meant to be 8x8.
Here are the two sets in memory during the first gameplay scene of Castlevania:
On the left is the "sprite" set. On the right is the "background" set. But you'll notice a very common sprite on the background side: The heart pickup. You'll also see it has a blank tile immediately to its right. This is because the bottom half of an 8x16 sprite MUST use the tile to the right of the top half. If that tile were not blank, the heart would have those not blank pixels drawn below it. You'll see some other things that are clearly sprites on the background side.
In fact, Castlevania only changes half of its background tile set for each "level" during gameplay. The other 128 tiles don't change. Some are reserved for sprites, some are for the HUD, some are (I'm not familiar enough with Castlevania.)
Here are some "gameplay" tilesets to demonstrate this:
The bottom 7 rows of 16 tiles do not change. The 8 tiles in the middle of the 2 rows above the bottom 7 rows also do not change. This means to create a proper Castlevania level, you must use only 128 unique background tiles in it!
Here are all the graphics from Castlevania. (Maybe. Being a CHR RAM game, some might be compressed, but it seems like they didn't bother.)