Tile-based graphics questions

[Jess Ragan]

This is pretty elementary, but I nevertheless feel compelled to ask. What's the difference between Pattern Tables (aka Character Base Blocks) and Name Tables (aka Screen Base Blocks)? I'm having trouble wrapping my head around these concepts. I believe that pattern tables are for the graphics and name tables are for their individual locations, but is there any way you can help clarify this for me? Thanks in advance.

[GroovyBee]

Do you have the link to where you've seen these terms used? They aren't terms I'd normally use when talking about tile based graphics.

[+selgus]

This is pretty elementary, but I nevertheless feel compelled to ask. What's the difference between Pattern Tables (aka Character Base Blocks) and Name Tables (aka Screen Base Blocks)? I'm having trouble wrapping my head around these concepts. I believe that pattern tables are for the graphics and name tables are for their individual locations, but is there any way you can help clarify this for me? Thanks in advance.

On the Genesis, the Pattern Tables are the actual character data. Your bitmap data for each sprite/character entry. The Name Table is which character to use at this location, and the attributes for said character (like if it is reflected on the X and/or Y axis). Each of these tables usually has a base address, and allows you to create compact graphics and reuse "Patterns" in your tiles.

--Selgus

[Nukey Shay]

The "name" table would be just a group of pointers to the actual data, right? These pointers could be vectors that indicate specific addresses (i.e. where the data starts for that bitmap or whatever), or as pointers to other tables (i.e. whether to display "time up" or "game over", for example).

[danwinslow]

Yes, I think the 'name table' would be the screen memory in a character mode? When you look at screen ram, you would see a bunch of character indexes, each one a 'pointer' into the character set data, which would be the 'pattern table'.

 

If you are using a softsprite type of mode, you would still have to have some kind of index or identification map that the screen is rendered from. In addition to pointing to the sprite data table, sometimes you also have other bits in the id that holds other attributes of the sprite, such as whether its hidden, reflected, and what game state the object that is represented by the sprite is in, ie., 'dead', 'exploding', etc.

[Rybags]

IIRC, the NES actually has a tile mode. Not sure how it works.

 

In theory, you could do it a number of ways... tokens that construct a 2x2, 4x4, whatever tile then blit to screen RAM. Token that just fetches and displays the data.

 

It's something that I don't think has many software implementations on the 8-bit Atari... given that we can scroll 4 characters over before needing to refresh screen contents, it's something that could be done easily in a game.

[youki]

On Colecovision, MSX , TI 99 and other computer/console using the same Video Display Processor.

 

Pattern is the table that defines the shape of each characters.

 

Name is the table that describe where these caracters are placed on the screen.

 

For instance in the pattern table , you define that the character number 65 will have the shape of a "A" or a "Pacman" or what you want.

 

then on the name table , which is for instance an array of 32 columns and 24 rows values , you put at column 10 rown 5 , the code 65 .

 

The result is that on the screen at column 10 row 5 , you will see the shape you defined in the pattern table for characters number 65 (a "A" or a "Pacamn" or else ).

[Jess Ragan]

Thanks guys, that helps clarify things quite a bit.