next »

[Kramlack]

I was just curious about restrictions on things like 8, 16, and 32-Bit games. Specifically, colour pallet restrictions. Things like are you limited to a set number of colours you use in total? Or are you limited to the number of colours on screen at a single time? Things like that.

I did a quick google search and only turned up something about having a 256 colour restriction on 8-Bit games. So I'm looking to TIG to help me clear up any misinformation, and just to learn some things, in general, on this matter.

Thanks in advance for the help.

[Inanimate]

All systems have a pallet, yeah. And sprites also have a limit to the number of color from that pallet they can have.

[Kramlack]

Thanks, but can I get some specifics numbers. Are you restricted to set number of colours on screen at one point (like say 8 for 8-Bit)? Or is the only restriction the number of colours used in the whole game (like 256 for 8-Bit)?

[JamesGecko]

Happy fun binary time!

The X-bit refers to the number of binary bits used to denote the color. In binary, each bit can either be one or zero. (Yay!)

One bit: 0, 1 (black & white, two colors, not very exciting.) One bit squared, or 1^2
Two bits: 00, 01, 10, 11. (4 colors. Note that this is every possible combination of 1 and 0 for a two-bit number.) 2^2
Three bits: (Never seen a game use this. Why would you bother?)
Four bits: 0000, 0001, 0010, 0011, 0100, 0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110, 1111. (Count 'em! that's 16 colors. Again, every possible combination of ones and zeros for four bits. 16 = 4^2)
Sixteen bits: (Ack! I'm not writing all that out. Rest assured that it's 16^2 = 256 colors.)

Sorry, that's probably not very clear.

Edit: A lot of older games were limited in the number of colors they could have on the screen at once, so they used palates to work within the limitations of the bit limit. I think? I haven't done much 2D console programming.

[Radix]

As a style thing? I just choose an old system and look up its technical limitations, maybe grab a palette from an emulated game for an authentic look.

[Kramlack]

@JamesGecko: Ah, thanks, that's clears a bit up and it was an interesting read, since I'd heard about how the term bit refers to memory rather than colour (that was all I'd heard though).

@Radix: Yeah, it's more of a style thing I'm going for, but I was just curious if there was a right and wrong way of going about it, so I wanted to understand from a more accurate stand point then just look and grab.

Thanks all.

[moi]

As a style thing? I just choose an old system and look up its technical limitations, maybe grab a palette from an emulated game for an authentic look.

That is the best way to do, here are a few other wikipedia pages to get started

http://en.wikipedia.org/wiki/List_of_8-bit_computer_hardware_palettes
http://en.wikipedia.org/wiki/List_of_16-bit_computer_hardware_palettes
http://en.wikipedia.org/wiki/List_of_videogame_consoles_palettes
http://en.wikipedia.org/wiki/List_of_home_computers_by_video_hardware
http://en.wikipedia.org/wiki/Attribute_clash

[Aquin]

And don't forget, if you really wanna be authentic, there is a limit of sprites per line on some system and other strange quirks for other systems.

Of course, if you're just going for "a generic look", there's no need to get that insane.

[Alec S.]

You know, one of these days, I want to try to make a game with Atari 2600 restrictions including the sprites per line limitations.

[Aquin]

Heh, I'm actually really intrigued by the ZX Spectrum's limitations.  I wonder if there's anything I can do there.

I'll have to think about it... I may even go homebrew.  This is obviously at least a year away (got a lot of other projects first.)

[PGGB]

Happy fun binary time!

The X-bit refers to the number of binary bits used to denote the color. In binary, each bit can either be one or zero. (Yay!)

One bit: 0, 1 (black & white, two colors, not very exciting.) One bit squared, or 1^2
Two bits: 00, 01, 10, 11. (4 colors. Note that this is every possible combination of 1 and 0 for a two-bit number.) 2^2
Three bits: (Never seen a game use this. Why would you bother?)
Four bits: 0000, 0001, 0010, 0011, 0100, 0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110, 1111. (Count 'em! that's 16 colors. Again, every possible combination of ones and zeros for four bits. 16 = 4^2)
Sixteen bits: (Ack! I'm not writing all that out. Rest assured that it's 16^2 = 256 colors.)

Sorry, that's probably not very clear.

Edit: A lot of older games were limited in the number of colors they could have on the screen at once, so they used palates to work within the limitations of the bit limit. I think? I haven't done much 2D console programming.

It's 2^16 and not 16^2. That makes 65536 values. 8-Bit has 256 values because it is 2^8. Same for all other bit-values.

[moi]

There isn't really a relation between the number of bits and the number of colors available
The 8 bits MSX2+ could display 19268 colors on screen, the PC engine 482, etc...
It all depends on the graphic chipset.

[pgil]

Yeah, the "bits" of the cpu has very little to do with the bit depth of the graphics. I'm pretty sure the Atari 2600 could show more colors on screen at once than the Sega Genesis. Really, just look for a machine's specs if you want a specific look. And if you just want a generic "8-bit" look, you could just make your own palette with a small amount of colors and just use 2-8 of those colors for each sprite.

[moi]

s youtube channel has a lot of very interesting videos about the TMS9918 chipset which equipped many early 8 bit computers (similar in capabilities to the zx spectrum). Not the best chipset to do mockups for but interesting to get an idea of things

[JamesGecko]

It's 2^16 and not 16^2. That makes 65536 values. 8-Bit has 256 values because it is 2^8. Same for all other bit-values.

Doh! Thanks for catching that.

[mcc]

It's also worth noting that in a surprising number of cases "8-bit" "16-bit" "32-bit" "64-bit" etc was just a marketing thing that really didn't have anything reasonable to do with the technical specs of the system.

[moi]

I like the NES "4 colours per 8x8 block" limitation. Just don't restrict yourself to the original NES crap colours. (increase 8x8 proportionally to the new resolution)

[gnat]

From a technical standpoint, in this day and age, there's very little reason to restrict ones game to 8 bit or 16 bit color if one is on the PC platform. 32 bit mode can do anything an equivalent 8 bit or 16 bit mode can do (one can have the 8 bit style in a 32 bit mode). The performance issues of yesteryear are now irrelevant.

[Alec S.]

These days old style restrictions are used generally for stylistic rather than technical reasons, unless it's a homebrew being developed for an old system.

[Loren Schmidt]

I think practical considerations and style are both great reasons to consider setting technical restrictions. Making all your assets within certain technical guidelines is also one tool to help make your game look more visually self-consistent. Certainly it's only one way to make games- there will always be a place for cutting-edge 2d and 3d games as well.

It can also be really effective to set your own restrictions, rather than basing them on old hardware. For instance Saucelifter uses flat shaded polygons very effectively. It's a very self-consistent look, even though it's not based on a specific piece of hardware.