NES Pallete Accuracy?

[Gaia]

As far back as the pixeltendo days, we ripped the NES games by means of grabbing a screenshot and plastering the sprite on paint and editing it out.

...Though it concerns me, are some of the NES sprites on the site's palletes actually correct? For example the Battletoads game rips seems to be using a different pallete altogether than what Metal Storm is using.

Even then some of the rips have muddled palletes as some of the monitors display the colors differently, and of course emulator differences (VirtuaNES, fceux, and nestopia immediately come to mind) as well.

There's also an issue in which is the -true- NES pallete out there as well, considering who compiled it, who ripped it, and who's using it to it's maximum potential.

So guys.. any ideas on what's an accurate NES pallete (other than plugging in an NES onto a PC, of course)?

[Gors]

personally (a pixelartist's view, not a ripper), considering that tSR is an archive of sprites for reference of graphics, I think it's too anal to complain about palette correctness for such a thing. Of course, when the palette is obviously off, that must be corrected, but when the discrepancy is produced by inevitable compression, or emulation differences, I don't really mind. Extracting the correct palette and protecting from compression is difficult, and so is rescanning all NES sheets to ensure all sprites use those colors. And the difference will not be too big either so I doubt there will come something good out of this.

So, an 'accurate' palette, as you put it, is any palette that has been extracted from the console at some point, disregarding any minucious changes that happen while saving it. Simple as that. It isn't practical for me to check the sheets and tell which one is accurate, but for me, it it's similar enough and follows the same tiling rules, it's A-OK.

Also, if you decide to go technical about NES sprites, you should re-check all SNES, GBA, Genesis, MSX, TG-16, Master System, C64 etc sprites in existence because they also have similar color limitations.

[Gaia]

Interesting, although I often see misconceptions of the term "8-bit" pop up more and more as I improve. A more realistic term would be "retro", right? Because if you look on some of the NES games (infamously, Mega Man), some of the games take advantage of layering, despite being you know, the usual three color-limit that most NES sprites commonly have.

I found this out by placing two different sprites together (one layered, one not) of similar colors. This is where (I find) the problem usually begins. And I'm not complaining at all per-say, given that curiousity killed the cat, so I apologize if it comes off as complaining.

[Gors]

I understand where you're coming from, but when ripping, you got to be as accurate as possible, given the tools you have. If your tools change the color millimetrically, then I think it's still passable, despite being innacurate. The sprite will still serve its purpose and no matter how the colors change, the 8x8 3 colors + alpha transparency will still apply.

[Sengir]

Gaia, the NES palette isn't subjective. It doesn't just depend on which emulator you use. I'm pretty certain that the method that stores NES graphics uses indices and not RGB values. That means that the palette is hardcoded into the NES somewhere.

Try looking at NES dev docs for an idea on how the colors are constructed (though honestly I don't think it matters to be that accurate)

[Gors]

oh right; that exists too. The palette is actually saved in the NES' graphic chip as pure data, therefore theoretically all NES will produce the same color, no exceptions, and the minimal differences will depend on the output (ie emulator, TV screen)'s configurations, not the NES itself.