the complete quote:
"PJ standards for creating pixel art dos not allow the usage of these 'dirty' tools. Basically, pixel art should be able to be drawn by manually choosing colours and placing them 1 pixel at a time... NOT by doing all of that and then using 'dirty' tools to create highlights and shades. There is NO way 728 colours is at all acceptable for a piece to be submitted at PJ."
I don't recall whether this was mentioned in the gallery:
NOT by doing all of that and then using 'dirty' tools to create highlights and shades
Anyway, in fact
what I did was the reverse. I made an outline, indexpainted it, then cleaned it up. And I stated that this was the case in the very first blurb IIRC. The reason the lavender looks messy after cleanup, is because well, lavender DOES look messy like that. My cleanup actually deliberately made things look significantly MORE messy, since before that it was not uniformly messy.
how am I contradicting myself? It is clear to me that I said "should be able to be drawn" I'm sorry if that leaves it up to interpretation but that means that the end result should look like it was done as such giving the end result a look and feel to it. The process is somewhat important but the end result is what is seen and judged upon. And you can't deny that a little of the process is important to know.
Certainly. And
that's why I objected so strenuously in this case; because
this is not consistent - otherwise, I would not have been punished for explaining my methods (as jalonso said "To be fair to others
because you freely admit to some index painting this will be sent back")
The above encapsulates the problem: Vagueness. When you (pixeljoint moderators in general) can and have approved pieces that clearly have indexpainting (eg some of iLKke's recent stuff), and disapprove a piece because it explicitly says "partly indexpainted",
you are saying lying by omission is good, and this clearly indicates a problem with Pixel Joint culture -
lying by omission is not any more good than lying by commission. As Helm said, some people
will take up this lesson and omit such details deliberately, and there is nothing good for the PixelJoint community coming out of that.
This needs to be addressed promptly and clearly.
I personally make a point of describing my methods, and as I said there, I will continue to do so. I'm only one person, who clearly does not reflect the PixelJoint culture in which information about processes magically changes a picture.
Yeah, except no. That's not clear wording at all. You're definately able to make smudgy shit AA by using 2000 colours by hand, something I did when I was around 13 magnifying non-pixel grass tiles from Heroes of Might and Magic art to 'find the secrets of making nice low-res images' and copy it.
Me too. Well, not copying, but when I was 17 and trying to figure out Rayman's shading, I had this 'recursive shading' idea which basically required that there were no square blocks of color larger than 2x2 (so if there were, you had to get intermediate colors and fill it in, until the 'thickness' of all color regions was approximately 1px. I could easily attain > 300 colors on a piece, all hand picked, when doing this, and the result looked like polished crap (cause I barely understood how to draw people then). That was pixel art for sure by pixelblink's definition (done just with single pixel brush IIRC, not even floodfill), but it was also the kind of ignorant crap level of quality that you see in newbies and is regularly required to be revised on PJ, despite how long it took to make.
You can definitely make any possible picture 1 pixel at a time.
It's more a question of whether you would be sane to do so.
IMPO, there are many things in pixel art in which, once you understand clearly what you want and how it works,
the only point to doing them 1px at a time is mental masturbation / machoness.
Which, I hope we can all agree is not what we want PixelJoint to be about. Techniques are one thing, pretentious ego-stroking quite another.
I don't think Ai's way of making pixel art is the optimal way and I both like and dislike the results according to different cases, but I enjoy his search for the "perfect automated process" for making indexed-palette-based art. And I'd like for him to be able to showcase his pieces on pixeljoint without ever having to worry about the validity of the pixel-artness.
Since it's relevant I'll clarify: Of course there is no such thing as the perfect automated process
What I'm looking for is more like the perfect process which *includes* automation.
Characteristics including :
a) A marked
lack of aforementioned machoness / pretension - exercising control at the right detail level (so you are not manually pixeling 50% checkerboard dithers or reshaping curves pixels at a time)
b) increased control at higher levels (the same kind of control that CG artists get from layering and effect layers, allowing easy recoloring, pattern overlays, or other (possibly partially applied) automated processing by virtue of the way the image is structured)
examples being:
1.
Using clone tool to apply dithering patterns, or path tool to draw smooth curves.
2. Using
resynthesizer on a sketch of a texture to get a starting point for making tiles
3.
Using gradient map to apply coloration to different parts (the original sprite being drawn in grayscale)I've even used gaussian blur, to smooth out a silhouette sketch I knew would be a bit messy.
I think most of us try to cheat with making the necessary frameworks (eg perspective, action line->skeleton->shapes-> detailed anatomy, volumetric shading -> angular shading) and this would be less of an issue if we allowed ourselves more freedom in the way we build up things from idea to finished product, to
do what seems logical to get things done rather than masochistically restricting our workflow. The underlying structure is *the* most important thing in any picture; rendering effects are just rendering effects,
pixel-precision is essentially very cheap to attain if all our frameworking and sketching/concept art is properly done.