These are hybrids.
Some things are hand pixelled some things are not.
Some things are pixelled then tweaked with tools/FX.

Here the idiots are the ones that don't ask.

Could you tell me which parts aren't pixel art then? I mean when I zoom in, I can tell that the backgrounds look more...antialiased than the foregrounds and it uses a lot more colors than most pixel art, but that's about it.

As you do more pixelling yourself and study more pixel art done by others, you'll develop a sense for what pixel art techniques used by a human tend to produce versus what colour reduction on a non-pixel art piece produces.
It tends to manifest mostly as well-defined, controlled clusters (pixel art) versus noisy clusters of less-contrasting colours (auto colour reduction), especially noticeable on objects meant to look smooth. Even when drawing something meant to be irregular and organic-looking, a (competent) pixel artist will have more cohesive clusters than a machine, because it tends to look better and read more easily.

I think most of the differences can be summarised thus: competent humans know when to eliminate unnecessary detail and colours, machines do not (yet).


Edited by eishiya - 15 December 2014 at 6:50am

With machines, it's garbage in, garbage out.
If you give it something pillow-shaded, you'll get something pillow-shaded out. If you give it something with nice shading, it'll give you something with that shading, the clusters will probably just be bad and the palette will likely be inefficient.
What this means is if you see pillow-shading, a human was most likely involved, and that human made bad decisions before handing the rest of the work off to a machine. A program can reduce colours for you, but it's not going to make your initial art for you or place your lights/shadows (unless you're dealing with something like 3D rendering or using SpriteLamp, in which case, you still have to make the models/initial art and place the lights - it's on you, the human, if the lighting's bad).

So for example, Rayman's pillow-shaded nose in the HUD, and the white blobby things in the first image, those are all a human's fault.


Edited by eishiya - 15 December 2014 at 7:32am

Came out a few months before Crash Bandicoot and Super Mario 64 (first 3D platformers); so...yea I'll go with that. It was right at the end of the era though (probably why the number of colors is so much higher than other pixel art.

I would say no. I have to Playstation and the game introduction is definably not pixel art. However then I plays it, I always more focused on the ridiculous high saturation.

The benefit to working that way is it allows animators who are used to non-PA animation to create animations for the game. It's quite possible that the artists doing the drawings/base animation and those doing the PA clean-up are different people. Starting big and shrinking down makes it easier to implement certain details and things like stretch-and-squish without over-/under-doing it, especially for pixel artists that aren't very experienced.

I think analyzing it in terms of lazy versus non-lazy is missing the point of it. This "hybrid" technique allows their artists to produce good-looking results without having to get better at PA animation, by being able to employ their animation skills independently of their PA skills.


Edited by eishiya - 07 January 2015 at 10:50am