I think the beauty of really classic pixel art - I'm talking NES and GB here - came directly from the struggle to overcome the limitations of the medium in that era. It's amazing that games like
Metroid and
Final Fantasy III achieved such distinct and memorable visual styles with a tiny palette to choose from and a maximum of a few colours per sprite. It's nothing short of incredible that games like
Link's Awakening or
Donkey Kong Land could have such artistic fidelity on probably the most restrictive screen of all time.
Then the SNES/GBC era was interesting because the limitations were effectively lifted. It was now possible to truly defy the expectations of the viewer.
Final Fantasy VI skewed the overworld with perspective effects;
Seiken Densetsu III established a new benchmark for detail, colour depth, and 'invisible tiling';
Mario Kart transposed sprite art into a 3D world.
The medium was finally subverted completely on PC by games like
Beneath a Steel Sky, which took hand-painted drawings and converted them to raster images of the correct bit-depth and colour palette.
This is essentially not sprite art any more. It has the superficial aesthetic of sprite art while having none of its form. This was really interesting and engaging at the time, when the game came out in the mid-90s.
But that's it for sprite art, isn't it? Contemporary pixel art just doesn't really scratch these itches for me. I'm not sure it can. The first sort of sprite art I like was based on its context; the second was based on the subversion of that context, which is no longer relevant; and the third was an innovation twenty-five years ago that has now become cliche. I'm just not sure what I'm supposed to see in modern attempts. It's all been done. Why are people still doing it?