Red Shark News, The end of Pixels

littleclaude

Red Shark News, The end of Pixels



"Imagine putting tracing paper over an image, drawing over all the visible lines, and then shading or colouring in the spaces between the lines with all the appropriate gradients." - Page 3

Interesting reading, not sure where this takes games? vector, nurbs, who knows?

http://www.redsharknews.com/technology/item/1282-there-will-probably-be-no-8k-there-will-probably-be-no-pixels-at-all-in-the-future?utm_source=www.lwks.com+subscribers&utm_campaign=109255ea9a-RSN_Nov19_2013&utm_medium=email&utm_term=0_079aaa3026-109255ea9a-75260873

ambershee

It's a largely speculative article with little to no sources, references or even understanding of the problem (he also contradicts his own article title in his conclusion...). The idea of vector based animation isn't new; it's already used in some cartoons as a prime example. It's a natural extension of vector based image formats - which you may realise are generally only used in certain circumstances.

There are reasons for this - firstly vector based images are not good at representing anything with noise or sharp edges. They work with respect to mathematically defined lines and gradients. The real world is full of surfaces with noise, and surfaces with sharp edges, so in order to get the kind of detail you'd want, you'd have to represent a hefty amount of data - which still only has a limited degree of accuracy, so you're not solving the problem that pixels have limited resolution, you're just smoothing out the image at the resolution you have with the same results as upscaling a normal image (only with curves instead of square edges). Resolution is still an issue for anything that isn't simple.

This is why we don't ever store things like photographs in vector formats, only simple images like flags, diagrams and logos. They have a finite amount of required detail and are easily represented, real world objects generally speaking do not have this.

Consider that the worse issue is that you cannot realistically record anything in a vector based format - it would have to already be digital in order to be converted. The direct result of this is that it falls over at the first hurdle - if you want a vector based image that scales nicely to 8k resolution, you've got to record it with a minimum of 8k resolution anyway, and somewhere along the lines that data is going to become digital as part of the vectorisation process.