Hacker News Reader

Comment by devindotcom | Hacker News Reader

There's actually a community of all-RGB pic creators: http://allrgb.com/

This is one of the most bizarre things I've seen in a long time.

Very psychedelic, I'd say.

haha..years ago when I used to do Win32/MFC development I was in the habit of writing code that painted random pixel colors on my windows desktop...you get the psychedelic effect if you don't do bit-block transfers but paint directly to the screen...it was always a fun time wasting activity.

I sometimes like to violate the browser canvas element in my spare time:

Chromium Version: http://webspekt.de/paintgen-client

node.js Version: http://webspekt.de/paintgen-server

My Server is pretty bad, so when in doubt, don't click.

Bizarre and very interesting. A lot of those look like an extreme version of dithering.

Eek, looks like this one is crumbling under our load too!

God dammit! We kill every site we stumble upon; maybe there should be a delay to show links/comments based on time-zones, that way we don't DDOS any site by mistake.

This solution proposed by jcs would work: https://lobste.rs/s/g0nxrf/we_have_the_ability_to_cache_and_.... (He’s not talking about HN, but it would work equally well for HN.)

The idea is to use the free CoralCDN (http://www.coralcdn.org/). HN could solve site-killing by prefetching every link with CoralCDN, so that the CDN would cache a version of the page. Then, if HN detects that a site (e.g. http://allrgb.com/) is down, it would link to the Coral version (e.g. http://allrgb.com.nyud.net/) until it’s up again. Or alternatively, instead of detecting when sites are down, it could show a control to replace any given link with the CoralCDN version and ask the user to toggle it if the link doesn’t work at first.

Very cool suggestion!

Would a really simple way to do this be to make a 4096x4096 square image, with 0,0,0 at the top left and 255,255,255 at the bottom right, then just go along from left to right incrementing by one (in base-256, essentially)?

Yeah, but it's not interesting or challenging to do that.

The start of the video shows a very similar approach.

Would be cool to re-apply all the algorithms to HSV space and see what comes out.

HSV is just a straight-forward transformation of RGB https://en.wikipedia.org/wiki/File:Hsl-and-hsv.svg

The set of “all colors” is exactly the same.

(Note, HSL/HSV are not suitable representations for almost any purpose, especially not for human user interfaces; it’s a tragedy that they are in such wide use.)

Whjy do you say that? I find them much more intuitive. For example, getting actually brown browns with RGB is tricky...with HSV it's trivial.

A rule of thumb if you have to find brown in RGB: use a bit of red, slightly less green, then a tiny bit of blue.

That kind of proves the point. How many of those rules of thumb are you going to have to formulate to express in RGB what's trivial to find in HSV?

Unfortunately I've already formulated them after working so much with RGB in low-level graphics code, pre-hsv() CSS, and some image editing software.

Actually, that sounds kind of like a fun challenge game: name a color, then try to guess RGB values for that color.

I've gotten pretty good at doing the reverse operation(given an RGB value, guess the color) because I'm red-green colorblind and that's how I figure out what color something is :)

The converse game exists: http://yizzle.com/whatthehex/

Edit: Actually, so does the “guess the RGB given the colour” direction, but I don’t find it as fun: http://mallory.jemts.com

jacobulus might have been referring to certain color spaces that are optimized for human use instead of for computer manipulation, such as HUSL (http://www.boronine.com/husl/). HUSL gives you a slider that actually changes the apparent brightness of a color to your eye: whereas HSL thinks that bright green is just as “light” as dark true blue, and HSV/HSB thinks that true blue is just as “bright” as white, HUSL colors with the same “lightness” actually do look equally light. Another human-optimized color space is CIELUV (https://en.wikipedia.org/wiki/CIELUV).

A color space is good for transforming colors along lines orthogonal to its axes. In RGB, you can't easily find another shade of a given hue. In HSV, this common task is trivial.

I was going to comment that HSL/HSV can be useful for generating colors that work well, but... the real problem is not luminance as much as the HSL values we commonly use. See http://en.wikipedia.org/wiki/HSL_and_HSV#Disadvantages where CIELAB L* is shown as a better representation. Though it's not perfectly matched to the human eye even so.

> See http://en.wikipedia.org/wiki/HSL_and_HSV#Disadvantages

Yeah, I wrote that. :-)

The mapping is not one-to-one though - for example when S is 0, all values of H give the same resulting colour, and worse when V is zero both H and S are useless. So an image exploring the full HSV space should look different, probably with a lot more of the grayscale.

I too have misgivings about HSV, the discontinuities created when generating hue ramps being particularly cruel. Using a rotation about the luma axis in YCbCr space gives much smoother rainbows but sadly not ones which hit all six primaries and secondaries. It's frustrating problem that perceptually H, S and V or L are what we think of when specifying a colour, but mapping them smoothly to RGB displays is so tough.

Reply on news.ycombinator.com