Live data from Hacker News

Asciifi - an ASCII art converter for a more civilized age

asciifi.com

21–28 of 28 posts

Re: Asciifi - an ASCII art converter for a more civilized age

#21
post #15

I don't consider auto generated ASCII figures (from aalib or libcaca or similar) to be art. To me, art implies there is an artist behind the work of art or the technique used to generate it. Real artists do their ASCII art 7 bits at a time.

In that case the developer of the converter (library) is the artist. I'm sure there's a lot of creative input involved in auto-converting a raster image into an ASCII figure.

Re: Asciifi - an ASCII art converter for a more civilized age

#23
post #9
post #7

Earlier quoted context omitted.

I wonder which algorithm you're using? From very quick testing, it looks like you're picking characters to represent various grayscale values, but I'm not sure that you're taking the shape of the characters into account like other converters do.

I used a combination of Tab Atkins' and the MooTools ASCII art ( https://github.com/khrome/MooAsciiArt.js ) algorithms. The four ASCII options came from those two libraries, as well (I played around with the values on various pictures until I found ones that I thought worked the best). You can also take a look at the Asciify and AsciiCharacter classes in the CoffeeScript file ( https://github.com/zachwill/asciifi/blo…

OK, and it seems that those algorithms only take the grayscale into account.

Compare http://paste.fulltxt.net/:gh (your output) with http://paste.fulltxt.net/h-,:wiS (aalib's output). This probably isn't the best choice of image, but you can still see how aalib approximates the edges of the drawing with all sort of characters.

Re: Asciifi - an ASCII art converter for a more civilized age

#24
post #14

Earlier quoted context omitted.

I did not, my browser window was not large enough to view both at once. Even so, this strikes me as a very strange use of a flash video.

It automatically scrolls down once you start playing the video.

Well, that much it does not do in my browser.

Re: Asciifi - an ASCII art converter for a more civilized age

#26
post #2

Haha, nice — I'm the developer if anyone has any questions (it's using a canvas element to convert images/data URIs to ASCII art). Also, here's a link to the Github repo: https://github.com/zachwill/asciifi

Great project -- I'd like to have the time for such side-projects

Re: Asciifi - an ASCII art converter for a more civilized age

#27
post #23
post #9

Earlier quoted context omitted.

I used a combination of Tab Atkins' and the MooTools ASCII art ( https://github.com/khrome/MooAsciiArt.js ) algorithms. The four ASCII options came from those two libraries, as well (I played around with the values on various pictures until I found ones that I thought worked the best). You can also take a look at the Asciify and AsciiCharacter classes in the CoffeeScript file ( https://github.com/zachwill/asciifi/blo…

OK, and it seems that those algorithms only take the grayscale into account. Compare http://paste.fulltxt.net/:gh (your output) with http://paste.fulltxt.net/h-,:wiS (aalib's output). This probably isn't the best choice of image, but you can still see how aalib approximates the edges of the drawing with all sort of characters.

>you can still see how aalib approximates the edges of the drawing with all sort of characters.

Wonder if that makes any sense for video, though?

Post reply on HN