pic

Case study: From ASCII Art to Subpixel ASCII+ Art
June 2017

When the character encoding standard “American Standard Code for Information Interchange” was introduced in 1963, hardly anybody could foresee that a little more than a decade later the world would loose itself in an early digital hype: ASCII art. Originally started as an underground movement, but not much later a major hit. For the younger generation not known with ASCII art (logically, that’s a pre-internet thing): ASCII art is an image displayed using letters. And not any possible letter, but just one of the 128 letters which are part of a character set called ASCII. That’s why it’s called ASCII art. This kind of visual art is not just made with any font, but with monospaced fonts. The uniform width allows for a pixel grid of a complete text paragraph. This matches the text directly with the pixels. If you say ASCII art, you automatically say monospaced fonts.

ASCII art is a versatile art, any possible image can be used, and meanwhile there are also several versions of ASCII art. Some more recent ways of creating ASCII art also include colours for example. But the early, should we say original ASCII art was plain black and white and strongly characterised by heavy limitations. The limitations of the early computer days were an important factor of the success of ASCII art. In the early computer days support for pictures was limited. The easiest way to create a picture back then, was to build it up out of letters yourself. Comparable to oversized emoticons. This method later on developed into photos which were automatically converted to letters, currently known as the most common form of ASCII art around. When computers progressed in the 1990s, more possibilities became available, the graphics improved and more non-monospaced fonts became at user’s disposal. Nice improvements, but bad for the ASCII art movement. ASCII art slowly lost its popularity. Meanwhile ASCII itself has been surpassed by Unicode. ASCII art itself has been declared dead by Microsoft in 1998, in favour of other formats like JPG and GIF and the promotion of proportional fonts. But ASCII art is still alive as ever, at least in our own studio. We love ASCII, we grew up with ASCII. We are ASCII kids. However, current technology allows for much more advanced ASCII art than in the 1970s. This is why we imagined to develop “subpixel ASCII art”. Same simplicity, but much more precise results.

20th century ASCII art
The traditional way of creating ASCII art is simple and straightforward. Every letter within a font has a certain greyscale value. While converting a pixel based image into text, the greyscale value of the pixel will correspond with the greyscale value of the glyph. A dark grey pixel will be converted into a glyph with a large surface (for example @), while a light grey pixel will become a character with a small surface, for example a punctuation mark like the comma. Easy. Sounds logical. All cool. No problem.

pic

This image shows how only 2 pixels are converted to text. If you would apply the same conversion for much more pixels, you’ll get the results of traditional ASCII art. Like this cat for example:

pic

21st century ASCII art
Current technology allows for a more sophisticated way of creating ASCII art. Only think of the possibilities of colors, combining different styles of the same font family, or subpixels. Just in case you have never heard about subpixels: every pixel on your screen actually consists of three pixels next to each other: red, green & blue (therefore it’s called RGB). Together they can make every color, including black (when all pixels are off) and white (when all pixels are on). By switching those subpixels on and off individually the resolution can be virtually tripled and used to render text sharper. Can the same approach of using parts of a pixel also be applied to ASCII art? We think so.

If only the overal color of a pixel is considered, two almost similar pixels can give very different end results. This T and 7 have more or less an identical grayscale, so both letters could be used to render both pixels, and many other letters could probably be used as well.

pic

However, the end result varies a lot. For an image it could make a big difference if the T or the 7 gets placed. Observing the subpixels could help in identifying the most optimal character for each pixel. Once a pixel is divided in 9 subpixels, pixels with a similar overal color suddenly show large mutual differences.

pic

In contradiction to subpixels in screen typography, this technique works in both directions, horizontal and vertical. So a L is selected for example for a pixel which is dark on the left and the bottom, and a T for a pixel which is dark on the top and in the middle. Although the overall color of these 2 pixels is similar, the end result of the pixels is different, and therefore the end result of the complete image will be significantly different between the old and new technique. Taking subpixels into consideration nonuples the actual resolution.

pic

Applying this subpixel technique to an image, in this case a white k on a black background, shows the advantages of this new technique. The diagonal stroke is displayed using a capital L, which automatically creates a diagonal line towards the right bottom. The shape of the L perfectly fits this location of the original image. Compare it to the diagonal stroke on top for example, which is displayed using ”. This glyph automatically creates a diagonal line towards the right top.

From ASCII to ASCII+
Fonts in the 1960s had a very limited character set, ASCII for example. Due to Unicode contemporary fonts include many more than 128 characters, offering a wider palette to select exactly the right character for each occasion. Having many more characters to choose from for each pixel contributes to a higher precision of the end result. So technically it’s not exactly ASCII art, but as described on Wikipedia the term ASCII art is also loosely used to refer to text based visual art in general. The term ASCII art became a byword for text based art, so to avoid confusion it’s wise to stick to that term. But because we cannot really speak any longer of ASCII art with this new technique, we named it ASCII+ art instead.

pic

The same tree, two different methods for text-based visual art. The tree on the left is made with the 20th century method; The tree on the right uses the new, subpixel method. Every trained typographer will recognise that the letters within the ASCII illustrations are not only selected by their overall color (like in traditional ASCII art), but also by the color of the subpixels. This will result in much more visible details of the final art.

pic

Especially graphic line art easily looses all sensibility in traditional ASCII art. This image full of curly lines is a thin wordmark (yes, it reads Underware), a good testcase to compare the different results.

pic

This is how the original image would look like in traditional ASCII art. Many details are lost, which makes the image – already hard to read – even harder to read.

pic

This text has exactly the same point size, the total amount of letters is identical. The traditional ASCII art destroys all the details & contrast, but with the subpixel generator everything looks great. Not only has the line itself a better appearance, even the contrast within the line remains visible.

Subpixel ASCII+ Art Generator
So far the theory. Practice is much more fun. Stay cool, and experience the difference between oldschool and contemporary text-based visual art yourself with a simple drag-and-drop tool. This tool has been optimised for Zeitung Mono, the monospaced font in the Zeitung font family. Go and play with the ASCII+ Art Generator.


pic



Want to read more about Subpixel ASCII+ Art, Zeitung Mono, web typography & more? Read the Zeitung newspaper.