Behind the tedious process of creating Chinese computer fonts

0
116

MIT Technology Review

But there are tens of thousands of Chinese characters, and a 5 x 7 grid was too small to be readable. Chinese required a grid of 16 by 16 or larger, ie at least 32 bytes of memory (256 bits) per character. If you imagine a font with 70,000 low-resolution Chinese characters, the total memory requirement would exceed two megabytes. Even a font containing only 8,000 of the most common Chinese characters would take around 256 kilobytes to hold the bitmaps. That was four times the total storage capacity of most commercial PCs in the early 1980s.

As severe as these memory challenges were, the most difficult problems in producing low-resolution Chinese typefaces in the 1970s and 1980s were aesthetics and design. Long before anyone sat down with a program like Gridmaster, the lion’s share of the work was done from the computer using pen, paper, and correction fluid.

Designers spent years designing bitmaps that met the needs for low memory and retained a bit of calligraphic elegance. Lily Huan-Ming Ling (凌焕銘) and Ellen Di Giovanni were among those who created this character set, be it by hand drawing bitmap drafts for certain Chinese characters or digitizing with Gridmaster.

Draft bitmap drawings of Chinese characters for the Sinotype III font.

LOUIS ROSEFLOWER COLLECTION, STANFORD UNIVERSITY LIBRARY SPECIAL COLLECTIONS

The core problem designers faced was translating between two radically different spellings for Chinese: the hand-drawn character, made with a pen or brush, and the bitmap glyph, made with an arrangement of pixels called the are arranged on two axes. Designers had to decide how (and if) they would try to recreate certain spelling features of handwritten Chinese, such as: B. Entry Strokes, Bar Taper, and Exit Strokes.

In the case of the Sinotype III font, the process of designing and digitizing Chinese low-resolution bitmaps has been thoroughly documented. One of the most fascinating archive sources from this period is a folder full of grids with hand-drawn diamond symbols – sketches that were later digitized into bitmaps for many thousands of Chinese characters. Each of these characters was carefully laid out and, for the most part, edited by Louis Rosenblum and GARF, using correction fluid to erase any “bits” that the editor disapproved of. Above the initial set of green diamonds, a second set of red diamonds then marked the “final” design. Only then did the data entry work begin.

A close-up of a draft bitmap drawing from at (背, back, back) showing edits made with correction fluid.

LOUIS ROSEFLOWER COLLECTION, STANFORD UNIVERSITY LIBRARY SPECIAL COLLECTIONS

Given the sheer number of bitmaps the team needed to design – at least 3,000 (and ideally many more) if the machine could meet consumer needs – one might assume the designers were looking for ways to streamline their work. For example, one possibility would have been to duplicate Chinese radicals – the basic components of a sign – when they appeared from one sign to another in roughly the same position, size, and orientation. For example, when producing the many dozen common Chinese characters that contain the “women’s radical” (女), the GARF team could only have created a standard bitmap and then replicated (and theoretically should have) replicated it in each character in where this radical appeared.

However, such mechanistic decisions were not made, as the archive materials show. On the contrary, Louis Rosenblum insisted that designers adjust each of these components – often in an almost imperceptible way – to ensure that they harmonize with the overall character in which they appear.

For example, in the bitmaps for juan (娟, graceful) and mian (娩, to be supplied) – which each contain the radical “woman” – this radical has been changed very slightly. In the sign juan, the middle section of the female radical occupies a horizontal span of six pixels compared to five pixels in the sign mian. At the same time, however, the lower right curve of the female radical in the sign mian extends only one pixel further outwards, and in the sign juan this line does not extend at all.

The bitmap characters for juan (grace, graceful) and mian (娩, to be supplied) from the Sinotype III font, recreated by the author.

LOUIS ROSEFLOWER COLLECTION, STANFORD UNIVERSITY LIBRARY SPECIAL COLLECTIONS

Throughout this scripture, this accuracy was the rule rather than the exception.

When we compare the design bitmap drawings with their final shapes, we see that more changes have been made. For example, in the draft version of luo (罗, collect, net) the lower left stroke extends down at a perfect 45 ° angle before tapping into the digitized version of a smear. In the final version, however, the curve was “flattened”, starting at 45 °, but then leveling it.

A comparison of two draft versions of the character luo (罗, collect, net).

LOUIS ROSEFLOWER COLLECTION, STANFORD UNIVERSITY LIBRARY SPECIAL COLLECTIONS

Despite the seemingly small space designers had to work in, they had to make an astonishing number of decisions. And each of those decisions influenced every other decision they made for a particular character, as adding just a pixel often changed the overall horizontal and vertical balance.

The unforgiving size of the grid affected the designers’ work in different, unexpected ways. We see this most clearly in the diabolical problem of symmetry. Symmetrical layouts – which occur in many Chinese characters – were particularly difficult to represent in low-resolution systems, since according to the rules of mathematics, the creation of symmetry requires unusually large spatial zones. Even-sized bitmap grids (like the 16×16 grid) made symmetry impossible. GARF managed to achieve symmetry by using only part of the overall grid in many cases: just a 15 by 15 region within the entire 16 by 16 grid. This reduced the usable area even further.

Symmetry and asymmetry in the characters Shan (山, mount), Zhong (中, middle), Ri (日, sun) and Tian (田, field).

LOUIS ROSEFLOWER COLLECTION, STANFORD UNIVERSITY LIBRARY SPECIAL COLLECTIONS

The story becomes even more complex when we start comparing the bitmap fonts made by different companies or creators for different projects. Consider the water radical (氵) as it appeared in the Sinotype III font (bottom and right) as opposed to another early Chinese font created by HC Tien (left), a Sino-American psychotherapist and entrepreneur, Chinese computer science experimented with it in the 1970s and 1980s.

A comparison of the water radical (氵) as it appeared in the Sinotype III font (right) with an early Chinese font by HC Tien (left).

LOUIS ROSEFLOWER COLLECTION, STANFORD UNIVERSITY LIBRARY SPECIAL COLLECTIONS

As insignificant as the above examples may seem, each was a further decision (among thousands) that the GARF design team had to make, be it during the drafting or digitizing phase.

Of course, low resolution did not stay “low” for long. The advancement of computers led to ever denser bitmaps, ever faster processing speeds, and ever lower costs for memory. In today’s age of 4K resolution, retina displays, and more, it can be difficult to appreciate the art – both aesthetic and technical – that went into the creation of early Chinese bitmap fonts, limited as they were. But it was such solutions to problems that ultimately made computers, new media and the Internet accessible to a sixth of the world’s population.

Source Link

Leave a Reply