Historical Font Generator

Blackletter is the family. Fraktur is one branch of it, alongside Textura, Rotunda and Schwabacher, and the differences are real: Textura is the tall, tightly packed script of the Gutenberg Bible, all vertical strokes and diamond feet, while Fraktur broke some of those strokes into curves. Fraktur was cut in Nuremberg around 1513 for Emperor Maximilian I and went on to be the ordinary way German was printed for four hundred years.

It ended abruptly. In January 1941 the Nazi party issued the Normalschrifterlass, a decree from Martin Bormann declaring blackletter to be "Schwabacher Jewish letters" and ordering a switch to roman type. The claim was invented, and the real motive was almost certainly that occupied Europe could not read Fraktur. Either way, German printing changed script inside a year and never went back, which is why blackletter now reads as historical rather than merely German.

"Old English" is a third name for the same thing and a misleading one: it is a Victorian trade name for a blackletter revival, not a face used for the Old English language. Old English was written in insular script, and it needed three letters modern English lost, thorn (þ), eth (ð) and wynn (ƿ). Unicode encodes fraktur in the Mathematical Alphanumeric Symbols, with five capitals living elsewhere: C, H, I, R and Z are ℭ, ℌ, ℑ, ℜ and ℨ in the Letterlike Symbols block, so on some devices those five sit at a slightly different weight to the rest of the word. The gothic font generator goes further into the download side of this look.

In typography, Egyptian means slab serif: a face where the serifs are the same weight as the stems, squared off rather than bracketed. The name is an accident of fashion. Slab serifs appeared in London around 1815, when Vincent Figgins showed one in a specimen, at the height of the Egyptomania that followed Napoleon's campaign, and the label stuck even though the letterforms owe nothing to Egypt. The best known member, Clarendon, was cut by Robert Besley in 1845 and registered under Britain's ornamental designs act, making it one of the first typefaces with any legal protection at all.

If you arrived looking for hieroglyphs, here is the honest answer. Unicode does encode them: the Egyptian Hieroglyphs block runs from U+13000 to U+1342F, 1,071 signs based on Gardiner's sign list, added in Unicode 5.2 in 2009. But it is a writing system, not a style of the Latin alphabet. There is no letter-for-letter mapping from English to hieroglyphs, because hieroglyphs write consonants, syllables and whole ideas rather than the sounds of English, and any tool offering to "convert your name to hieroglyphs" is picking signs that vaguely resemble letters. Very few devices have a font covering the block, so even correct hieroglyphs will show as empty boxes for most readers.

The Unicode styles above are the slab-serif answer, not the hieroglyph one: heavy weights and small capitals carry the blocky, even-stroke feel as far as copy and paste text can. For the real thing, the Egyptian shelf below holds both meanings, genuine slab serifs and Egyptian-revival display faces.

Every typewriter face is monospaced for one reason: the carriage advanced the same distance no matter which typebar struck, so an i and an m had to occupy the same width. That single constraint produced everything people recognise as typewriter type. The i and the l grew long serifs to fill their cell, the m and w were squeezed narrow, and the whole face acquired an even, slightly awkward colour that no proportional design has.

Two faces carry most of the association. Courier was designed by Howard Kettler at IBM in 1955, and because IBM never secured exclusive rights to it, it spread everywhere and became the default look of a manuscript. ITC American Typewriter, drawn by Joel Kaden and Tony Stan in 1974, is the other one, and it is a deliberate fake: it looks like typewriting but is proportionally spaced, which is why it feels warmer.

Unicode has a monospace alphabet, U+1D670 onwards, with digits, and it is one of the few styles on this site that has digits at all. One caveat worth knowing. Those characters are a separate mathematical set, not an instruction to the renderer, so whether they truly line up in columns depends on the font your reader happens to have. Underline is included above because underlining was a real typewriter operation, done by backspacing and striking the underscore over the same letter.

Collegiate lettering is a specific thing rather than a general sporty feeling: heavy slab-serif capitals, usually with a stripe or a second outline running inside the stroke so the letter can be sewn or printed in two colours. It comes out of American athletic clothing, where the letters were cut from felt and chenille and applied to a jacket, and that manufacturing origin explains most of its rules. The forms are capitals only, because a lowercase letter has no room for an inline. The strokes are thick, because a thin one falls apart when it is cut from cloth.

Unicode has no inline or two-layer alphabet, so this is the weakest approximation on the page and the styles above are only a weight match. If you want the real look, the collegiate shelf below is the place, and note that most of those faces ship capitals only, so typing lowercase will get you nothing. For jersey numbering specifically the jersey font generator is closer to the mark.

Eighties display type has three recognisable moves: extreme condensation, a hard forward slant, and chrome or gradient fills over a grid. None of those are Unicode features. But there is one genuine connection, and it is the fullwidth forms above. Vaporwave, which is eighties nostalgia as an aesthetic, adopted spaced-out Latin text because it was copying Japanese-language interfaces of the period, where Latin letters occupy a full character cell. So the fullwidth block, encoded for a technical reason to do with CJK text layout, ended up as the closest thing to an eighties style in the standard.

The retro shelf below is the largest group on this page and it spans more than one decade: sign painting, motel and diner lettering, condensed movie-poster faces, and the chrome-and- wedge eighties look. If what you want is specifically the swollen seventies poster style, go to the groovy section instead, which is a different lineage and a different shelf. For the very late nineties and the turn of the century, the Y2K font generator picks up where this leaves off.

Groovy lettering follows rules, which is why it is instantly recognisable and why a nearest-lookalike font usually fails. The counters, the enclosed white spaces inside letters, are swollen almost shut. The fitting is extremely tight, letters touching or overlapping. And the baseline is negotiable: in the poster work this style comes from, letters were stretched and bent to fill whatever shape was left, which is a lettering decision rather than something a font can do on its own.

It has a real lineage. The psychedelic poster artists of the mid-sixties were openly reworking Vienna Secession lettering from sixty years earlier, and the softer commercial version that followed leaned on Cooper Black, drawn by Oswald Cooper in 1922 and revived so heavily in the seventies that it now reads as a seventies face. Unicode has nothing in this direction at all, which is why the styles above are a bold-and-round guess rather than a match. The psychedelic shelf below is the honest answer.

Strip away the association and pirate lettering is eighteenth-century jobbing print: a rough roman with heavy bracketed serifs, uneven inking, and the mixed sizes of a wanted bill or a ship's notice. That is why blackletter and small capitals above get you closer than anything self-described as pirate. The two symbols that do exist in Unicode are the skull and crossbones ☠ at U+2620 and the anchor ⚓ at U+2693, both copyable straight from this line.

The backend has only six faces tagged pirate and several are novelty hand lettering rather than period work, so there is no pirate group on the shelf below. If you want the weathered nautical look, the western font generator has the same wanted-poster ancestry with far more to choose from, and it is also the right page if you came here after looking for a cowboy font.

There are two very different situations inside this one word, and confusing them is why elvish text so often arrives as a row of empty boxes.

Runes are encoded. The Runic block, U+16A0 to U+16F8, was added in Unicode 3.0 in 1999 and covers the Elder Futhark, the Anglo-Saxon futhorc and the Younger Futhark together. The first six characters of the Elder Futhark are ᚠ ᚢ ᚦ ᚨ ᚱ ᚲ, which is where the name futhark comes from, and you can copy them from this line. They are ordinary Unicode, so they behave like any other text, though font coverage is patchy outside desktop systems.

Tengwar, the elvish script Tolkien invented, is not encoded. It has been proposed to Unicode repeatedly and has never been accepted, so the fonts that exist put the letterforms in the Private Use Area instead, an address range Unicode reserves for private agreements. The ConScript Unicode Registry, an unofficial convention, places Tengwar at U+E000 to U+E07F and Cirth at U+E080 to U+E0FF. The consequence is concrete: text typed with one of those fonts only looks like Tengwar on a machine that has the same font installed, it is not searchable, and pasted anywhere else it is a stream of undefined characters. The fantasy shelf below includes several genuine Tengwar faces, so treat them as artwork tools rather than as text.

Japanese type divides much as Latin type does, into two broad classes. Mincho corresponds to serif: modulated strokes with a small triangular finish, the equivalent of a serif, used for body text. Gothic corresponds to sans-serif: even strokes, no finish, used for signage and interfaces. Ryumin and Kozuka Mincho are Mincho families, Shin Go and Kozuka Gothic are Gothic ones, which is why the names on the shelf below look opaque if you are reading them as English descriptions.

The fullwidth Latin letters in the styles above are the genuine connection to Japanese typesetting. Japanese is set on a square grid where every character occupies one em cell, so a Latin letter mixed into Japanese text is either widened to fill a cell, zenkaku, or left narrow, hankaku. Unicode encoded both, and the fullwidth forms at U+FF21 onwards are the widened ones. Circled letters have a similar origin in Japanese enclosed alphanumerics. Neither turns your text into Japanese; they are Latin letters shaped by Japanese layout.

A real Japanese font is a different order of thing from a Latin one. It has to cover two kana syllabaries plus thousands of kanji: the JIS X 0208 standard alone defines 6,879 characters, and a professional family goes well beyond that. That is why these files run to several megabytes where a Latin face is usually under a hundred kilobytes, and why serving one on a web page without cutting it down is a mistake. The font subsetter exists for exactly this case.