Base85 / ASCII85 Encoder

Ascii85 encodes four bytes into five characters, against base64's four for three. In practice that means 25 per cent growth instead of 33 — it sounds small, but in a PostScript file or a PDF stream, where binary data lives embedded in text, the difference shows in the final size. The price is a wider alphabet, including punctuation that gets in the way inside URLs.

Type the text and get the encoding in the Adobe variant, wrapped in opening and closing markers. The maths treats each group of four bytes as a 32-bit number and writes it in base 85, starting at the exclamation mark. There is one shortcut: a whole group of zeros becomes a lone letter z, saving four characters in data with plenty of empty space.

The final group, when it does not hold four full bytes, is padded with zeros and then truncated — it comes out one character longer than the real byte count, which is how the decoder knows how many bytes to recover. Check against a known example: the text "Man " encodes to 9jqo^, and "sure." to F*2M7/c. Outside PDF and PostScript, Ascii85 rarely turns up: base64 won by being safe in more places.