Encodings
basE91
Joachim Henke basE91 with 13 or 14 bits per two printable characters. About 23 percent overhead
Encoding / basE91
basE91
Squeezes printable ASCII hardest.
- Overhead
- +23% on random bytes
- Padding
- none
- Standard
- Joachim Henke, basE91 reference implementation (base91.sourceforge.net)
- Family
- 1 in base91
family 1 / 20
Sample
- Out
- TPwJh>Io2Tv!8PE
- Back
- "hello world!"
ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789!#$%&()*+,./:;<=>?@[]^_`{|}~"
Access
- Import
import { base91 } from "@agntn/encodings/base91" - CLI
encodings encode base91 'hello world!' - Tryplayground with the sample above
The squeeziest one here. 91 printable ASCII characters, every one except -, \ and ', and a variable 13 or 14 bits per pair. About 23% overhead, against 33% for base64.
encode("base91", "Hello, World!"); // ">OwJh>}AQ;r@@Y?F"
There's no RFC. Joachim Henke's reference C code is the spec, and this follows it bit for bit. One difference. The reference skips characters outside the alphabet, this decoder throws. Why? Silently skipping input is how you lose a byte and never notice.