[ up | introduction · alphabet · comparison · license · download · screenshot · links ]

basE91 encoding

• introduction

basE91 is an advanced method for encoding binary data as ASCII characters. It is similar to UUencode or base64, but is more efficient. The overhead produced by basE91 depends on the input data. It amounts at most to 23% (versus 33% for base64) and can range down to 14%, which typically occurs on 0-byte blocks. This makes basE91 very useful for transferring larger files over binary unsafe connections like e-mail or terminal lines.

• alphabet

As the name suggests, basE91 needs 91 characters to represent the encoded binary data in ASCII. From the 94 printable ASCII characters (0x21-0x7E), the following three ones have been omitted to build the basE91 alphabet:

- (dash, 0x2D)
\ (backslash, 0x5C)
' (apostrophe, 0x27)

The translation table is composed of the remaining characters as shown below.

0	A	0x41	13	N	0x4E	26	a	0x61	39	n	0x6E	52	0	0x30	65	%	0x25	78	>	0x3E
1	B	0x42	14	O	0x4F	27	b	0x62	40	o	0x6F	53	1	0x31	66	&	0x26	79	?	0x3F
2	C	0x43	15	P	0x50	28	c	0x63	41	p	0x70	54	2	0x32	67	(	0x28	80	@	0x40
3	D	0x44	16	Q	0x51	29	d	0x64	42	q	0x71	55	3	0x33	68	)	0x29	81	[	0x5B
4	E	0x45	17	R	0x52	30	e	0x65	43	r	0x72	56	4	0x34	69	*	0x2A	82	]	0x5D
5	F	0x46	18	S	0x53	31	f	0x66	44	s	0x73	57	5	0x35	70	+	0x2B	83	^	0x5E
6	G	0x47	19	T	0x54	32	g	0x67	45	t	0x74	58	6	0x36	71	,	0x2C	84	_	0x5F
7	H	0x48	20	U	0x55	33	h	0x68	46	u	0x75	59	7	0x37	72	.	0x2E	85	`	0x60
8	I	0x49	21	V	0x56	34	i	0x69	47	v	0x76	60	8	0x38	73	/	0x2F	86	{	0x7B
9	J	0x4A	22	W	0x57	35	j	0x6A	48	w	0x77	61	9	0x39	74	:	0x3A	87	\|	0x7C
10	K	0x4B	23	X	0x58	36	k	0x6B	49	x	0x78	62	!	0x21	75	;	0x3B	88	}	0x7D
11	L	0x4C	24	Y	0x59	37	l	0x6C	50	y	0x79	63	#	0x23	76	<	0x3C	89	~	0x7E
12	M	0x4D	25	Z	0x5A	38	m	0x6D	51	z	0x7A	64	$	0x24	77	=	0x3D	90	"	0x22

• comparison

The table below shows a comparison of base64 and basE91, each method applied to 1 MiB binary data. The plain row lists the results with unformatted output and the MIME values show the amount of data that is produced, when you attach a file to an email (MIME requires a line break <CR><LF> after 76 characters).

	binary	base64	basE91
	binary	base64	worst case	random data	best case
plain	1048576 Byte =1024 KiB 100%	1398104 Byte +341 KiB 133.33%	1290556 Byte +236 KiB 123.08%	1289478 Byte +235 KiB 122.97%	1198373 Byte +146 KiB 114.29%
MIME	1048576 Byte =1024 KiB 100%	1434898 Byte +377 KiB 136.84%	1324518 Byte +269 KiB 126.32%	1323412 Byte +268 KiB 126.21%	1229911 Byte +177 KiB 117.29%

Using basE91 you need to transfer only 85.7% to 92.3% of the data, you would have to transfer with base64.

• license

basE91 has been developed by Joachim Henke, and is released as free software under the terms of the BSD license.

• download

Source code package (C, Java, PHP, 8086 assembly, AWK): base91-0.6.0.tar.gz (15 KiB)

Pre-compiled binaries for
GNU/Linux: base91-0.6.0-linux-i386.tar.gz (7 KiB),
Mac OS X: base91-0.6.0-osx-universal.tar.gz (11 KiB)

Java byte code: base91.jar (4 KiB)
To generate an e-mail attachment, run java -jar base91.jar file.dat, which encodes file.dat to pure ASCII data and writes this output to file.dat_b91.txt. Type java -jar base91.jar -h for all options, and read the included README file for further information.

• screenshot

basE91 command line tool

• links

SourceForge.net project page
freshmeat.net project pag e