Hawkynt

LZFX

Improved LZF variant with better compression ratios while maintaining high speed. Uses hash-based LZ77 matching with 13-bit offset encoding and simple token format. Designed for applications requiring fast compression with minimal memory overhead.

Properties

Property	Value
Category	Compression Algorithms
Sub-category	Dictionary
Security status	Not classified
Complexity	Intermediate
Inventor	Andrew Collette
Year	2008
Origin	🇺🇸 United States
Source	algorithms/compression/lzfx.js

Security

Status: not classified — treat as unverified.

No vulnerabilities are recorded for this implementation.

Documentation

• LZFX Original Project
• LZFX GitHub Repository
• LZF Compression Filter for HDF5
• pcompress LZFX Implementation

References

• Original LZF by Marc Lehmann
• LZ77 Algorithm
• LZFX Format Specification

Test vectors

6 vectors ship with this algorithm and run in the test suite. Byte values are hexadecimal.

Vector 1 — All literals - no compression

Field	Value
input	41424344
expected	040000000341424344

Vector 2 — Repetition - AAAA

Field	Value
input	41414141
expected	0400000000412000

Vector 3 — Long repetition - 10 A’s

Field	Value
input	41414141414141414141
expected	0a0000000041e00000

Vector 4 — Pattern repetition - ABCABCABC

Field	Value
input	414243414243414243
expected	09000000024142438002

Vector 5 — Long text compression

Field	Value
input	48656c6c6f20576f726c64212048656c6c6f20576f726c6421
expected	190000000c48656c6c6f20576f726c642120e0030c

Vector 6 — Highly repetitive data

Field	Value
input	42424242424242424242424242424242 42424242424242424242424242424242 42424242424242424242424242424242 42424242424242424242424242424242 42424242424242424242424242424242 42424242424242424242424242424242 42424242
expected	640000000042e05a00

---

← All algorithms

This site is open source. Improve this page.