Deflate · Gzip Lv6 · Brotli Modeling · Shannon Entropy
Gzip & Brotli Compression Calculator
Benchmark compression ratios for web code, stylesheets, and JSON APIs. Measure byte-level savings, examine Shannon entropy, and simulate real-world mobile transfer speeds.
Sample Assets:
783 Bytes
Uncompressed Size
783 B
Original wire payload
Gzip (Deflate Lv6)
422 B
-46.1% Reduction
Brotli (Estimated)
346 B
-55.8% Reduction
Shannon Entropy
4.75 b/B
Max 8.0 (Random)
Relative Bandwidth Footprint
Uncompressed Original783 B (100%)
Gzip Compressed422 B (53.9%)
Brotli Compressed346 B (44.2%)
Network Transfer Time Simulation
| Network Profile | Raw Transfer Time | Gzip Transfer Time | Time Saved |
|---|---|---|---|
| Slow 3G (400 Kbps) | 15.7 ms | 8.4 ms | -7.2 ms |
| Fast 3G (1.6 Mbps) | 3.9 ms | 2.1 ms | -1.8 ms |
| 4G LTE (15 Mbps) | 0.4 ms | 0.2 ms | -0.2 ms |
| 5G / Fiber (100 Mbps) | 0.1 ms | 0.0 ms | -0.0 ms |
Lossless Web Compression: LZ77, Huffman Coding, Brotli & Shannon Entropy
How the Deflate (Gzip) Algorithm Compresses
Standard Gzip (RFC 1952) is built upon the DEFLATE algorithm, combining two distinct techniques:
- LZ77 Sliding Window: Replaces repeated substrings with backward references consisting of a distance and length pointer (e.g.
"the same word as 120 bytes ago for 14 characters"). - Huffman Coding: Replaces fixed 8-bit byte representations with variable-bit prefix codes, assigning fewer bits to high-frequency characters (like
eor space). - Overhead on Tiny Files: For strings smaller than ~150 bytes, Gzip can actuallyincrease file size due to the 10-byte header, 8-byte footer, and Huffman tree definitions.
Brotli vs. Gzip in Modern Browsers
Developed by Google and supported by all evergreen browsers via the br content-encoding header:
- Predefined Static Dictionary: Brotli embeds a 120 KB static dictionary containing over 13,000 common words, HTML tags (
<html>,<div>), and common JavaScript phrases. - 15% - 25% Smaller Bundles: Because common web patterns are already recognized in the static dictionary, Brotli compresses web assets substantially more effectively than Gzip.
- Woff2 Webfonts: The modern WOFF2 font format internally utilizes Brotli compression.
Understanding Shannon Entropy
Formulated by Claude Shannon in 1948, information entropy measures the average amount of information produced by each byte: $H = -\sum p_i \log_2(p_i)$. A pure repetitive string has near-zero entropy (highly compressible), while encrypted or already-compressed data approaches 8.0 bits per byte, representing mathematical randomness where no lossless compression can reduce size further.