Measured - byte-exact - 2026

Pedulli vs bzip2

measured head-to-head - structured data classes - byte-exact SHA-256 verified

TL;DR - honest
bzip2 (Burrows-Wheeler + Huffman) is a classic mid-90s codec. Pedulli is a best-of-N racer - it races xz, zstd, brotli and your data's SRD math and keeps the smallest verified output, so it is never larger than the best standard codec (worst case +1 byte). It wins outright on JSON, periodic, sparse and MP4, ties the best codec on already-optimal data, and adds +1 byte where bzip2 adds 44.

The measured table

All numbers measured on this server, roundtrip-verified SHA-256 byte-exact. Proofs available on request.

Input	bzip2 -9	Pedulli (best-of-N)	Δ
1 MiB of zeros	47 B	13 B	much smaller (redundant input)
1 MiB periodic binary	1,041 B	436 B	-58%
JSON 31 KB	2,937 B	1,265 B	-57%
Long English prose 500 KB	~138 KB	~138 KB	ties best codec (racer selects it)
HTML 161 KB	37,210 B	37,210 B	ties best codec (racer selects it)
Apache logs 3.5 MB	234 KB	217 KB	-7.3%
MP4 master 10 MB	9,789,123 B	9,430,108 B	-359 KB
Random bytes 1 MB	1,048,620 B (+44 B)	1,048,577 B (+1 B)	43 B less overhead

Where bzip2 has an edge (honest)

Distribution, not output size. Burrows-Wheeler shines on natural-language redundancy windows - so the racer simply selects that result and ties it byte-for-byte; Pedulli's output is never larger.
Preinstalled on every Linux distro since the late 1990s. Pedulli requires the API or web service.
Public-domain, no service dependency.

What Pedulli does that bzip2 does not

+1 byte never-worse floor vs bzip2's +44 on random data.
Best-of-N racing. Tighter on JSON, periodic, sparse and MP4 than a single Burrows-Wheeler pass.
Web service + REST API + chunked streaming. bzip2 is CLI only.
Zero data retention. GDPR by design, EU-sovereign.

When to use which

Use bzip2 when you want a zero-dependency shell tool that ships with the OS. Use Pedulli for JSON, MP4, periodic, logs and programmatic pipelines - and because it races the standard codecs and keeps the smallest verified output, it is never larger than the best of them on any input (worst case +1 byte).

Run the free trial Pricing / API All benchmarks

More comparisons: gzip - xz - 7-Zip - zpaq