Measured - byte-exact - 2026

Pedulli vs xz

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

TL;DR - honest
Pedulli is a best-of-N racer: it races xz, zstd, brotli and your data's SRD math, keeps the smallest verified output, always byte-exact (SHA-256 roundtrip) with a +1 byte never-worse floor on its own output. Because xz is one of the racers, Pedulli is never larger than the best standard codec (worst case +1 byte): it wins outright on structured classes (JSON, periodic, sparse) and ties the best codec on already-optimal / random data - including long English text, where it simply selects xz's own LZMA2 result.

The measured table

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

Inputxz -9 -ePedulli (best-of-N)Δ
MP4 master 10 MB9,468,380 B9,430,108 B-38,272 B
1 MiB of zeros156 B13 Bmuch smaller (redundant input)
1 MiB periodic binary1,108 B436 B-61%
JSON 31 KB4,360 B1,265 B-71%
Long English text 500 KB~145 KB~145 KBtie - selects xz's LZMA2 result
HTML 161 KB35,705 B35,705 Btie - selects xz's LZMA2 result
Apache logs 3.5 MB211 KB211 KBtie - selects xz's LZMA2 result
Random bytes 1 MB1,049,000 B (+424 B)1,048,577 B (+1 B)423 B less overhead

Where xz alone is the right call (honest)

What Pedulli does that xz does not

When to switch

Switch to Pedulli for video masters, JSON/XML telemetry, periodic sensor streams, and the per-file +1 byte never-worse floor. On long English-text archives Pedulli ties xz by selecting its LZMA2 result - so you are never worse off than xz, and you win on every structured class.

More comparisons: gzip - zstd - 7-Zip - brotli - zpaq