TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Modern LZ Compression (2019)

122 pointsby imadralmost 4 years ago

10 comments

retracalmost 4 years ago
There&#x27;s quite a lot of retro modern LZ activity too! LZ turns out to be amazing on old machines, often only a couple times slower than a block copy. Optimal compressors and control over the algorithm have led to some very tight demos.<p><a href="https:&#x2F;&#x2F;www.brutaldeluxe.fr&#x2F;products&#x2F;crossdevtools&#x2F;lz4&#x2F;index.html" rel="nofollow">https:&#x2F;&#x2F;www.brutaldeluxe.fr&#x2F;products&#x2F;crossdevtools&#x2F;lz4&#x2F;index...</a> LZ4 Data Compression - a rather long and in-depth article looking at LZ4 on the 65816 for the Apple IIgs with a decompressor that exploits that processor&#x27;s block copy.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;emmanuel-marty&#x2F;lzsa" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;emmanuel-marty&#x2F;lzsa</a> - LZSA - a LZ4-like modern LZ that&#x27;s more efficient both in speed and compression to LZ4 (at least on the 8 bitters it targets) - includes a neat chart of speed&#x2F;compression trade-offs on a ZX Spectrum with a variety of algorithms
dangalmost 4 years ago
Discussed at the time:<p><i>Modern LZ Compression</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=19064791" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=19064791</a> - Feb 2019 (4 comments)
nathellalmost 4 years ago
&quot;Managing Gigabytes&quot; by Witten et al is _the_ book that got me into the field of compression back in the day. Granted, it’s a bit dated now as there’s been a lot of progress in the field, but I’d still heartily recommend it to newcomers.
nickdothuttonalmost 4 years ago
Only partly related, since it deals with a specific type of data (English language Wikipedia), some of you might enjoy reading about the Hutter Prize: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Hutter_Prize" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Hutter_Prize</a>
akkartikalmost 4 years ago
Does anyone have recommendations for learning the xz format? Particularly as used by the ZIM archival format: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;ZIM_(file_format)" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;ZIM_(file_format)</a>
评论 #28218586 未加载
nayukiover 3 years ago
&gt; First, we just shorten any symbols that are longer than our maximum code length — 11 — to that value. This means that our tree will no longer be a Huffman tree, so we need to do a fixup pass to redistribute the error we&#x27;ve introduced.<p>This can in fact be solved directly and optimally: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Package-merge_algorithm" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Package-merge_algorithm</a> ; <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Huffman_coding#Length-limited_Huffman_coding" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Huffman_coding#Length-limited_...</a>
评论 #28245295 未加载
pwrrrover 3 years ago
I just made a huffman encoder on the c64, for fun. I need understand how you go from the variable length codes of huffman, to suddenly fixed lenght codes, because you don&#x27;t want the codes to be above a certain length. hmm...
meiji163almost 4 years ago
I wonder if LZ would still be standard, if not for the inertia of gzip&#x2F;zip? There are surely better and comparably fast algorithms (paq, ppm, etc.)
评论 #28220019 未加载
评论 #28217541 未加载
评论 #28218196 未加载
评论 #28217610 未加载
评论 #28217574 未加载
评论 #28217389 未加载
rurbanalmost 4 years ago
I had to implement recently an oldstyle lz77 en&#x2F;decoder to handle an old fileformat, and it was surprisingly simple. Even the encoder
user-the-namealmost 4 years ago
&quot;Modern&quot;, but uses <i>Huffman coding</i>? No LZ implementation aiming for high compression in the last decade has used Huffman.
评论 #28216841 未加载
评论 #28216547 未加载