Zlib
From Just Solve the File Format Problem
(Difference between revisions)
Dan Tobias (Talk | contribs) (→Identification) |
(Added sample files) |
||
| (18 intermediate revisions by 3 users not shown) | |||
| Line 1: | Line 1: | ||
{{FormatInfo | {{FormatInfo | ||
| + | |name=zlib | ||
|formattype=electronic | |formattype=electronic | ||
|subcat=Compression | |subcat=Compression | ||
| + | |released=1995 | ||
}} | }} | ||
| − | '''zlib''' is a | + | {{DISPLAYTITLE:zlib}} |
| + | '''zlib''' format is a container and extension of the [[DEFLATE]] compressed data format. It adds a lightweight container, a small amount of configurability, and error checking. | ||
| − | + | zlib is also the name of a software library that supports the zlib format, and other formats related to DEFLATE. | |
| − | + | There are many formats and protocols that use, or may use, zlib format. A notable example is [[PNG]]. | |
== See also == | == See also == | ||
* [[Adler-32]] | * [[Adler-32]] | ||
| + | * [[DEFLATE]] | ||
== Identification == | == Identification == | ||
| − | The first | + | The first two bytes can help to identify a zlib datastream. The following tests are ordered starting with those that are most certain to be true. |
| − | If the first two bytes are interpreted as a [[Endianness|big-endian]] integer, it is always a multiple of 31. | + | * If the first two bytes are interpreted as a [[Endianness|big-endian]] integer, it is always a multiple of 31. (Always true) |
| + | * The first byte has eight possible values: <code>0x08</code>, <code>0x18</code>, <code>0x28</code>, ..., <code>0x78</code>. (Always true, unless the zlib format has been extended to support compression methods other than Deflate.) | ||
| + | * Almost always, the <code>0x20</code> bit of the second byte is 0. (A value of 1 is legal in zlib v3.3+, but indicates use of a preset dictionary, so the zlib stream is not as portable.) | ||
| + | * Most often, the Deflate window size is 32K, and there is no preset dictionary. With these assumptions, the first byte is <code>0x78</code>, and the second is <code>0x01</code> (if using fastest compression), <code>0x5e</code> (fast), <code>0x9c</code> (default), or <code>0xda</code> (max), with <code>0x9c</code> and <code>0xda</code> being the most common. | ||
| + | |||
| + | == Specifications == | ||
| + | * ZLIB Specification v3.1 - ''No copy found. There are dead links to it in e.g. the [https://zlib.net/fossils/zlib-0.95.tar.gz zlib 0.95] README file.'' | ||
| + | * [https://web.archive.org/web/20030426211434/http://mirror.cs.wisc.edu/pub/mirrors/ghost/gzip_doc/zlib-3.2.doc ZLIB Specification v3.2] (from archive.org) | ||
| + | * RFC 1950: ZLIB Compressed Data Format Specification version 3.3 | ||
== Software == | == Software == | ||
| − | * [ | + | * [https://zlib.net zlib data compression library] |
| + | * [https://github.com/cloudflare/zlib CloudFlare fork of zlib] | ||
| + | * [https://github.com/zlib-ng/zlib-ng zlib-ng] | ||
| + | * Puff - Lightweight decompression software included with [https://zlib.net/ zlib] (contrib/puff) | ||
| + | * [https://github.com/richgel999/miniz miniz] | ||
| + | * [https://github.com/google/zopfli zopfli] - zlib compressor highly optimized for file size | ||
| + | * [http://aluigi.altervista.org/mytoolz.htm#offzip Offzip] | ||
| + | * {{Deark}} | ||
| − | == | + | See also [[DEFLATE#Software]]. |
| − | * | + | |
| + | == Sample files == | ||
| + | * {{DexvertSamples|archive/zlib}} | ||
| − | == | + | == Links == |
* [[Wikipedia:Zlib|Wikipedia article]] | * [[Wikipedia:Zlib|Wikipedia article]] | ||
Latest revision as of 15:06, 28 December 2023
zlib format is a container and extension of the DEFLATE compressed data format. It adds a lightweight container, a small amount of configurability, and error checking.
zlib is also the name of a software library that supports the zlib format, and other formats related to DEFLATE.
There are many formats and protocols that use, or may use, zlib format. A notable example is PNG.
Contents |
[edit] See also
[edit] Identification
The first two bytes can help to identify a zlib datastream. The following tests are ordered starting with those that are most certain to be true.
- If the first two bytes are interpreted as a big-endian integer, it is always a multiple of 31. (Always true)
- The first byte has eight possible values:
0x08,0x18,0x28, ...,0x78. (Always true, unless the zlib format has been extended to support compression methods other than Deflate.) - Almost always, the
0x20bit of the second byte is 0. (A value of 1 is legal in zlib v3.3+, but indicates use of a preset dictionary, so the zlib stream is not as portable.) - Most often, the Deflate window size is 32K, and there is no preset dictionary. With these assumptions, the first byte is
0x78, and the second is0x01(if using fastest compression),0x5e(fast),0x9c(default), or0xda(max), with0x9cand0xdabeing the most common.
[edit] Specifications
- ZLIB Specification v3.1 - No copy found. There are dead links to it in e.g. the zlib 0.95 README file.
- ZLIB Specification v3.2 (from archive.org)
- RFC 1950: ZLIB Compressed Data Format Specification version 3.3
[edit] Software
- zlib data compression library
- CloudFlare fork of zlib
- zlib-ng
- Puff - Lightweight decompression software included with zlib (contrib/puff)
- miniz
- zopfli - zlib compressor highly optimized for file size
- Offzip
- Deark
See also DEFLATE#Software.
