Re: [libpng:feature-requests] #32 Modern compressor

Златовратский Павел <scondo-JGs/[email protected]>
Newsgroups gmane.comp.graphics.png.general
Message-ID <[email protected]>
Well, I proposed LZMA only because it's pretty popular(among other 
high-level compression).
My own favourite is ZPAQ. It has documentation and use some kind of 
virtual machine to keep description of decompression algorithm along 
with data.
http://www.mattmahoney.net/dc/zpaq206.pdf

Here some dirty test of brotli, gzip (with 7-zip) and zpaq on some 
picture from wikipedia:

02.01.2017  07:39         3 396 302 Marmolada,_Italy.jpg
02.01.2017  07:50        20 842 445 Marmolada_none_.png
02.01.2017  08:21        16 256 069 Marmolada_none_IDAT.brotli
02.01.2017  08:19        20 811 891 Marmolada_none_IDAT.chunk
02.01.2017  08:21        30 500 652 Marmolada_none_IDAT.rawchunk
02.01.2017  08:23        19 349 879 Marmolada_none_IDAT.rawchunk.gz
02.01.2017  08:35        19 676 279 Marmolada_none_IDAT.zpaq2
02.01.2017  08:34        12 127 402 Marmolada_none_IDAT.zpaq5
02.01.2017  07:47        15 459 765 Marmolada_paeth_.png
02.01.2017  08:04        13 766 930 Marmolada_paeth_IDAT.brotli
02.01.2017  07:48        15 459 691 Marmolada_paeth_IDAT.chunk
02.01.2017  08:04        30 500 652 Marmolada_paeth_IDAT.rawchunk
02.01.2017  08:07        15 086 510 Marmolada_paeth_IDAT.rawchunk.gz
02.01.2017  08:17        16 379 977 Marmolada_paeth_IDAT.zpaq2
02.01.2017  08:22        12 207 144 Marmolada_paeth_IDAT.zpaq5

Brotli is pretty good on default(max) settings, ZPAQ is not good when 
it's fast, but beat about anyone when it's slow. Also ZPAQ's own context 
model on max level works better without PNG filters.

I think Brotli is fast way to improve compression as it's already RFC 
and based on deflate (so it will be easier to add for developers who 
doesn't rely on external libraries).
ZPAQ on the other side is long way to achieve maximum compression.

02.01.2017 1:40, Chris Lilley пишет:
>
> LZMA is a non-starter. It was considered for WOFF 2 ( a font compression
> scheme used for webfonts) and rejected because there is no adequate
> documentation, just a sample implementation that everyone else must
> reverse engineer.
>
> This is why we used Brotli compression instead. It would be a good
> candidate for PNG 2.0 also.
>


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
png-mng-misc mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/png-mng-misc
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.