Hacker News Reader

Comment by pjscott | Hacker News Reader

I was skeptical about the claims of competitive compression, but in a handful of experiments it looks like farbfeld+bzip2 got slightly better compression than png (after running through pngcrush). Decoding took an unfortunate number of milliseconds, though.

(Give it a try! The tools are right there in the git repo, and quite easy to use.)

View on HN

I'm sure the designers of PNG were aware that slower entropy coders that gave better compression ratios existed. But there's a reason they didn't choose them: bzip2 is 5x slower than DEFLATE even on the lowest setting. Scanline filtering, on the other hand, is fast, amenable to SIMD, and even parallelizable if the image is properly encoded. It's also easy: the only thing I don't remember off the top of my head is the order ties are broken in the Paeth filter.

bzip2 is an especially poor choice in 2016 given that it's inferior to LZMA in both speed and quality.

thing is, bzip2 is widespread, unlike lzma not everybody has.

on a technical side, lzma is a better choice. I might add a lzma recommendation, however, in soma cases lzma performs worse than bzip2.

also, decompression speeds matter, not compression speeds.

in the end, use what you prefer. the compression is not mandated by the spec. the main use for farbfeld is also rather thought to be as a piping format. it could also be used for storage, but that's not the main point.

It's worth noting that lzma decompression speeds are generally much faster than bzip2, but at a significant increase in memory usage.

> thing is, bzip2 is widespread, unlike lzma not everybody has.

What year is this, 2010? xz/lzma is everywhere.

bzip2 is 5x slower at decompression.

Left over Up over UpperLeft for Paeth

That's a,b,c when your pixels are labeled

    c b
    a d

the same order you write the predictor p = a+b-c.

Reply on news.ycombinator.com