FiniteStateEntropy VS libvips

It's obsolete. It's limited to 32KB LZ window with huffman coding. Zstd can use a much larger window (8MB recommended) and a much better entropy coder: https://github.com/Cyan4973/FiniteStateEntropy

tANS block based : FSE

Sounds similar to: https://github.com/Cyan4973/FiniteStateEntropy

https://arxiv.org/abs/1311.2540

> The modern data compression is mainly based on two approaches to entropy coding: Huffman (HC) and arithmetic/range coding (AC). The former is much faster, but approximates probabilities with powers of 2, usually leading to relatively low compression rates. The latter uses nearly exact probabilities - easily approaching theoretical compression rate limit (Shannon entropy), but at cost of much larger computational cost.

FiniteStateEntropy - Two highly efficient compression codecs optimized for modern CPUs. BSD-2-Clause

After some research, I found libvips, a demand-driven, horizontally threaded image processing library. It is designed to run quickly while using as little as memory as possible.

Sharp is a fast and efficient image optimization Node.js module that makes use of the native libvips library.

h2non/bimg can handle both if the underlying libvips is compiled with support for both formats.

if you're using linux (which it doesn't seem so) there's also vispdisp https://github.com/jcupitt/vipsdisp which is based on https://github.com/libvips/libvips which will likely take over how images are decoded in the future for everything, at least methodology wise.

The magic happens by way of a library called Libvips, which contains an ultra-high-speed low-memory image resizer.

imagor uses one of the most efficient image processing library libvips. It is typically 4-8x faster than using the quickest ImageMagick and GraphicsMagick settings.

Try the solutions here: https://github.com/libvips/libvips/issues/2479

in that case maybe take a look at https://github.com/libvips/libvips

libvips v8.6+ or ImageMagick for image analysis and transformations