Re: Best existing binary compressor method?

On Feb 18, 6:17 pm, Industrial One <industrial_...@xxxxxxxxxxx> wrote:
On Feb 18, 6:35 am, Einstein <michae...@xxxxxxxxx> wrote:

Arithmetic is the best. Ironically, it is the simplest to code, yet
achieves the best compression ratio.

OTOH, it takes fucking FOREVER.

"Arithmetic" is not a compression routine.

You can make a general statement that data compression is achieved by
combining a model and an encoder. The model looks for redundancy, and
uses the encoder to take advantage of that redundancy. (Yes, you can
nitpick this generalization to death, but it's a useful place to
start.)

Arithmetic coding is used in data compression to encode data very
efficiently, using nearly the optimal number of digits dictated by
entropy of the model. But by itself, it has nothing to do with data
compression, despite the fact that the wikipiedia article says:
"Arithmetic coding is a method for lossless data compression."

You can find arithmetic coding used as an encoder for many different
compression algorithms, and to provide a reasonable answer to this
(unanswerable) question, you need to specify which algorithm you are
suggesting. LZARI? Lossless JPEG? PPMZ? All quite different.

|
| Mark Nelson - http://marknelson.us
|
.