2 ms·
It's not that crazy if you separate it into two parts. A) Use arithmetic coding to encode symbols as fractions instead of using an integer number of bits for e
by Dylan16807 4y ago
It's not that crazy if you separate it into two parts.
A) Use arithmetic coding to encode symbols as fractions instead of using an integer number of bits for each symbol. https://en.wikipedia.org/wiki/Arithmetic_coding https://en.wikipedia.org/wiki/Arithmetic_coding http://www.ws.binghamton.edu/fowler/fowler%20personal%20page/EE523_files/Ch_04%20Arithmetic%20Coding%20%28PPT%29.pdf http://www.ws.binghamton.edu/fowler/fowler%20personal%20page...
B) Shuffle the bits around so that you can use faster math and lookup tables instead of directly operating on fractions.