Originally Posted by Contrastique
That sounds quite odd to me since you will loose half of the size but keep the same qual...Don't know how it exactly works though but you will have to sacrifice something in order to let the files shrink so dramatically. Not saying you are wrong, but it just sounds odd.
It works a bit like this:
Suppose you've got a CD filled with music. Since it's digital it exists of zeros and ones. So an example of this would be:
1111111111111111 <-- Those are 16-bits. Yes, there are. You can count them if you want to.
What a compression does is the following. It will look for repeating paterns in the code. Since there are 16-bits which are the exact same it will save it in another way. A common way to do this is like this:
4 x 1111
Still 16 bits, but written in another way. This you will not have to save the full 16 bits but a substitute of the same thing.
Not only ALAC and FLAC use this, but winRAR, winzip and 7z do the exact same thing. They are looking for repeating paterns and save them in another way.